"""Implements the world class for the SCML2020 world """
import copy
import functools
import itertools
import logging
import math
import numbers
import random
from abc import abstractmethod
from collections import defaultdict, namedtuple
import sys
from dataclasses import dataclass, field
from typing import (
Optional,
Dict,
Any,
Union,
Tuple,
Callable,
List,
Set,
Collection,
Type,
Iterable,
)
import numpy as np
from negmas import (
Contract,
Action,
Breach,
AgentWorldInterface,
Agent,
RenegotiationRequest,
Negotiator,
AgentMechanismInterface,
MechanismState,
Issue,
Entity,
SAONegotiator,
SAOController,
PassThroughSAONegotiator,
)
from negmas.helpers import instantiate, unique_name, get_class, get_full_type_name
from negmas.situated import World, TimeInAgreementMixin, BreachProcessing
from scml.scml2019.utils import _realin
__all__ = [
"FactoryState",
"SCML2020Agent",
"AWI",
"SCML2020World",
"FinancialReport",
"FactoryProfile",
"INFINITE_COST",
"ANY_LINE",
"ANY_STEP",
"NO_COMMAND",
"Factory",
]
ANY_STEP = -1
"""Used to indicate any time-step"""
ANY_LINE = -1
"""Used to indicate any line"""
NO_COMMAND = -1
"""A constant indicating no command is scheduled on a factory line"""
INFINITE_COST = sys.maxsize // 2
"""A constant indicating an invalid cost for lines incapable of running some process"""
ContractInfo = namedtuple(
"ContractInfo", ["q", "u", "product", "is_seller", "partner", "contract"]
)
"""Information about a contract including a pointer to it"""
CompensationRecord = namedtuple(
"CompensationRecord", ["product", "quantity", "money", "seller_bankrupt", "factory"]
)
"""A record of delayed compensation used when a factory goes bankrupt to keep honoring its future contracts to the
limit possible"""
@dataclass
class ExogenousContract:
"""Represents a contract to be revealed at revelation_time to buyer and seller between them that is not agreed upon
through negotiation but is endogenously given"""
product: int
"""Product"""
quantity: int
"""Quantity"""
unit_price: int
"""Unit price"""
time: int
"""Delivery time"""
revelation_time: int
"""Time at which to reveal the contract to both buyer and seller"""
seller: int = -1
"""Seller index in the agents array (-1 means "system")"""
buyer: int = -1
"""Buyer index in the agents array (-1 means "system")"""
[docs]@dataclass
class FinancialReport:
"""A report published periodically by the system showing the financial standing of an agent"""
__slots__ = [
"agent_id",
"step",
"cash",
"assets",
"breach_prob",
"breach_level",
"is_bankrupt",
"agent_name",
]
agent_id: str
"""Agent ID"""
step: int
"""Simulation step at the beginning of which the report was published."""
cash: int
"""Cash in the agent's wallet. Negative numbers indicate liabilities."""
assets: int
"""Value of the products in the agent's inventory @ catalog prices. """
breach_prob: float
"""Number of times the agent breached a contract over the total number of contracts it signed."""
breach_level: float
"""Sum of the agent's breach levels so far divided by the number of contracts it signed."""
is_bankrupt: bool
"""Whether the agent is already bankrupt (i.e. incapable of doing any more transactions)."""
agent_name: str
"""Agent name for printing purposes"""
def __str__(self):
bankrupt = "BANKRUPT" if self.is_bankrupt else ""
return (
f"{self.agent_name} @ {self.step} {bankrupt}: Cash: {self.cash}, Assets: {self.assets}, "
f"breach_prob: {self.breach_prob}, breach_level: {self.breach_level} "
f"{'(BANKRUPT)' if self.is_bankrupt else ''}"
)
[docs]@dataclass
class FactoryProfile:
"""Defines all private information of a factory"""
__slots__ = [
"costs",
"exogenous_sales",
"exogenous_supplies",
"exogenous_sale_prices",
"exogenous_supply_prices",
]
costs: np.ndarray
"""An n_lines * n_processes array giving the cost of executing any process (INVALID_COST indicates infinity)"""
exogenous_sales: np.ndarray
"""A n_steps * n_products array giving guaranteed sales of different products for the whole simulation time"""
exogenous_supplies: np.ndarray
"""A n_steps * n_products array giving guaranteed sales of different products for the whole simulation time"""
exogenous_sale_prices: np.ndarray
"""A n_steps * n_products array giving guaranteed unit prices for the `exogenous_quantities` . It will be zero
for times and products for which there are no guaranteed quantities (i.e. (exogenous_quantities[...] == 0) =>
(exogenous_prices[...] == 0) )"""
exogenous_supply_prices: np.ndarray
"""A n_steps * n_products array giving guaranteed unit prices for the `exogenous_quantities` . It will be zero
for times and products for which there are no guaranteed quantities (i.e. (exogenous_quantities[...] == 0) =>
(exogenous_prices[...] == 0) )"""
@property
def n_lines(self):
return self.costs.shape[0]
@property
def n_products(self):
return self.exogenous_sales.shape[1]
@property
def n_steps(self):
return self.exogenous_sales.shape[0]
@property
def n_processes(self):
return self.costs.shape[1]
@dataclass
class Failure:
"""A production failure"""
__slots__ = ["is_inventory", "line", "step", "process"]
is_inventory: bool
"""True if the cause of failure was insufficient inventory. If False, the cause was insufficient funds. Note that
if both conditions were true, only insufficient funds (is_inventory=False) will be reported."""
line: int
"""The line at which the failure happened"""
step: int
"""The step at which the failure happened"""
process: int
"""The process that failed to execute (if `exogenous_contract_failure` and `is_inventory` , then this will be the
process that would have generated the needed product. and if `exogenous_contract_failure` and not `is_inventory`
, then it is not valid)"""
[docs]@dataclass
class FactoryState:
inventory: np.ndarray
"""An n_products vector giving current quantity of every product in storage"""
balance: int
"""Current balance in the wallet"""
commands: np.ndarray
"""n_steps * n_lines array giving the process scheduled on each line at every step for the
whole simulation"""
inventory_changes: np.ndarray
"""Changes in the inventory in the last step"""
balance_change: int
"""Change in the balance in the last step"""
contracts: List[List[ContractInfo]]
"""The An n_steps list of lists containing the contracts of this agent by time-step"""
@property
def n_lines(self) -> int:
return self.commands.shape[1]
@property
def n_steps(self) -> int:
return self.commands.shape[0]
@property
def n_products(self) -> int:
return len(self.inventory)
@property
def n_processes(self) -> int:
return len(self.inventory) - 1
[docs]class Factory:
"""A simulated factory"""
def __init__(
self,
profile: FactoryProfile,
initial_balance: int,
inputs: np.ndarray,
outputs: np.ndarray,
catalog_prices: np.ndarray,
world: "SCML2020World",
compensate_before_past_debt: bool,
buy_missing_products: bool,
production_buy_missing: bool,
exogenous_buy_missing: bool,
exogenous_penalty: float,
exogenous_no_bankruptcy: bool,
exogenous_no_borrow: bool,
production_penalty: float,
production_no_bankruptcy: bool,
production_no_borrow: bool,
agent_id: str,
agent_name: Optional[str] = None,
confirm_production: bool = True,
initial_inventory: Optional[np.ndarray] = None,
):
self.confirm_production = confirm_production
self.production_buy_missing = production_buy_missing
self.exogenous_buy_missing = exogenous_buy_missing
self.compensate_before_past_debt = compensate_before_past_debt
self.buy_missing_products = buy_missing_products
self.exogenous_penalty = exogenous_penalty
self.exogenous_no_bankruptcy = exogenous_no_bankruptcy
self.exogenous_no_borrow = exogenous_no_borrow
self.production_penalty = production_penalty
self.production_no_bankruptcy = production_no_bankruptcy
self.production_no_borrow = production_no_borrow
self.catalog_prices = catalog_prices
self.initial_balance = initial_balance
self.__profile = profile
self.world = world
self.profile = copy.deepcopy(profile)
"""The readonly factory profile (See `FactoryProfile` )"""
self.commands = NO_COMMAND * np.ones(
(profile.n_steps, profile.n_lines), dtype=int
)
"""An n_steps * n_lines array giving the process scheduled for each line at every step. -1 indicates an empty
line. """
# self.predicted_inventory = profile.exogenous_quantities.copy()
"""An n_steps * n_products array giving the inventory content at different steps. For steps in the past and
present, this is the *actual* value of the inventory at that time. For steps in the future, this is a
*prediction* of the inventory at that step."""
self._balance = initial_balance
"""Current balance"""
self._inventory = (
np.zeros(profile.n_products, dtype=int)
if initial_inventory is None
else initial_inventory
)
"""Current inventory"""
# self.predicted_balance = initial_balance - np.sum(
# profile.exogenous_quantities * profile.exogenous_prices, axis=-1
# )
"""An n_steps vector giving the wallet balance at different steps. For steps in the past and
present, this is the *actual* value of the balance at that time. For steps in the future, this is a
*prediction* of the balance at that step."""
self.agent_id = agent_id
"""A unique ID for the agent owning the factory"""
self.inputs = inputs
"""An n_process array giving the number of inputs needed for each process
(of the product with the same index)"""
self.outputs = outputs
"""An n_process array giving the number of outputs produced by each process
(of the product with the next index)"""
self.inventory_changes = np.zeros(len(inputs) + 1, dtype=int)
"""Changes in the inventory in the last step"""
self.balance_change = 0
"""Change in the balance in the last step"""
self.min_balance = self.world.bankruptcy_limit
"""The minimum balance possible"""
self.is_bankrupt = False
"""Will be true when the factory is bankrupt"""
self.agent_name = (
self.world.agents[agent_id].name
if agent_name is None and world
else agent_name
)
"""SCML2020Agent names used for logging purposes"""
self.contracts: List[List[ContractInfo]] = [[] for _ in range(world.n_steps)]
"""A list of lists of contracts per time-step (len == n_steps)"""
@property
def state(self) -> FactoryState:
return FactoryState(
self._inventory.copy(),
self._balance,
self.commands,
self.inventory_changes,
self.balance_change,
[copy.copy(_.contract) for times in self.contracts for _ in times],
)
@property
def current_inventory(self) -> np.ndarray:
"""Current inventory contents"""
return self._inventory
@property
def current_balance(self) -> int:
"""Current wallet balance"""
return self._balance
[docs] def schedule_production(
self,
process: int,
repeats: int,
step: Union[int, Tuple[int, int]] = ANY_STEP,
line: int = ANY_LINE,
override: bool = True,
method: str = "latest",
partial_ok: bool = False,
) -> Tuple[np.ndarray, np.ndarray]:
"""
Orders production of the given process on the given step and line.
Args:
process: The process index
repeats: How many times to repeat the process
step: The simulation step or a range of steps. The special value ANY_STEP gives the factory the freedom to
schedule production at any step in the present or future.
line: The production line. The special value ANY_LINE gives the factory the freedom to use any line
override: Whether to override any existing commands at that line at that time.
method: When to schedule the command if step was set to a range. Options are latest, earliest, all
partial_ok: If true, it is OK to produce only a subset of repeats
Returns:
Tuple[np.ndarray, np.ndarray] The steps and lines at which production is scheduled.
Remarks:
- You cannot order production in the past or in the current step
- Ordering production, will automatically update inventory and balance for all simulation steps assuming
that this production will be carried out. At the indicated `step` if production was not possible (due
to insufficient funds or insufficient inventory of the input product), the predictions for the future
will be corrected.
"""
if self.is_bankrupt:
return np.empty(0, dtype=int), np.empty(0, dtype=int)
steps, lines = self.available_for_production(
repeats, step, line, override, method
)
if len(steps) < 1:
return np.empty(0, dtype=int), np.empty(0, dtype=int)
if len(steps) < repeats:
if not partial_ok:
return np.empty(0, dtype=int), np.empty(0, dtype=int)
repeats = len(steps)
self.order_production(process, steps[:repeats], lines[:repeats])
return steps, lines
[docs] def order_production(
self, process: int, steps: np.ndarray, lines: np.ndarray
) -> None:
"""
Orders production of the given process
Args:
process: The process to run
steps: The time steps to run the process at as an np.ndarray
lines: The corresponding lines to run the process at
Remarks:
- len(steps) must equal len(lines)
- No checks are done in this function. It is expected to be used after calling `available_for_production`
"""
if self.is_bankrupt:
return
if len(steps) > 0:
self.commands[steps, lines] = process
[docs] def available_for_production(
self,
repeats: int,
step: Union[int, Tuple[int, int]] = ANY_STEP,
line: int = ANY_LINE,
override: bool = True,
method: str = "latest",
) -> Tuple[np.ndarray, np.ndarray]:
"""
Finds available times and lines for scheduling production.
Args:
repeats: How many times to repeat the process
step: The simulation step or a range of steps. The special value ANY_STEP gives the factory the freedom to
schedule production at any step in the present or future.
line: The production line. The special value ANY_LINE gives the factory the freedom to use any line
override: Whether to override any existing commands at that line at that time.
method: When to schedule the command if step was set to a range. Options are latest, earliest, all
Returns:
Tuple[np.ndarray, np.ndarray] The steps and lines at which production is scheduled.
Remarks:
- You cannot order production in the past or in the current step
- Ordering production, will automatically update inventory and balance for all simulation steps assuming
that this production will be carried out. At the indicated `step` if production was not possible (due
to insufficient funds or insufficient inventory of the input product), the predictions for the future
will be corrected.
"""
if self.is_bankrupt:
return np.empty(shape=0, dtype=int), np.empty(shape=0, dtype=int)
current_step = self.world.current_step
if not isinstance(step, tuple):
if step < 0:
step = (current_step, self.profile.n_steps)
else:
step = (step, step + 1)
else:
step = (step[0], step[1] + 1)
step = (max(current_step, step[0]), step[1])
if step[1] <= step[0]:
return np.empty(shape=0, dtype=int), np.empty(shape=0, dtype=int)
if override:
if line < 0:
steps, lines = np.nonzero(
self.commands[step[0] : step[1], :] >= NO_COMMAND
)
else:
steps = np.nonzero(
self.commands[step[0] : step[1], line] >= NO_COMMAND
)[0]
lines = [line]
else:
if line < 0:
steps, lines = np.nonzero(
self.commands[step[0] : step[1], :] == NO_COMMAND
)
else:
steps = np.nonzero(
self.commands[step[0] : step[1], line] == NO_COMMAND
)[0]
lines = [line]
steps += step[0]
possible = min(repeats, len(steps))
if possible < repeats:
return np.empty(shape=0, dtype=int), np.empty(shape=0, dtype=int)
if method.startswith("l"):
steps, lines = steps[-possible + 1 :], lines[-possible + 1 :]
elif method == "all":
pass
else:
steps, lines = steps[:possible], lines[:possible]
return steps, lines
[docs] def cancel_production(self, step: int, line: int) -> bool:
"""
Cancels pre-ordered production given that it did not start yet.
Args:
step: Step to cancel at
line: Line to cancel at
Returns:
True if step >= self.current_step
Remarks:
- Cannot cancel a process in the past or present.
"""
if self.is_bankrupt:
return False
if step < self.world.current_step or line < 0:
return False
self.commands[step, line] = NO_COMMAND
return True
[docs] def step(
self, accepted_sales: np.ndarray, accepted_supplies: np.ndarray
) -> List[Failure]:
"""
Override this method to modify stepping logic.
Args:
accepted_sales: Sales per product accepted by the factory manager
accepted_supplies: Supplies per product accepted by the factory manager
Returns:
"""
if self.is_bankrupt:
return []
step = self.world.current_step
profile = self.__profile
failures = []
initial_balance = self._balance
initial_inventory = self._inventory.copy()
# buy guaranteed supplies as much as possible
# if it is possible to pay for all the supplies, do that directly without checks
# otherwise do normal transactions to check for bankruptcy (either way we do not report breaches)
if np.max(accepted_supplies) > 0:
prices = profile.exogenous_supply_prices[step, :] * accepted_supplies
supply_money = np.sum(prices)
if self._balance - supply_money >= self.min_balance:
self._balance -= supply_money
self._inventory += accepted_supplies
else:
for p, (q, u) in enumerate(
zip(
accepted_supplies.tolist(),
profile.exogenous_supply_prices[step, :].tolist(),
)
):
self.buy(
p,
q,
u,
self.exogenous_buy_missing,
self.exogenous_penalty,
self.exogenous_no_bankruptcy,
self.exogenous_no_borrow,
)
if self.is_bankrupt:
break
if self.is_bankrupt:
return []
# Sell guaranteed sales as much as possible
if np.max(accepted_sales) > 0:
in_inventory = (self._inventory - accepted_sales) >= 0
if np.all(in_inventory):
self._balance += np.sum(
accepted_sales * profile.exogenous_sale_prices[step, :]
)
self._inventory -= accepted_sales
else:
for p, (q, u) in enumerate(
zip(
accepted_sales.tolist(),
profile.exogenous_sale_prices[step, :].tolist(),
)
):
if q < 1:
continue
self.buy(
p,
-q,
u,
self.exogenous_buy_missing,
self.exogenous_penalty,
self.exogenous_no_bankruptcy,
self.exogenous_no_borrow,
)
if self.is_bankrupt:
break
if self.is_bankrupt:
return []
if self.confirm_production:
self.commands[step, :] = self.world.agents[
self.agent_id
].confirm_production(
self.commands[step, :],
self.current_balance,
self.current_inventory.copy(),
)
# do production
for line in np.nonzero(self.commands[step, :] != NO_COMMAND)[0]:
p = self.commands[step, line]
cost = profile.costs[line, p]
ins, outs = self.inputs[p], self.outputs[p]
# if execution will lead to bankruptcy or the cost is infinite, ignore this command
if self._balance - cost < self.min_balance or cost == INFINITE_COST:
failures.append(
Failure(is_inventory=False, line=line, step=step, process=p)
)
# self._register_failure(step, p, cost, ins, outs)
continue
inp, outp = p, p + 1
# if we do not have enough inputs, ignore this command
if self._inventory[inp] < ins:
failures.append(
Failure(is_inventory=True, line=line, step=step, process=p)
)
continue
# execute the command
self.store(
inp,
-ins,
0,
self.production_buy_missing,
self.production_penalty,
self.production_no_bankruptcy,
self.production_no_borrow,
)
self.store(
outp,
outs,
0,
self.production_buy_missing,
self.production_penalty,
self.production_no_bankruptcy,
self.production_no_borrow,
)
assert self._balance >= self.min_balance
assert np.min(self._inventory) >= 0
self.inventory_changes = self._inventory - initial_inventory
self.balance_change = self._balance - initial_balance
return failures
[docs] def store(
self,
product: int,
quantity: int,
unit_price: int,
buy_missing: bool,
penalty: float,
no_bankruptcy: bool = False,
no_borrowing: bool = False,
) -> int:
"""
Stores the given amount of product (signed) to the factory.
Args:
product: Product
quantity: quantity to store/take out (-ve means take out)
unit_price: Unit price
buy_missing: If the quantity is negative and not enough product exists in the market, it buys the product
from the spot-market at an increased price of penalty
penalty: The fraction of unit_price added because we are buying from the spot market. Only effectivec if
quantity is negative and not enough of the product exists in the inventory
no_bankruptcy: Never bankrupt the agent on this transaction
no_borrowing: Never borrow for this transaction
Returns:
The quantity actually stored or taken out (always positive)
"""
if self.is_bankrupt:
self.world.logwarning(
f"{self.agent_name} received a transaction "
f"(product: {product}, q: {quantity}, u:{unit_price}) after being bankrupt"
)
return 0
available = self._inventory[product]
if available + quantity >= 0:
self._inventory[product] += quantity
self.inventory_changes[product] += quantity
return quantity if quantity > 0 else -quantity
# we have an inventory breach here. We know that quantity < 0
assert quantity < 0
quantity = -quantity
if not buy_missing:
# if we are not buying from the spot market, pay the penalty for missing products and transfer all available
to_pay = int(
np.ceil(penalty * (quantity - available) / quantity)
* max(self.catalog_prices[product], unit_price)
)
self.pay(to_pay, no_bankruptcy, no_borrowing)
self._inventory[product] = 0
self.inventory_changes[product] -= available
return available
# we have an inventory breach and should try to buy missing quantity from the spot market
real_price = (quantity - available) * max(
self.catalog_prices[product], unit_price
)
to_pay = int(np.ceil(real_price * (1 + penalty)))
paid = int(self.pay(to_pay, no_bankruptcy, no_borrowing) / (1 + penalty))
paid_for = paid // unit_price
assert self._inventory[product] + paid_for >= 0, (
f"{self.agent_name} had {self._inventory[product]} and paid for {paid_for} ("
f"original quantity {quantity})"
)
self._inventory[product] += paid_for
self.inventory_changes[product] += paid_for
return self.store(
product, -quantity, unit_price, False, penalty, no_bankruptcy, no_borrowing
)
[docs] def buy(
self,
product: int,
quantity: int,
unit_price: int,
buy_missing: bool,
penalty: float,
no_bankruptcy: bool = False,
no_borrowing: bool = False,
) -> Tuple[int, int]:
"""
Executes a transaction to buy/sell involving adding quantity and paying price (both are signed)
Args:
product: The product transacted on
quantity: The quantity (added)
unit_price: The unit price (paid)
buy_missing: If true, attempt buying missing products from the spot market
penalty: The penalty as a fraction to be paid for breaches
no_bankruptcy: If true, this transaction can never lead to bankruptcy
no_borrowing: If true, this transaction can never lead to borrowing
Returns:
Tuple[int, int] The actual quantities bought and the total cost
"""
if self.is_bankrupt:
self.world.logwarning(
f"{self.agent_name} received a transaction "
f"(product: {product}, q: {quantity}, u:{unit_price}) after being bankrupt"
)
return 0, 0
if quantity < 0:
# that is a sell contract
taken = self.store(
product,
quantity,
unit_price,
buy_missing,
penalty,
no_bankruptcy,
no_borrowing,
)
paid = self.pay(-taken * unit_price, no_bankruptcy, no_borrowing)
return taken, paid
# that is a buy contract
paid = self.pay(quantity * unit_price, no_bankruptcy, no_borrowing)
stored = self.store(
product,
quantity * paid // unit_price,
unit_price,
buy_missing,
penalty,
no_bankruptcy,
no_borrowing,
)
return stored, paid
[docs] def pay(
self, money: int, no_bankruptcy: bool = False, no_borrowing: bool = False
) -> int:
"""
Pays money
Args:
money: amount to pay
no_bankruptcy: If true, this transaction can never lead to bankruptcy
no_borrowing: If true, this transaction can never lead to borrowing
Returns:
The amount actually paid
"""
if self.is_bankrupt:
self.world.logwarning(
f"{self.agent_name} was asked to pay {money} after being bankrupt"
)
return 0
new_balance = self._balance - money
if new_balance < self.min_balance:
if no_bankruptcy:
money = self._balance - self.min_balance
else:
money = self.bankrupt(money)
elif no_borrowing and new_balance < 0:
money = self._balance
self._balance -= money
self.balance_change -= money
return money
[docs] def bankrupt(self, required: int) -> int:
"""
Bankruptcy processing for the given agent
Args:
required: The money required after the bankruptcy is processed
Returns:
The amount of money to pay back to the entity that should have been paid `money`
"""
self.world.logdebug(
f"bankrupting {self.agent_name} (has: {self._balance}, needs {required})"
)
# sell everything on the agent's inventory
total = int(np.sum(self._inventory * self.catalog_prices))
pay_back = min(required, total)
available = total - required
# If past debt is paid before compensation pay it
original_balance = self._balance
if not self.compensate_before_past_debt:
available += original_balance
self.world.compensate(available, self)
self.is_bankrupt = True
return pay_back
[docs]class AWI(AgentWorldInterface):
"""The Agent SCML2020World Interface for SCML2020 world allowing a single process per agent"""
# --------
# Actions
# --------
[docs] def request_negotiations(
self,
is_buy: bool,
product: int,
quantity: Union[int, Tuple[int, int]],
unit_price: Union[int, Tuple[int, int]],
time: Union[int, Tuple[int, int]],
controller: SAOController,
partners: List[str] = None,
extra: Dict[str, Any] = None,
) -> bool:
"""
Requests a negotiation
Args:
is_buy: If True the negotiation is about buying otherwise selling.
product: The product to negotiate about
quantity: The minimum and maximum quantities. Passing a single value q is equivalent to passing (q,q)
unit_price: The minimum and maximum unit prices. Passing a single value u is equivalent to passing (u,u)
time: The minimum and maximum delivery step. Passing a single value t is equivalent to passing (t,t)
controller: The controller to manage the complete set of negotiations
partners: ID of all the partners to negotiate with.
extra: Extra information accessible through the negotiation annotation to the caller
Returns:
`True` if the partner accepted and the negotiation is ready to start
Remarks:
- All negotiations will use the following issues **in order**: quantity, time, unit_price
- Negotiations with bankrupt agents or on invalid products (see next point) will be automatically rejected
- Valid products for a factory are the following (any other products are not valid):
1. Buying an input product (i.e. product $\in$ `my_input_products`
1. Seeling an output product (i.e. product $\in$ `my_output_products`
"""
if extra is None:
extra = dict()
if (product not in self.my_input_products and is_buy) or (
product not in self.my_output_products and not is_buy
):
self._world.logwarning(
f"{self.agent.name} requested negotiation on {product} "
f"({'buying' if is_buy else 'selling'}) but this is not in "
f"its ({'inputs' if is_buy else 'outputs'})"
)
return False
if partners is None:
partners = (
self.all_suppliers[product] if is_buy else self.all_consumers[product]
)
negotiators = [
controller.create_negotiator(PassThroughSAONegotiator) for _ in partners
]
results = [
self.request_negotiation(
is_buy, product, quantity, unit_price, time, partner, negotiator, extra
)
if not self._world.a2f[partner].is_bankrupt
else False
for partner, negotiator in zip(partners, negotiators)
]
return any(results)
[docs] def request_negotiation(
self,
is_buy: bool,
product: int,
quantity: Union[int, Tuple[int, int]],
unit_price: Union[int, Tuple[int, int]],
time: Union[int, Tuple[int, int]],
partner: str,
negotiator: SAONegotiator,
extra: Dict[str, Any] = None,
) -> bool:
"""
Requests a negotiation
Args:
is_buy: If True the negotiation is about buying otherwise selling.
product: The product to negotiate about
quantity: The minimum and maximum quantities. Passing a single value q is equivalent to passing (q,q)
unit_price: The minimum and maximum unit prices. Passing a single value u is equivalent to passing (u,u)
time: The minimum and maximum delivery step. Passing a single value t is equivalent to passing (t,t)
partner: ID of the partner to negotiate with.
negotiator: The negotiator to use for this negotiation (if the partner accepted to negotiate)
extra: Extra information accessible through the negotiation annotation to the caller
Returns:
`True` if the partner accepted and the negotiation is ready to start
Remarks:
- All negotiations will use the following issues **in order**: quantity, time, unit_price
- Negotiations with bankrupt agents or on invalid products (see next point) will be automatically rejected
- Valid products for a factory are the following (any other products are not valid):
1. Buying an input product (i.e. product $\in$ `my_input_products`
1. Seeling an output product (i.e. product $\in$ `my_output_products`
"""
if extra is None:
extra = dict()
if (product not in self.my_input_products and is_buy) or (
product not in self.my_output_products and not is_buy
):
self._world.logwarning(
f"{self.agent.name} requested negotiation on {product} "
f"({'buying' if is_buy else 'selleing'}) but this is not in "
f"its ({'inputs' if is_buy else 'outputs'})"
)
return False
if self._world.a2f[partner].is_bankrupt:
return False
def values(x: Union[int, Tuple[int, int]]):
if not isinstance(x, Iterable):
return int(x), int(x)
return int(x[0]), int(x[1])
self._world.logdebug(
f"{self.agent.name} requested to {'buy' if is_buy else 'sell'} {product} to {partner}"
f" q: {quantity}, u: {unit_price}, t: {time}"
)
annotation = {
"product": product,
"is_buy": is_buy,
"buyer": self.agent.id if is_buy else partner,
"seller": partner if is_buy else self.agent.id,
"caller": self.agent.id,
}
issues = [
Issue(values(quantity), name="quantity", value_type=int),
Issue(values(time), name="time", value_type=int),
Issue(values(unit_price), name="unit_price", value_type=int),
]
partners = [self.agent.id, partner]
extra["negotiator_id"] = negotiator.id
req_id = self.agent.create_negotiation_request(
issues=issues,
partners=partners,
negotiator=negotiator,
annotation=annotation,
extra=dict(**extra),
)
return self.request_negotiation_about(
issues=issues, partners=partners, req_id=req_id, annotation=annotation
)
[docs] def schedule_production(
self,
process: int,
repeats: int,
step: Union[int, Tuple[int, int]] = ANY_STEP,
line: int = ANY_LINE,
override: bool = True,
method: str = "latest",
) -> Tuple[np.ndarray, np.ndarray]:
"""
Orders the factory to run the given process at the given line at the given step
Args:
process: The process to run
repeats: How many times to repeat the process
step: The simulation step or a range of steps. The special value ANY_STEP gives the factory the freedom to
schedule production at any step in the present or future.
line: The production line. The special value ANY_LINE gives the factory the freedom to use any line
override: Whether to override existing production commands or not
method: When to schedule the command if step was set to a range. Options are latest, earliest
Returns:
Tuple[int, int] giving the steps and lines at which production is scheduled.
Remarks:
- The step cannot be in the past. Production can only be ordered for current and future steps
- ordering production of process -1 is equivalent of `cancel_production` only if both step and line are
given
"""
return self._world.a2f[self.agent.id].schedule_production(
process, repeats, step, line, override, method
)
[docs] def order_production(
self, process: int, steps: np.ndarray, lines: np.ndarray
) -> None:
"""
Orders production of the given process
Args:
process: The process to run
steps: The time steps to run the process at as an np.ndarray
lines: The corresponding lines to run the process at
Remarks:
- len(steps) must equal len(lines)
- No checks are done in this function. It is expected to be used after calling `available_for_production`
"""
return self._world.a2f[self.agent.id].order_production(process, steps, lines)
[docs] def available_for_production(
self,
repeats: int,
step: Union[int, Tuple[int, int]] = ANY_STEP,
line: int = ANY_LINE,
override: bool = True,
method: str = "latest",
) -> Tuple[np.ndarray, np.ndarray]:
"""
Finds available times and lines for scheduling production.
Args:
repeats: How many times to repeat the process
step: The simulation step or a range of steps. The special value ANY_STEP gives the factory the freedom to
schedule production at any step in the present or future.
line: The production line. The special value ANY_LINE gives the factory the freedom to use any line
override: Whether to override any existing commands at that line at that time.
method: When to schedule the command if step was set to a range. Options are latest, earliest, all
Returns:
Tuple[np.ndarray, np.ndarray] The steps and lines at which production is scheduled.
Remarks:
- You cannot order production in the past or in the current step
- Ordering production, will automatically update inventory and balance for all simulation steps assuming
that this production will be carried out. At the indicated `step` if production was not possible (due
to insufficient funds or insufficient inventory of the input product), the predictions for the future
will be corrected.
"""
return self._world.a2f[self.agent.id].available_for_production(
repeats, step, line, override, method
)
[docs] def cancel_production(self, step: int, line: int) -> bool:
"""
Cancels any production commands on that line at this step
Args:
step: The step to cancel production at (must be in the future).
line: The production line
Returns:
success/failure
Remarks:
- The step cannot be in the past or the current step. Cancellation can only be ordered for future steps
"""
return self._world.a2f[self.agent.id].cancel_production(step, line)
# ---------------------
# Information Gathering
# ---------------------
@property
def state(self) -> FactoryState:
"""Receives the factory state"""
return self._world.a2f[self.agent.id].state
@property
def profile(self) -> FactoryProfile:
"""Gets the profile (static private information) associated with the agent"""
profile = self._world.a2f[self.agent.id].profile
s = min(self.n_steps, self.current_step+self._world.exogenous_horizon)
return FactoryProfile(
profile.costs, profile.exogenous_sales[:s], profile.exogenous_supplies[:s], profile.exogenous_sale_prices[:s]
, profile.exogenous_supply_prices[:s]
)
@property
def all_suppliers(self) -> List[List[str]]:
"""Returns a list of agent IDs for all suppliers for every product"""
return self._world.suppliers
@property
def all_consumers(self) -> List[List[str]]:
"""Returns a list of agent IDs for all consumers for every product"""
return self._world.consumers
@property
def catalog_prices(self) -> np.ndarray:
"""Returns the catalog prices of all products"""
return self._world.catalog_prices
@property
def inputs(self) -> np.ndarray:
"""Returns the number of inputs to every production process"""
return self._world.process_inputs
@property
def outputs(self) -> np.ndarray:
"""Returns the number of outputs to every production process"""
return self._world.process_outputs
@property
def n_products(self) -> int:
"""Returns the number of products in the system"""
return len(self._world.catalog_prices)
@property
def n_processes(self) -> int:
"""Returns the number of processes in the system"""
return self.n_products - 1
@property
def my_input_product(self) -> int:
"""Returns a list of products that are inputs to at least one process the agent can run"""
return self._world.agent_inputs[self.agent.id][0]
@property
def my_output_product(self) -> int:
"""Returns a list of products that are outputs to at least one process the agent can run"""
return self._world.agent_outputs[self.agent.id][0]
@property
def my_input_products(self) -> np.ndarray:
"""Returns a list of products that are inputs to at least one process the agent can run"""
return self._world.agent_inputs[self.agent.id]
@property
def my_output_products(self) -> np.ndarray:
"""Returns a list of products that are outputs to at least one process the agent can run"""
return self._world.agent_outputs[self.agent.id]
@property
def my_suppliers(self) -> List[str]:
"""Returns a list of IDs for all of the agent's suppliers (agents that can supply at least one product it may
need).
Remarks:
- If the agent have multiple input products, suppliers of a specific product $p$ can be found using:
**self.all_suppliers[p]**.
"""
return self._world.agent_suppliers[self.agent.id]
@property
def my_consumers(self) -> List[str]:
"""Returns a list of IDs for all the agent's consumers (agents that can consume at least one product it may
produce).
Remarks:
- If the agent have multiple output products, consumers of a specific product $p$ can be found using:
**self.all_consumers[p]**.
"""
return self._world.agent_consumers[self.agent.id]
@property
def n_lines(self) -> int:
"""The number of lines in the corresponding factory. You can read `state` to get this among other information"""
return self.state.n_lines
@property
def n_products(self) -> int:
"""Number of products in the world"""
return self.state.n_products
@property
def n_processes(self) -> int:
"""Number of processes in the world"""
return self.state.n_processes
[docs]class SCML2020Agent(Agent):
"""Base class for all SCML2020 agents (factory managers)"""
def __init__(self, **kwargs):
super().__init__(**kwargs)
def _respond_to_negotiation_request(
self,
initiator: str,
partners: List[str],
issues: List[Issue],
annotation: Dict[str, Any],
mechanism: AgentMechanismInterface,
role: Optional[str],
req_id: Optional[str],
) -> Optional[Negotiator]:
return self.respond_to_negotiation_request(
initiator, issues, annotation, mechanism
)
[docs] def set_renegotiation_agenda(
self, contract: Contract, breaches: List[Breach]
) -> Optional[RenegotiationRequest]:
return None
[docs] def respond_to_renegotiation_request(
self, contract: Contract, breaches: List[Breach], agenda: RenegotiationRequest
) -> Optional[Negotiator]:
return None
[docs] def on_neg_request_rejected(self, req_id: str, by: Optional[List[str]]):
pass
[docs] def on_neg_request_accepted(self, req_id: str, mechanism: AgentMechanismInterface):
pass
[docs] @abstractmethod
def on_contract_nullified(
self, contract: Contract, compensation_money: int, new_quantity: int
) -> None:
"""
Called whenever a contract is nullified (because the partner is bankrupt)
Args:
contract: The contract being nullified
compensation_money: The compensation money that is already added to the agent's wallet
new_quantity: The new quantity that will actually be executed for this contract at its delivery time.
Remarks:
- compensation_money and new_quantity will never be both nonzero
- compensation_money and new_quantity may both be zero which means that the contract will be cancelled
without compensation
- compensation_money will be nonzero iff immediate_compensation is enabled for this world
"""
[docs] @abstractmethod
def on_failures(self, failures: List[Failure]) -> None:
"""
Called whenever there are failures either in production or in execution of guaranteed transactions
Args:
failures: A list of `Failure` s.
"""
[docs] @abstractmethod
def confirm_exogenous_sales(
self, quantities: np.ndarray, unit_prices: np.ndarray
) -> np.ndarray:
"""
Called to confirm the amount of guaranteed sales the agent is willing to accept
Args:
quantities: An n_products vector giving the maximum quantity that can sold (without negotiation)
unit_prices: An n_products vector giving the guaranteed unit prices
Returns:
An n_products vector specifying the quantities to be sold (up to the given `quantities` limit).
"""
[docs] @abstractmethod
def confirm_exogenous_supplies(
self, quantities: np.ndarray, unit_prices: np.ndarray
) -> np.ndarray:
"""
Called to confirm the amount of guaranteed supplies the agent is willing to accept
Args:
quantities: An n_products vector giving the maximum quantity that can bought (without negotiation)
unit_prices: An n_products vector giving the guaranteed unit prices
Returns:
An n_products vector specifying the quantities to be bought (up to the given `quantities` limit).
"""
[docs] @abstractmethod
def respond_to_negotiation_request(
self,
initiator: str,
issues: List[Issue],
annotation: Dict[str, Any],
mechanism: AgentMechanismInterface,
) -> Optional[Negotiator]:
"""
Called whenever another agent requests a negotiation with this agent.
Args:
initiator: The ID of the agent that requested this negotiation
issues: Negotiation issues
annotation: Annotation attached with this negotiation
mechanism: The `AgentMechanismInterface` interface to the mechanism to be used for this negotiation.
Returns:
None to reject the negotiation, otherwise a negotiator
"""
[docs] def confirm_production(
self, commands: np.ndarray, balance: int, inventory
) -> np.ndarray:
"""
Called just before production starts at every time-step allowing the agent to change what is to be
produced in its factory
Args:
commands: an n_lines vector giving the process to be run at every line (NO_COMMAND indicates nothing to be
processed
balance: The current balance of the factory
inventory: an n_products vector giving the number of items available in the inventory of every product type.
Returns:
an n_lines vector giving the process to be run at every line (NO_COMMAND indicates nothing to be
processed
Remarks:
- The inventory will contain zero items of all products that the factory does not buy or sell
- The default behavior is to just retrun commands confirming production of everything.
"""
return commands
def integer_cut(n: int, l: int, l_m: Union[int, List[int]]) -> List[int]:
"""
Generates l random integers that sum to n where each of them is at least l_m
Args:
n: total
l: number of levels
l_m: minimum per level
Returns:
"""
if not isinstance(l_m, Iterable):
l_m = [l_m] * l
sizes = np.asarray(l_m)
if n < sizes.sum():
raise ValueError(
f"Cannot generate {l} numbers summing to {n} with a minimum summing to {sizes.sum()}"
)
while sizes.sum() < n:
sizes[random.randint(0, l - 1)] += 1
return sizes.tolist()
def realin(rng: Union[Tuple[float, float], float]) -> float:
"""
Selects a random number within a range if given or the input if it was a float
Args:
rng: Range or single value
Returns:
the real within the given range
"""
if isinstance(rng, float):
return rng
if abs(rng[1] - rng[0]) < 1e-8:
return rng[0]
return rng[0] + random.random() * (rng[1] - rng[0])
def intin(rng: Union[Tuple[int, int], int]) -> int:
"""
Selects a random number within a range if given or the input if it was an int
Args:
rng: Range or single value
Returns:
the int within the given range
"""
if isinstance(rng, int):
return rng
if rng[0] == rng[1]:
return rng[0]
return random.randint(rng[0], rng[1])
def make_array(x: Union[np.ndarray, Tuple[int, int], int], n, dtype=int) -> np.ndarray:
"""Creates an array with the given choices"""
if not isinstance(x, Iterable):
return np.ones(n, dtype=dtype) * x
if isinstance(x, tuple) and len(x) == 2:
if dtype == int:
return np.random.randint(x[0], x[1] + 1, n, dtype=dtype)
return x[0] + np.random.rand(n) * (x[1] - x[0])
x = list(x)
if len(x) == n:
return np.array(x)
return np.array(list(random.choices(x, k=n)))
class _SystemAgent(SCML2020Agent):
"""Implements an agent for handling system operations"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.id = "SYSTEM"
self.name = "SYSTEM"
def on_contract_nullified(self, contract: Contract, compensation_money: int, new_quantity: int) -> None:
pass
def on_failures(self, failures: List[Failure]) -> None:
pass
def confirm_exogenous_sales(self, quantities: np.ndarray, unit_prices: np.ndarray) -> np.ndarray:
return quantities
def confirm_exogenous_supplies(self, quantities: np.ndarray, unit_prices: np.ndarray) -> np.ndarray:
return quantities
def respond_to_negotiation_request(self, initiator: str, issues: List[Issue], annotation: Dict[str, Any],
mechanism: AgentMechanismInterface) -> Optional[Negotiator]:
pass
def step(self):
pass
def init(self):
pass
def on_negotiation_failure(self, partners: List[str], annotation: Dict[str, Any],
mechanism: AgentMechanismInterface, state: MechanismState) -> None:
pass
def on_negotiation_success(self, contract: Contract, mechanism: AgentMechanismInterface) -> None:
pass
def on_contract_executed(self, contract: Contract) -> None:
pass
def on_contract_breached(self, contract: Contract, breaches: List[Breach], resolution: Optional[Contract]) -> None:
pass
def sign_all_contracts(self, contracts: List[Contract]) -> List[Optional[str]]:
"""Signs all contracts"""
return [self.id] * len(contracts)
[docs]class SCML2020World(TimeInAgreementMixin, World):
"""A Supply Chain SCML2020World simulation as described for the SCML league of ANAC @ IJCAI 2020.
Args:
process_inputs: An n_processes vector specifying the number of inputs from each product needed to execute
each process.
process_outputs: An n_processes vector specifying the number of inputs from each product generated by
executing each process.
catalog_prices: An n_products vector (i.e. n_processes+1 vector) giving the catalog price of all products
profiles: An n_agents list of `FactoryProfile` objects specifying the private profile of the factory
associated with each agent.
agent_types: An n_agents list of strings/ `SCML2020Agent` classes specifying the type of each agent
agent_params: An n_agents dictionaries giving the parameters of each agent
initial_balance: The initial balance in each agent's wallet. All agents will start with this same value.
breach_penalty: The total penalty paid upon a breach will be calculated as (breach_level * breach_penalty *
contract_quantity * contract_unit_price).
exogenous_supply_limit: An n_steps * n_products array giving the total supply available of each product over time.
Only affects guaranteed supply.
exogenous_sales_limit: An n_steps * n_products array giving the total sales to happen for each product over time.
Only affects guaranteed sales.
financial_report_period: The number of steps between financial reports. If < 1, it is a fraction of n_steps
borrow_on_breach: If true, agents will be forced to borrow money on breach as much as possible to honor the
contract
interest_rate: The interest at which loans grow over time (it only affect a factory when its balance is
negative)
bankruptcy_limit: The maximum amount that be be borrowed (including interest). The balance of any factory cannot
go lower than - borrow_limit or the agent will go bankrupt immediately
compensation_fraction: Fraction of a contract to be compensated (at most) if a partner goes bankrupt. Notice
that this fraction is not guaranteed because the bankrupt agent may not have enough
assets to pay all of its standing contracts to this level of compensation. In such
cases, a smaller fraction will be used.
compensate_immediately: If true, compensation will happen immediately when an agent goes bankrupt and in
in money. This means that agents with contracts involving the bankrupt agent will
just have these contracts be nullified and receive monetary compensation immediately
. If false, compensation will not happen immediately but at the contract execution
time. In this case, agents with contracts involving the bankrupt agent will be
informed of the compensation fraction (instead of the compensation money) at the
time of bankruptcy and will receive the compensation in kind (money if they are
sellers and products if they are buyers) at the normal execution time of the
contract. In the special case of no-compensation (i.e. `compensation_fraction` is
zero or the bankrupt agent has no assets), the two options will behave similarity.
compensate_before_past_debt: If true, then compensations will be paid before past debt is considered,
otherwise, the money from liquidating bankrupt agents will first be used to
pay past debt then whatever remains will be used for compensation. Notice that
in all cases, the trigger of bankruptcy will be paid before compensation and
past debts.
exogenous_force_max: If true, agents are not asked to confirm guaranteed transactions and they
are carried out up to bankruptcy
exogenous_no_borrow: If true, agents will not borrow if they fail to satisfy an external transaction. The
transaction will just fail silently
exogenous_no_bankruptcy: If true, agents will not go bankrupt because of an external transaction. The
transaction will just fail silently
exogenous_penalty: The penalty paid for failure to honor external contracts
exogenous_horizon: The horizon for revealing external contracts
production_no_borrow: If true, agents will not borrow if they fail to satisfy its production need to execute
a scheduled production command
production_no_bankruptcy: If true, agents will not go bankrupt because of an production related transaction.
production_penalty: The penalty paid when buying from spot-market to satisfy production needs
production_confirm: If true, the factory will confirm running processes at every time-step just before running
them by calling `confirm_production` on the agent controlling it.
compact: If True, no logs will be kept and the whole simulation will use a smaller memory footprint
n_steps: Number of simulation steps (can be considered as days).
time_limit: Total time allowed for the complete simulation in seconds.
neg_n_steps: Number of negotiation steps allowed for all negotiations.
neg_time_limit: Total time allowed for a complete negotiation in seconds.
neg_step_time_limit: Total time allowed for a single step of a negotiation. in seconds.
negotiation_speed: The number of negotiation steps that pass in every simulation step. If 0, negotiations
will be guaranteed to finish within a single simulation step
signing_delay: The number of simulation steps to pass between a contract is concluded and signed
name: The name of the simulations
**kwargs: Other parameters that are passed directly to `SCML2020World` constructor.
"""
def __init__(
self,
# SCML2020 specific parameters
process_inputs: np.ndarray,
process_outputs: np.ndarray,
catalog_prices: np.ndarray,
profiles: List[FactoryProfile],
agent_types: List[Type[SCML2020Agent]],
agent_params: List[Dict[str, Any]] = None,
exogenous_contracts: Collection[ExogenousContract] = (),
initial_balance: Union[np.ndarray, Tuple[int, int], int] = 1000,
# breach processing parameters
buy_missing_products=False,
borrow_on_breach=True,
bankruptcy_limit=1.0,
breach_penalty=0.15,
interest_rate=0.05,
financial_report_period=5,
# compensation parameters (for victims of bankrupt agents)
compensation_fraction=1.0,
compensate_immediately=False,
compensate_before_past_debt=True,
# external contracts parameters
exogenous_force_max=True,
exogenous_buy_missing=False,
exogenous_no_borrow=False,
exogenous_no_bankruptcy=False,
exogenous_penalty=0.15,
exogenous_supply_limit: np.ndarray = None,
exogenous_sales_limit: np.ndarray = None,
exogenous_horizon: int = None,
# production failure parameters
production_confirm=True,
production_buy_missing=False,
production_no_borrow=True,
production_no_bankruptcy=False,
production_penalty=0.15,
# General SCML2020World Parameters
compact=False,
no_logs=False,
n_steps=1000,
time_limit=60 * 90,
# mechanism params
neg_n_steps=20,
neg_time_limit=2 * 60,
neg_step_time_limit=60,
negotiation_speed=21,
# simulation parameters
signing_delay=0,
force_signing=False,
batch_signing=True,
name: str = None,
# debugging parameters
agent_name_reveals_position: bool = True,
agent_name_reveals_type: bool = True,
**kwargs,
):
if exogenous_horizon is None:
exogenous_horizon = n_steps
self.exogenous_horizon = exogenous_horizon
self.buy_missing_products = buy_missing_products
self.production_buy_missing = production_buy_missing
self.exogenous_buy_missing = exogenous_buy_missing
kwargs["log_to_file"] = not no_logs
if compact:
kwargs["log_screen_level"] = logging.CRITICAL
kwargs["log_file_level"] = logging.ERROR
kwargs["log_negotiations"] = False
kwargs["log_ufuns"] = False
# kwargs["save_mechanism_state_in_contract"] = False
kwargs["save_cancelled_contracts"] = False
kwargs["save_resolved_breaches"] = False
kwargs["save_negotiations"] = False
self.compact = compact
if negotiation_speed == 0:
negotiation_speed = neg_n_steps + 1
super().__init__(
bulletin_board=None,
breach_processing=BreachProcessing.NONE,
awi_type="scml.scml2020.AWI",
mechanisms={"negmas.sao.SAOMechanism": {}},
default_signing_delay=signing_delay,
n_steps=n_steps,
time_limit=time_limit,
negotiation_speed=negotiation_speed,
neg_n_steps=neg_n_steps,
neg_time_limit=neg_time_limit,
neg_step_time_limit=neg_step_time_limit,
force_signing=force_signing,
batch_signing=batch_signing,
name=name,
**kwargs,
)
self.bulletin_board.record(
"settings", buy_missing_products, "buy_missing_products"
)
self.bulletin_board.record("settings", borrow_on_breach, "borrow_on_breach")
self.bulletin_board.record("settings", bankruptcy_limit, "bankruptcy_limit")
self.bulletin_board.record("settings", breach_penalty, "breach_penalty")
self.bulletin_board.record(
"settings", financial_report_period, "financial_report_period"
)
self.bulletin_board.record("settings", interest_rate, "interest_rate")
self.bulletin_board.record("settings", exogenous_horizon, "exogenous_horizon")
self.bulletin_board.record(
"settings", compensation_fraction, "compensation_fraction"
)
self.bulletin_board.record(
"settings", compensate_immediately, "compensate_immediately"
)
self.bulletin_board.record(
"settings", compensate_before_past_debt, "compensate_before_past_debt"
)
self.bulletin_board.record("settings", exogenous_force_max, "exogenous_force_max")
self.bulletin_board.record(
"settings", exogenous_buy_missing, "exogenous_buy_missing"
)
self.bulletin_board.record("settings", exogenous_no_borrow, "exogenous_no_borrow")
self.bulletin_board.record(
"settings", exogenous_no_bankruptcy, "exogenous_no_bankruptcy"
)
self.bulletin_board.record("settings", exogenous_penalty, "exogenous_penalty")
self.bulletin_board.record("settings", production_confirm, "production_confirm")
self.bulletin_board.record(
"settings", production_buy_missing, "production_buy_missing"
)
self.bulletin_board.record(
"settings", production_no_borrow, "production_no_borrow"
)
self.bulletin_board.record(
"settings", production_no_bankruptcy, "production_no_bankruptcy"
)
self.bulletin_board.record("settings", production_penalty, "production_penalty")
self.bulletin_board.record("settings", len(exogenous_contracts) > 0, "has_exogenous_contracts")
if self.info is None:
self.info = {}
self.info.update(
process_inputs=process_inputs,
process_outputs=process_outputs,
catalog_prices=catalog_prices,
agent_types_final=[get_full_type_name(_) for _ in agent_types],
agent_params_final=agent_params,
initial_balance_final=initial_balance,
buy_missing_products=buy_missing_products,
production_buy_missing=production_buy_missing,
exogenous_buy_missing=exogenous_buy_missing,
borrow_on_breach=borrow_on_breach,
bankruptcy_limit=bankruptcy_limit,
breach_penalty=breach_penalty,
financial_report_period=financial_report_period,
interest_rate=interest_rate,
compensation_fraction=compensation_fraction,
compensate_immediately=compensate_immediately,
compensate_before_past_debt=compensate_before_past_debt,
exogenous_force_max=exogenous_force_max,
exogenous_no_borrow=exogenous_no_borrow,
exogenous_no_bankruptcy=exogenous_no_bankruptcy,
exogenous_penalty=exogenous_penalty,
exogenous_supply_limit=exogenous_supply_limit,
exogenous_sales_limit=exogenous_sales_limit,
production_no_borrow=production_no_borrow,
production_no_bankruptcy=production_no_bankruptcy,
production_penalty=production_penalty,
compact=compact,
no_logs=no_logs,
n_steps=n_steps,
time_limit=time_limit,
neg_n_steps=neg_n_steps,
neg_time_limit=neg_time_limit,
neg_step_time_limit=neg_step_time_limit,
negotiation_speed=negotiation_speed,
signing_delay=signing_delay,
agent_name_reveals_position=agent_name_reveals_position,
agent_name_reveals_type=agent_name_reveals_type,
)
TimeInAgreementMixin.init(self, time_field="time")
self.breach_penalty = breach_penalty
self.bulletin_board.add_section("reports_time")
self.bulletin_board.add_section("reports_agent")
self.exogenous_no_borrow = exogenous_no_borrow
self.exogenous_no_bankruptcy = exogenous_no_bankruptcy
self.exogenous_penalty = exogenous_penalty
self.production_no_borrow = production_no_borrow
self.production_no_bankruptcy = production_no_bankruptcy
self.production_penalty = production_penalty
self.compensation_fraction = compensation_fraction
if not isinstance(agent_types, Iterable):
agent_types = [agent_types] * len(profiles)
assert len(profiles) == len(agent_types)
self.profiles = profiles
self.catalog_prices = catalog_prices
self.process_inputs = process_inputs
self.process_outputs = process_outputs
self.n_products = len(catalog_prices)
self.n_processes = len(process_inputs)
self.borrow_on_breach = borrow_on_breach
self.interest_rate = interest_rate
self.exogenous_force_max = exogenous_force_max
self.compensate_before_past_debt = compensate_before_past_debt
self.confirm_production = production_confirm
self.financial_reports_period = (
financial_report_period
if financial_report_period >= 1
else int(0.5 + financial_report_period * n_steps)
)
self.compensation_fraction = compensation_fraction
self.compensate_immediately = compensate_immediately
initial_balance = make_array(initial_balance, len(profiles), dtype=int)
agent_types = [get_class(_) for _ in agent_types]
self.bankruptcy_limit = (
-bankruptcy_limit
if isinstance(bankruptcy_limit, int)
else -int(0.5 + bankruptcy_limit * initial_balance.mean())
)
assert self.n_products == self.n_processes + 1
if exogenous_supply_limit is None:
self.supply_limit = sys.maxsize * np.ones(
(n_steps, self.n_products), dtype=int
)
else:
self.supply_limit = exogenous_supply_limit
if exogenous_sales_limit is None:
self.sales_limit = sys.maxsize * np.ones(
(n_steps, self.n_products), dtype=int
)
else:
self.sales_limit = exogenous_sales_limit
n_agents = len(profiles)
if agent_name_reveals_position or agent_name_reveals_type:
default_names = [f"{_:02}" for _ in range(n_agents)]
else:
default_names = [unique_name("", add_time=False) for _ in range(n_agents)]
if agent_name_reveals_type:
for i, at in enumerate(agent_types):
default_names[i] += f"{at.__name__[:3]}"
agent_levels = [
int(np.nonzero(np.max(p.costs != INFINITE_COST, axis=0).flatten())[0])
for p in profiles
]
if agent_name_reveals_position:
for i, l in enumerate(agent_levels):
default_names[i] += f"@{l:01}"
if agent_params is None:
agent_params = [dict(name=name) for i, name in enumerate(default_names)]
elif isinstance(agent_params, dict):
a = copy.copy(agent_params)
agent_params = []
for i, name in enumerate(default_names):
b = copy.deepcopy(a)
b["name"] = name
agent_params.append(b)
elif len(agent_params) == 1:
a = copy.copy(agent_params[0])
agent_params = []
for i, _ in enumerate(default_names):
b = copy.deepcopy(a)
b["name"] = name
agent_params.append(b)
else:
if agent_name_reveals_type or agent_name_reveals_position:
for i, (ns, ps) in enumerate(zip(default_names, agent_params)):
agent_params[i] = dict(**ps)
agent_params[i]["name"] = ns
n_processes = len(process_inputs)
n_products = n_processes + 1
agent_types.append(_SystemAgent)
agent_params.append({})
initial_balance = initial_balance.tolist() + [sys.maxsize // 4]
profiles.append(FactoryProfile(INFINITE_COST * np.ones(n_processes, dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
))
agents = []
for i, (atype, aparams) in enumerate(zip(agent_types, agent_params)):
a = instantiate(atype, **aparams)
self.join(a, i)
agents.append(a)
self.agent_types = [
get_class(_)
._type_name()
.replace("_agent", "")
.replace("_factory", "")
.replace("_manager", "")
for _ in agent_types
]
self.agent_params = [
{k: v for k, v in _.items() if k != "name"} for _ in agent_params
]
self.agent_unique_types = [
f"{t}{hash(p)}" if len(p) > 0 else t
for t, p in zip(self.agent_types, self.agent_params)
]
self.factories = [
Factory(
world=self,
profile=profile,
initial_balance=initial_balance[i],
inputs=process_inputs,
outputs=process_outputs,
agent_id=agents[i].id,
catalog_prices=catalog_prices,
compensate_before_past_debt=self.compensate_before_past_debt,
buy_missing_products=self.buy_missing_products,
exogenous_buy_missing=self.exogenous_buy_missing,
production_buy_missing=self.production_buy_missing,
exogenous_penalty=self.exogenous_penalty,
exogenous_no_bankruptcy=self.exogenous_no_bankruptcy,
exogenous_no_borrow=self.exogenous_no_borrow,
production_penalty=self.production_penalty,
production_no_borrow=self.production_no_borrow,
production_no_bankruptcy=self.production_no_bankruptcy,
confirm_production=self.confirm_production,
initial_inventory=None if i < len(profiles) - 1 else sys.maxsize // 4 * np.ones(n_products, dtype=int)
)
for i, profile in enumerate(profiles)
]
self.a2f = dict(zip((_.id for _ in agents), self.factories))
self.afp = list(zip(agents, self.factories, profiles))
self.f2i = self.a2i = dict(zip((_.id for _ in agents), range(n_agents)))
self.i2a = agents
self.i2f = self.factories
self.breach_prob = dict(zip((_.id for _ in agents), itertools.repeat(0.0)))
self._breach_level = dict(zip((_.id for _ in agents), itertools.repeat(0.0)))
self.agent_n_contracts = dict(zip((_.id for _ in agents), itertools.repeat(0)))
self.suppliers: List[List[str]] = [[] for _ in range(n_products)]
self.consumers: List[List[str]] = [[] for _ in range(n_products)]
self.agent_processes: Dict[str, List[int]] = defaultdict(list)
self.agent_inputs: Dict[str, List[int]] = defaultdict(list)
self.agent_outputs: Dict[str, List[int]] = defaultdict(list)
self.agent_consumers: Dict[str, List[str]] = defaultdict(list)
self.agent_suppliers: Dict[str, List[str]] = defaultdict(list)
for p in range(n_processes):
for agent_id, profile in zip(self.agents.keys(), profiles):
if agent_id == "SYSTEM":
continue
if np.all(profile.costs[:, p] == INFINITE_COST):
continue
self.suppliers[p + 1].append(agent_id)
self.consumers[p].append(agent_id)
self.agent_processes[agent_id].append(p)
self.agent_inputs[agent_id].append(p)
self.agent_outputs[agent_id].append(p + 1)
for p in range(n_products):
for a in self.suppliers[p]:
self.agent_consumers[a] = self.consumers[p]
for a in self.consumers[p]:
self.agent_suppliers[a] = self.suppliers[p]
self.agent_processes = {k: np.array(v) for k, v in self.agent_processes.items()}
self.agent_inputs = {k: np.array(v) for k, v in self.agent_inputs.items()}
self.agent_outputs = {k: np.array(v) for k, v in self.agent_outputs.items()}
self._n_production_failures = 0
self.__n_nullified = 0
self.__n_bankrupt = 0
self.penalties = 0
# self.is_bankrupt: Dict[str, bool] = dict(
# zip(self.agents.keys(), itertools.repeat(False))
# )
self.compensation_balance = 0
self.compensation_records: Dict[str, List[CompensationRecord]] = defaultdict(
list
)
self.exogenous_contracts: Dict[int: List[Contract]] = defaultdict(list)
for c in exogenous_contracts:
seller_id = agents[c.seller].id if c.seller >= 0 else "SYSTEM"
buyer_id = agents[c.buyer].id if c.buyer >= 0 else "SYSTEM"
contract = Contract(
agreement={
"time": c.time,
"quantity": c.quantity,
"unit_price": c.unit_price,
},
partners = [buyer_id, seller_id],
issues=[],
signatures=[],
signed_at=-1,
to_be_signed_at=c.revelation_time,
annotation={
"seller": seller_id,
"buyer": buyer_id,
"caller": "SYSTEM",
"is_buy": random.random() > 0.5,
"product": c.product,
},
)
self.exogenous_contracts[c.revelation_time].append(contract)
self.compensation_factory = Factory(
FactoryProfile(
np.zeros((n_steps, n_processes), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
np.zeros((n_steps, n_products), dtype=int),
),
initial_balance=0,
inputs=self.process_inputs,
outputs=self.process_outputs,
world=self,
agent_id="COMPENSATION",
agent_name="COMPENSATION",
catalog_prices=catalog_prices,
compensate_before_past_debt=True,
buy_missing_products=False,
production_buy_missing=False,
exogenous_buy_missing=False,
exogenous_penalty=0.0,
exogenous_no_bankruptcy=True,
exogenous_no_borrow=True,
production_penalty=0.0,
production_no_borrow=True,
production_no_bankruptcy=True,
confirm_production=False,
)
[docs] @classmethod
def generate(
cls,
agent_types: List[Type[SCML2020Agent]],
agent_params: List[Dict[str, Any]] = None,
n_steps: Union[Tuple[int, int], int] = 100,
n_processes: Union[Tuple[int, int], int] = 4,
n_lines: Union[np.ndarray, Tuple[int, int], int] = 10,
n_agents_per_process: Union[np.ndarray, Tuple[int, int], int] = 3,
process_inputs: Union[np.ndarray, Tuple[int, int], int] = 1,
process_outputs: Union[np.ndarray, Tuple[int, int], int] = 1,
production_costs: Union[np.ndarray, Tuple[int, int], int] = (1, 10),
profit_means: Union[np.ndarray, Tuple[float, float], float] = (0.1, 0.2),
profit_stddevs: Union[np.ndarray, Tuple[float, float], float] = 0.05,
max_productivity: Union[np.ndarray, Tuple[float, float], float] = 1.0,
initial_balance: Optional[Union[np.ndarray, Tuple[int, int], int]] = None,
cost_increases_with_level=True,
horizon: Union[Tuple[float, float], float] = (0.5, 0.9),
equal_exogenous_supply=False,
equal_exogenous_sales=False,
use_exogenous_contracts=True,
exogenous_control: Union[Tuple[float, float], float] = (0.0, 1.0),
cash_availability: Union[Tuple[float, float], float] = 2.5,
profit_basis=np.mean,
force_signing=False,
**kwargs,
) -> Dict[str, Any]:
"""
Generates the configuration for a world
Args:
agent_types: All agent types
agent_params: Agent parameters used to initialize them
n_steps: Number of simulation steps
n_processes: Number of processes in the production chain
n_lines: Number of lines per factory
process_inputs: Number of input units per process
process_outputs: Number of output units per process
production_costs: Production cost per factory
profit_means: Mean profitability per production level (i.e. process).
profit_stddevs: Std. Dev. of the profitability of every level (i.e. process).
max_productivity: Maximum possible productivity per level (i.e. process).
initial_balance: The initial balance of all agents
n_agents_per_process: Number of agents per process
cost_increases_with_level: If true, production cost will be higher for processes nearer to the final
product.
profit_basis: The statistic used when controlling catalog prices by profit arguments. It can be np.mean,
np.median, np.min, np.max or any Callable[[list[float]], float] and is used to summarize
production costs at every level.
horizon: The horizon used for revealing external supply/sales as a fraction of n_steps
equal_exogenous_supply: If true, external supply will be distributed equally among all agents in the first
layer
equal_exogenous_sales: If true, external sales will be distributed equally among all agents in the last
layer
use_exogenous_contracts: If true, external supply and sales are revealed as exogenous contracts instead
of using exogenous_* parts of the factory profile
cash_availability: The fraction of the total money needs of the agent to work at maximum capacity that
is available as `initial_balance` . This is only effective if `initial_balance` is set
to `None` .
force_signing: Whether to force contract signatures (exogenous contracts are treated in the same way).
exogenous_control: How much control does the agent have over exogenous contract signing. Only effective if
force_signing is False and use_exogenous_contracts is True
**kwargs:
Returns:
world configuration as a Dict[str, Any]. A world can be generated from this dict by calling SCML2020World(**d)
Remarks:
- Most parameters (i.e. `process_inputs` , `process_outputs` , `n_agents_per_process` , `costs` ) can
take a single value, a tuple of two values, or a list of values.
If it has a single value, it is repeated for all processes/factories as appropriate. If it is a
tuple of two numbers $(i, j)$, each process will take a number sampled from a uniform distribution
supported on $[i, j]$ inclusive. If it is a list of values, of the length `n_processes` , it is used as
it is otherwise, it is used to sample values for each process.
"""
info = dict(
n_steps=n_steps,
n_processes=n_processes,
n_lines=n_lines,
force_signing=force_signing,
n_agents_per_process=n_agents_per_process,
process_inputs=process_inputs,
process_outputs=process_outputs,
production_costs=production_costs,
profit_means=profit_means,
profit_stddevs=profit_stddevs,
max_productivity=max_productivity,
initial_balance=initial_balance,
cost_increases_with_level=cost_increases_with_level,
equal_exogenous_sales=equal_exogenous_sales,
equal_exogenous_supply=equal_exogenous_supply,
cash_availability=cash_availability,
profit_basis="min"
if profit_basis == np.min
else "mean"
if profit_basis == np.mean
else "max"
if profit_basis == np.max
else "median"
if profit_basis == np.median
else "unknown",
)
n_processes = intin(n_processes)
n_steps = intin(n_steps)
exogenous_control = realin(exogenous_control)
np.errstate(divide="ignore")
n_startup = n_processes
horizon = max(1, min(n_steps, int(realin(horizon) * n_steps)))
process_inputs = make_array(process_inputs, n_processes, dtype=int)
process_outputs = make_array(process_outputs, n_processes, dtype=int)
n_agents_per_process = make_array(n_agents_per_process, n_processes, dtype=int)
profit_means = make_array(profit_means, n_processes, dtype=float)
profit_stddevs = make_array(profit_stddevs, n_processes, dtype=float)
max_productivity = make_array(
max_productivity, n_processes * n_steps, dtype=float
).reshape((n_processes, n_steps))
n_agents = n_agents_per_process.sum()
assert n_agents >= n_processes
n_products = n_processes + 1
production_costs = make_array(production_costs, n_agents, dtype=int)
if initial_balance is not None:
initial_balance = make_array(initial_balance, n_agents, dtype=int)
if not isinstance(agent_types, Iterable):
agent_types = [agent_types] * n_agents
if agent_params is None:
agent_params = dict()
if isinstance(agent_params, dict):
agent_params = [copy.copy(agent_params) for _ in range(n_agents)]
else:
assert len(agent_params) == 1
agent_params = [copy.copy(agent_params[0]) for _ in range(n_agents)]
elif len(agent_types) != n_agents:
if agent_params is None:
agent_params = [dict()] * len(agent_types)
if isinstance(agent_params, dict):
agent_params = [
copy.copy(agent_params) for _ in range(len(agent_types))
]
assert len(agent_types) == len(agent_params)
tp = random.choices(list(range(len(agent_types))), k=n_agents)
agent_types = [copy.copy(agent_types[_]) for _ in tp]
agent_params = [copy.copy(agent_params[_]) for _ in tp]
else:
if agent_params is None:
agent_params = [dict()] * len(agent_types)
if isinstance(agent_params, dict):
agent_params = [
copy.copy(agent_params) for _ in range(len(agent_types))
]
agent_types = list(agent_types)
agent_params = list(agent_params)
assert len(agent_types) == len(agent_params)
# generate production costs making sure that every agent can do exactly one process
n_agents_cumsum = n_agents_per_process.cumsum().tolist()
first_agent = [0] + n_agents_cumsum[:-1]
last_agent = n_agents_cumsum[:-1] + [n_agents]
costs = INFINITE_COST * np.ones((n_agents, n_lines, n_processes), dtype=int)
for p, (f, l) in enumerate(zip(first_agent, last_agent)):
costs[f:l, :, p] = production_costs[f:l].reshape((l - f), 1)
process_of_agent = np.empty(n_agents, dtype=int)
for i, (f, l) in enumerate(zip(first_agent, last_agent)):
process_of_agent[f: l] = i
if cost_increases_with_level:
production_costs[f: l] = np.round(production_costs[f: l] * math.sqrt(i+1)).astype(int)
# generate external contract amounts (controlled by productivity):
# - generate total amount of input to the market (it will end up being an n_products list of n_steps vectors)
quantities = [
np.round(n_lines * n_agents_per_process[0] * max_productivity[0, :]).astype(
int
)
]
# - make sure there is a cool-down period at the end in which no more input is added that cannot be converted
# into final products in time
quantities[0][-n_startup:] = 0
# - for each level, find the amount of the output product that can be produced given the input amount and
# productivity
for p in range(n_processes):
agents = n_agents_per_process[p]
lines = n_lines * agents
quantities.append(
np.minimum(
(quantities[-1] // process_outputs[p]) * process_inputs[p],
(
np.round(lines * max_productivity[p, :]).astype(int)
// process_inputs[p]
)
* process_outputs[p],
)
)
# * shift quantities one step to account for the one step needed to move the produce to the next level. This
# step results from having production happen after contract execution.
quantities[-1][1:] = quantities[-1][:-1]
quantities[-1][0] = 0
assert quantities[-1][-1] == 0 or p >= n_startup - 1
assert quantities[-1][0] == 0
assert np.sum(quantities[-1] == 0) >= n_startup
# - divide the quantity at every level between factories
if equal_exogenous_supply:
exogenous_supplies = np.maximum(
1, np.round(quantities[0] / n_agents_per_process[0]).astype(int)
).tolist()
exogenous_supplies = [
np.array([exogenous_supplies[p]] * n_agents_per_process[p])
for p in range(n_processes)
]
else:
exogenous_supplies = []
for s in range(n_steps):
exogenous_supplies.append(
integer_cut(quantities[0][s], n_agents_per_process[0], 0)
)
assert sum(exogenous_supplies[-1]) == quantities[0][s]
if equal_exogenous_sales:
exogenous_sales = np.maximum(
1, np.round(quantities[-1] / n_agents_per_process[-1]).astype(int)
).tolist()
exogenous_sales = [
np.array([exogenous_sales[p]] * n_agents_per_process[p])
for p in range(n_processes)
]
else:
exogenous_sales = []
for s in range(n_steps):
exogenous_sales.append(
integer_cut(quantities[-1][s], n_agents_per_process[-1], 0)
)
assert sum(exogenous_sales[-1]) == quantities[-1][s]
# - now exogenous_supplies and exogenous_sales are both n_steps lists of n_agents_per_process[p] vectors (jagged)
# assign prices to the quantities given the profits
catalog_prices = np.zeros(n_products, dtype=int)
catalog_prices[0] = 10
supply_prices = np.zeros((n_agents_per_process[0], n_steps), dtype=int)
supply_prices[:, :] = catalog_prices[0]
sale_prices = np.zeros((n_agents_per_process[-1], n_steps), dtype=int)
manufacturing_costs = np.zeros((n_processes, n_steps), dtype=int)
for p in range(n_processes):
manufacturing_costs[p, :] = profit_basis(
costs[first_agent[p] : last_agent[p], :, p]
)
manufacturing_costs[p, :p] = 0
manufacturing_costs[p, p - n_startup :] = 0
profits = np.zeros((n_processes, n_steps))
for p in range(n_processes):
profits[p, :] = np.random.randn() * profit_stddevs[p] + profit_means[p]
input_costs = np.zeros((n_processes, n_steps), dtype=int)
for step in range(n_steps):
input_costs[0, step] = np.sum(
exogenous_supplies[step] * supply_prices[:, step][:]
)
input_quantity = np.zeros((n_processes, n_steps), dtype=int)
input_quantity[0, :] = quantities[0]
active_lines = np.hstack(
[(n_lines * n_agents_per_process).reshape((n_processes, 1))] * n_steps
)
assert active_lines.shape == (n_processes, n_steps)
active_lines[0, :] = input_quantity[0, :] // process_inputs[0]
output_quantity = np.zeros((n_processes, n_steps), dtype=int)
output_quantity[0, :] = active_lines[0, :] * process_outputs[0]
manufacturing_costs[0, :-n_startup] *= active_lines[0, :-n_startup]
total_costs = input_costs + manufacturing_costs
output_total_prices = np.ceil(total_costs * (1 + profits)).astype(int)
for p in range(1, n_processes):
input_costs[p, p:] = output_total_prices[p - 1, p - 1 : -1]
input_quantity[p, p:] = output_quantity[p - 1, p - 1 : -1]
active_lines[p, :] = input_quantity[p, :] // process_inputs[p]
output_quantity[p, :] = active_lines[p, :] * process_outputs[p]
manufacturing_costs[p, p : p - n_startup] *= active_lines[
p, p : p - n_startup
]
total_costs[p, :] = input_costs[p, :] + manufacturing_costs[p, :]
output_total_prices[p, :] = np.ceil(
total_costs[p, :] * (1 + profits[p, :])
).astype(int)
sale_prices[:, n_startup:] = np.ceil(
output_total_prices[-1, n_startup - 1 : -1]
/ output_quantity[-1, n_startup - 1 : -1]
).astype(int)
product_prices = np.zeros((n_products, n_steps))
product_prices[0, :-n_startup] = catalog_prices[0]
product_prices[1:, 1:] = np.ceil(
np.divide(
output_total_prices.astype(float),
output_quantity.astype(float),
out=np.zeros_like(output_total_prices, dtype=float),
where=output_quantity != 0,
)
).astype(int)[:, :-1]
catalog_prices = np.ceil(
[
profit_basis(product_prices[p, p : p + n_steps - n_startup])
for p in range(n_products)
]
).astype(int)
profiles = []
nxt = 0
for l in range(n_processes):
for a in range(n_agents_per_process[l]):
esales = np.zeros((n_steps, n_products), dtype=int)
esupplies = np.zeros((n_steps, n_products), dtype=int)
esale_prices = np.zeros((n_steps, n_products), dtype=int)
esupply_prices = np.zeros((n_steps, n_products), dtype=int)
if l == 0:
esupplies[:, 0] = [exogenous_supplies[s][a] for s in range(n_steps)]
esupply_prices[:, 0] = supply_prices[a, :]
if l == n_processes - 1:
esales[:, -1] = [exogenous_sales[s][a] for s in range(n_steps)]
esale_prices[:, -1] = sale_prices[a, :]
profiles.append(
FactoryProfile(
costs=costs[nxt],
exogenous_sales=esales,
exogenous_supplies=esupplies,
exogenous_sale_prices=esale_prices,
exogenous_supply_prices=esupply_prices,
)
)
nxt += 1
max_income = (
output_quantity * catalog_prices[1:].reshape((n_processes, 1)) - total_costs
)
assert nxt == n_agents
if initial_balance is None:
# every agent at every level will have just enough to do all the needed to do cash_availability fraction of
# production (even though it may not have enough lines to do so)
cash_availability = _realin(cash_availability)
balance = np.ceil(
np.sum(total_costs, axis=1) # / n_agents_per_process
).astype(int)
initial_balance = []
for b, a in zip(balance, n_agents_per_process):
initial_balance += [int(math.ceil(b * cash_availability))] * a
b = np.sum(initial_balance)
info.update(
dict(
product_prices=product_prices,
active_lines=active_lines,
input_quantities=input_quantity,
output_quantities=output_quantity,
expected_productivity=float(np.sum(active_lines))
/ np.sum(n_lines * n_steps * n_agents_per_process),
expected_n_products=np.sum(active_lines, axis=-1),
expected_income=max_income,
expected_welfare=float(np.sum(max_income)),
expected_income_per_step=max_income.sum(axis=0),
expected_income_per_process=max_income.sum(axis=-1),
expected_mean_profit=float(np.sum(max_income) / b)
if b != 0
else np.sum(max_income),
expected_profit_sum=float(n_agents * np.sum(max_income) / b)
if b != 0
else n_agents * np.sum(max_income),
)
)
exogenous = []
if use_exogenous_contracts:
for indx, profile in enumerate(profiles):
input_product = process_of_agent[indx]
for step, (sale, price) in enumerate(
zip(profile.exogenous_sales[input_product + 1, :], profile.exogenous_sale_prices[ input_product +1, :])
):
if sale == 0:
continue
if force_signing or exogenous_control <= 0.0:
exogenous.append(
ExogenousContract(
product=input_product + 1,
quantity=sale,
unit_price=price,
time=step,
revelation_time=max(0, step - horizon),
seller=indx,
buyer=-1,
)
)
else:
n_contracts = int(1 + exogenous_control * (sale - 1))
per_contract = integer_cut(sale, n_contracts, 0)
for q in per_contract:
if q == 0:
continue
exogenous.append(
ExogenousContract(
product=input_product + 1,
quantity=q,
unit_price=price,
time=step,
revelation_time=max(0, step - horizon),
seller=indx,
buyer=-1,
)
)
for step, (supply, price) in enumerate(
zip(profile.exogenous_supplies[input_product, :], profile.exogenous_supply_prices[input_product, :])
):
if supply == 0:
continue
if force_signing or exogenous_control <= 0.0:
exogenous.append(
ExogenousContract(
product=input_product,
quantity=supply,
unit_price=price,
time=step,
revelation_time=max(0, step - horizon),
seller=-1,
buyer=indx,
)
)
else:
n_contracts = int(1 + exogenous_control * (supply - 1))
per_contract = integer_cut(supply, n_contracts, 0)
for q in per_contract:
if q == 0:
continue
exogenous.append(
ExogenousContract(
product=input_product,
quantity=q,
unit_price=price,
time=step,
revelation_time=max(0, step - horizon),
seller=-1,
buyer=indx,
)
)
profile.exogenous_sales = np.zeros_like(profile.exogenous_sales)
profile.exogenous_supplies = np.zeros_like(profile.exogenous_supplies)
profile.exogenous_sale_prices = np.zeros_like(
profile.exogenous_sale_prices
)
profile.exogenous_supply_prices = np.zeros_like(
profile.exogenous_supply_prices
)
return dict(
process_inputs=process_inputs,
process_outputs=process_outputs,
catalog_prices=catalog_prices,
profiles=profiles,
exogenous_contracts=exogenous,
agent_types=agent_types,
agent_params=agent_params,
initial_balance=initial_balance,
n_steps=n_steps,
info=info,
force_signing=force_signing,
**kwargs,
)
[docs] def get_private_state(self, agent: "SCML2020Agent") -> dict:
return vars(self.a2f[agent.id].state)
[docs] def add_financial_report(
self, agent: SCML2020Agent, factory: Factory, reports_agent, reports_time
) -> None:
"""
Records a financial report for the given agent in the agent indexed reports and time indexed reports
Args:
agent: The agent
factory: Its factory
reports_agent: A dictionary of financial reports indexed by agent id
reports_time: A dictionary of financial reports indexed by time
Returns:
"""
bankrupt = factory.is_bankrupt
inventory = (
int(np.sum(self.catalog_prices * factory.current_inventory))
if not bankrupt
else 0
)
report = FinancialReport(
agent_id=agent.id,
step=self.current_step,
cash=factory.current_balance,
assets=inventory,
breach_prob=self.breach_prob[agent.id],
breach_level=self._breach_level[agent.id],
is_bankrupt=bankrupt,
agent_name=agent.name,
)
repstr = str(report).replace("\n", " ")
self.logdebug(f"{agent.name}: {repstr}")
if reports_agent.get(agent.id, None) is None:
reports_agent[agent.id] = {}
reports_agent[agent.id][self.current_step] = report
if reports_time.get(self.current_step, None) is None:
reports_time[self.current_step] = {}
reports_time[self.current_step][agent.id] = report
[docs] def simulation_step_before_execution(self):
s = self.current_step
# register exogenous contracts as concluded
# -----------------------------------------
for contract in self.exogenous_contracts[s]:
self.on_contract_concluded(contract, to_be_signed_at=contract.to_be_signed_at)
# pay interests for negative balances
# -----------------------------------
if self.interest_rate > 0.0:
for agent, factory, _ in self.afp:
if factory.current_balance < 0 and not factory.is_bankrupt:
to_pay = -int(
math.ceil(self.interest_rate * factory.current_balance)
)
factory.pay(to_pay)
[docs] def simulation_step_after_execution(self):
s = self.current_step
# publish financial reports
# -------------------------
if self.current_step % self.financial_reports_period == 0:
reports_agent = self.bulletin_board.data["reports_agent"]
reports_time = self.bulletin_board.data["reports_time"]
for agent, factory, _ in self.afp:
if agent.id == "SYSTEM":
continue
self.add_financial_report(agent, factory, reports_agent, reports_time)
# do external transactions and step factories
# -------------------------------------------
if self.exogenous_force_max:
for a, f, p in self.afp:
if f.is_bankrupt:
continue
f.step(p.exogenous_sales[s, :], p.exogenous_supplies[s, :])
else:
afp_randomized = [
self.afp[_] for _ in np.random.permutation(np.arange(len(self.afp)))
]
for a, f, p in afp_randomized:
if f.is_bankrupt:
continue
supply = a.confirm_exogenous_supplies(
p.exogenous_supplies[s].copy(), p.exogenous_supply_prices[s].copy()
)
sales = a.confirm_exogenous_sales(
p.exogenous_sales[s].copy(), p.exogenous_sale_prices[s].copy()
)
f.step(sales, supply)
# remove contracts saved in factories for this step
for factory in self.factories:
factory.contracts[self.current_step] = []
[docs] def contract_size(self, contract: Contract) -> float:
return contract.agreement["quantity"] * contract.agreement["unit_price"]
[docs] def contract_record(self, contract: Contract) -> Dict[str, Any]:
c = {
"id": contract.id,
"seller_name": self.agents[contract.annotation["seller"]].name,
"buyer_name": self.agents[contract.annotation["buyer"]].name,
"seller_type": self.agents[
contract.annotation["seller"]
].__class__.__name__,
"buyer_type": self.agents[contract.annotation["buyer"]].__class__.__name__,
"delivery_time": contract.agreement["time"],
"quantity": contract.agreement["quantity"],
"unit_price": contract.agreement["unit_price"],
"signed_at": contract.signed_at,
"nullified_at": contract.nullified_at,
"concluded_at": contract.concluded_at,
"signatures": "|".join(str(_) for _ in contract.signatures),
"issues": contract.issues if not self.compact else None,
"seller": contract.annotation["seller"],
"buyer": contract.annotation["buyer"],
"product_name": "p" + str(contract.annotation["product"]),
}
if not self.compact:
c.update(contract.annotation)
c["n_neg_steps"] = contract.mechanism_state.step if contract.mechanism_state else 0
return c
[docs] def breach_record(self, breach: Breach) -> Dict[str, Any]:
return {
"perpetrator": breach.perpetrator,
"perpetrator_name": breach.perpetrator,
"level": breach.level,
"type": breach.type,
"time": breach.step,
}
[docs] def execute_action(
self, action: Action, agent: "SCML2020Agent", callback: Callable = None
) -> bool:
if action.type == "schedule":
s, _ = self.a2f[agent.id].schedule_production(
process=action.params["process"],
step=action.params.get("step", (self.current_step, self.n_steps - 1)),
line=action.params.get("line", -1),
override=action.params.get("override", True),
method=action.params.get("method", "latest"),
)
return s >= 0
elif action.type == "cancel":
return self.a2f[agent.id].cancel_production(
step=action.params.get("step", -1), line=action.params.get("line", -1)
)
[docs] def post_step_stats(self):
self._stats["n_contracts_nullified_now"].append(self.__n_nullified)
self._stats["n_bankrupt"].append(self.__n_bankrupt)
market_size = 0
self._stats[f"_balance_society"].append(self.penalties)
internal_market_size = self.penalties
prod = []
for a, f, _ in self.afp:
if a.id == "SYSTEM":
continue
self._stats[f"balance_{a.name}"].append(f.current_balance)
for p in a.awi.my_input_products:
self._stats[f"inventory_{a.name}_input_{p}"].append(
f.current_inventory[p]
)
for p in a.awi.my_output_products:
self._stats[f"inventory_{a.name}_output_{p}"].append(
f.current_inventory[p]
)
prod.append(
np.sum(f.commands[self.current_step, :] != NO_COMMAND)
/ f.profile.n_lines
)
self._stats[f"productivity_{a.name}"].append(prod[-1])
self._stats[f"assets_{a.name}"].append(
np.sum(f.current_inventory * self.catalog_prices)
)
self._stats[f"bankrupt_{a.name}"].append(f.is_bankrupt)
if not f.is_bankrupt:
market_size += f.current_balance
self._stats["productivity"].append(float(np.mean(prod)))
self._stats["market_size"].append(market_size)
self._stats["production_failures"].append(
self._n_production_failures / len(self.factories)
if len(self.factories) > 0
else np.nan
)
self._stats["_market_size_total"].append(market_size + internal_market_size)
self._stats["bankruptcy"] = np.sum(self.stats["n_bankrupt"]) / len(self.agents)
# self._stats["business"] = np.sum(self.stats["business_level"])
[docs] def pre_step_stats(self):
self._n_production_failures = 0
self.__n_nullified = 0
self.__n_bankrupt = 0
@property
def productivity(self) -> float:
"""Fraction of production lines occupied during the simulation"""
return np.mean(self.stats["productivity"])
[docs] def welfare(self, include_bankrupt: bool = False) -> float:
"""Total welfare of all agents"""
return sum(
f.current_balance - f.initial_balance
for f in self.factories
if include_bankrupt or not f.is_bankrupt
)
[docs] def relative_welfare(self, include_bankrupt: bool = False) -> Optional[float]:
"""Total welfare relative to expected value. Returns None if no expectation is found in self.info"""
if "expected_income" not in self.info.keys():
return None
return self.welfare(include_bankrupt) / np.sum(self.info["expected_income"])
@property
def relative_productivity(self) -> Optional[float]:
"""Productivity relative to the expected value. Will return None if self.info does not have
the expected productivity"""
if "expected_productivity" not in self.info.keys():
return None
return self.productivity / self.info["expected_productivity"]
@property
def bankruptcy_rate(self) -> float:
"""The fraction of factories that went bankrupt"""
return sum([f.is_bankrupt for f in self.factories]) / len(self.factories)
@property
def num_bankrupt(self) -> float:
"""The fraction of factories that went bankrupt"""
return sum([f.is_bankrupt for f in self.factories])
[docs] def order_contracts_for_execution(
self, contracts: Collection[Contract]
) -> Collection[Contract]:
return sorted(contracts, key=lambda x: x.annotation["product"])
def _execute(
self,
product: int,
q: int,
p: int,
u: int,
buyer_factory: Factory,
seller_factory: Factory,
has_breaches: bool,
):
"""Executes the contract"""
self.logdebug(
f"Transferring {q} of {product} at price {u} ({'breached' if has_breaches else ''})"
)
if q == 0 or u == 0:
self.logwarning(
f"{buyer_factory.agent_name} bought {q} from {seller_factory.agent_name} at {u} dollars"
f" ({'with breaches' if has_breaches else 'no breaches'})!! Zero quantity or unit price"
)
if has_breaches:
money = (
p
if buyer_factory.current_balance - p > self.bankruptcy_limit
else max(0, buyer_factory.current_balance - self.bankruptcy_limit)
)
quantity = min(seller_factory.current_inventory[product], q)
if quantity == 0 or money == 0:
return
u, q = min(money // quantity, u), min(quantity, money // u)
assert q >= 0, f"executing with quantity {q}"
if q != 0:
buyer_factory.buy(product, q, u, False, 0.0)
seller_factory.buy(product, -q, u, False, 0.0)
def __register_contract(self, agent_id: str, level: float) -> None:
"""Registers execution of the contract in the agent's stats"""
n_contracts = self.agent_n_contracts[agent_id] - 1
self.breach_prob[agent_id] = (
self.breach_prob[agent_id] * n_contracts + (level > 0)
) / (n_contracts + 1)
self._breach_level[agent_id] = (
self.breach_prob[agent_id] * n_contracts + level
) / (n_contracts + 1)
[docs] def record_bankrupt(self, factory: Factory) -> None:
"""Records agent bankruptcy"""
agent_id = factory.agent_id
# announce bankruptcy
reports_agent = self.bulletin_board.data["reports_agent"]
reports_time = self.bulletin_board.data["reports_time"]
self.add_financial_report(
self.agents[agent_id], factory, reports_agent, reports_time
)
self.__n_bankrupt += 1
[docs] def on_contract_concluded(self, contract: Contract, to_be_signed_at: int) -> None:
if (
any(self.a2f[_].is_bankrupt for _ in contract.partners)
or contract.agreement["time"] >= self.n_steps
):
return
super().on_contract_concluded(contract, to_be_signed_at)
[docs] def on_contract_signed(self, contract: Contract):
# we need to cancel this contract if a partner was bankrupt (that is necessary only for
# force_signing case as in this case the two partners will be assued to sign no matter what is
# their bankruptcy status
if (
any(self.a2f[_].is_bankrupt for _ in contract.partners)
or contract.agreement["time"] >= self.n_steps
):
self.ignore_contract(contract)
return
super().on_contract_signed(contract)
self.logdebug(f"SIGNED {str(contract)}")
t = contract.agreement["time"]
u, q = contract.agreement["unit_price"], contract.agreement["quantity"]
product = contract.annotation["product"]
agent, partner = contract.partners
is_seller = agent == contract.annotation["seller"]
self.a2f[agent].contracts[t].append(
ContractInfo(q, u, product, is_seller, partner, contract)
)
self.a2f[partner].contracts[t].append(
ContractInfo(q, u, product, not is_seller, agent, contract)
)
[docs] def nullify_contract(self, contract: Contract, new_quantity: int):
self.__n_nullified += 1
contract.nullified_at = self.current_step
contract.annotation["new_quantity"] = new_quantity
def __register_breach(
self, agent_id: str, level: float, contract_total: float, factory: Factory
) -> int:
"""
Registers a breach of the given level on the given agent. Assume that the contract is already added
to the agent_contracts
Args:
agent_id: The perpetrator of the breach
level: The breach level
contract_total: The total of the contract breached (quantity * unit_price)
factory: The factory corresponding to the perpetrator
Returns:
If nonzero, the agent should go bankrupt and this amount taken from them
"""
self.logdebug(
f"{self.agents[agent_id].name} breached {level} of {contract_total}"
)
if factory.is_bankrupt:
return 0
if level <= 0:
return 0
penalty = int(math.ceil(level * contract_total))
if factory.current_balance - penalty < self.bankruptcy_limit:
return penalty
if penalty > 0:
factory.pay(penalty)
self.penalties += penalty
return 0
[docs] def start_contract_execution(self, contract: Contract) -> Optional[Set[Breach]]:
self.logdebug(f"Executing {str(contract)}")
# get contract info
breaches = set()
if self.compensate_immediately and (
contract.nullified_at >= 0
or any(self.a2f[a].is_bankrupt for a in contract.partners)
):
return None
product = contract.annotation["product"]
buyer_id, seller_id = (
contract.annotation["buyer"],
contract.annotation["seller"],
)
buyer, buyer_factory = self.agents[buyer_id], self.a2f[buyer_id]
seller, seller_factory = self.agents[seller_id], self.a2f[seller_id]
q, u, t = (
contract.agreement["quantity"],
contract.agreement["unit_price"],
contract.agreement["time"],
)
if q <= 0 or u <= 0:
self.logwarning(
f"Contract {str(contract)} has zero quantity of unit price!!! will be ignored"
)
return breaches
# if the contract is already nullified, take care of it
if contract.nullified_at >= 0:
self.compensation_factory._inventory[product] = 0
self.compensation_factory._balance = 0
for c in self.compensation_records.get(contract.id, []):
q = min(q, c.quantity)
if c.product >= 0 and c.quantity > 0:
assert c.product == product
self.compensation_factory._inventory[product] += c.quantity
self.compensation_factory._balance += c.money
if c.seller_bankrupt:
seller_factory = self.compensation_factory
else:
buyer_factory = self.compensation_factory
elif seller_factory == buyer_factory:
# means that both the seller and buyer are bankrupt
return None
p = q * u
assert t == self.current_step
self.agent_n_contracts[buyer_id] += 1
self.agent_n_contracts[seller_id] += 1
missing_product = q - seller_factory.current_inventory[product]
missing_money = p - buyer_factory.current_balance
# if there are no breaches, just execute the contract
if missing_money <= 0 and missing_product <= 0:
self._execute(
product, q, p, u, buyer_factory, seller_factory, has_breaches=False
)
self.__register_contract(seller_id, 0)
self.__register_contract(buyer_id, 0)
return breaches
# if there is a product breach (the seller does not have enough products), register it
if missing_product <= 0:
self.__register_contract(seller_id, 0)
else:
product_breach_level = missing_product / q
breaches.add(
Breach(
contract=contract,
perpetrator=seller_id,
victims=buyer_id,
level=product_breach_level,
type="product",
)
)
self.__register_contract(seller_id, product_breach_level)
self.__register_breach(seller_id, product_breach_level, p, seller_factory)
if self.borrow_on_breach and seller_factory != self.compensation_factory:
paid_for = seller_factory.store(
product,
-missing_product,
u,
self.buy_missing_products,
self.breach_penalty,
)
missing_product -= paid_for
# if there is a money breach (the buyer does not have enough money), register it
if missing_money < 0:
self.__register_contract(buyer_id, 0)
else:
money_breach_level = missing_money / p
breaches.add(
Breach(
contract=contract,
perpetrator=buyer_id,
victims=seller_id,
level=money_breach_level,
type="money",
)
)
self.__register_contract(buyer_id, money_breach_level)
self.__register_breach(buyer_id, money_breach_level, p, buyer_factory)
if self.borrow_on_breach and buyer_factory != self.compensation_factory:
# find out the amount to be paid to borrow the needed money
to_pay = math.ceil(missing_money * self.breach_penalty)
paid = buyer_factory.pay(to_pay)
missing_money -= (missing_money * paid) // to_pay
# execute the contract to the limit possible
self._execute(
product,
q,
p,
u,
buyer_factory,
seller_factory,
has_breaches=missing_product > 0 or missing_money > 0,
)
# return the list of breaches
return breaches
[docs] def complete_contract_execution(
self, contract: Contract, breaches: List[Breach], resolution: Contract
) -> None:
pass
[docs] def compensate(self, available: int, factory: Factory) -> None:
"""
Called by a factory when it is going bankrupt after liquidation
Args:
available: The amount available from liquidation
factory: The factory being bankrupted
Returns:
"""
agent_id = factory.agent_id
# get all future contracts of the bankrupt agent that are not executed
contracts = list(
itertools.chain(
*(factory.contracts[s] for s in range(self.current_step, self.n_steps))
)
)
owed = 0
total_owed = 0
nulled_contracts = []
for contract in contracts:
total_owed += contract.q * contract.u
if (
self.a2f[contract.partner].is_bankrupt
or contract.contract.nullified_at >= 0
):
continue
nulled_contracts.append(contract)
owed += contract.q * contract.u
if available <= 0 or owed <= 0:
self.record_bankrupt(factory)
return
# calculate compensation fraction
if available >= owed:
fraction = self.compensation_fraction
else:
fraction = self.compensation_fraction * available / owed
# calculate compensation and pay it as needed
for contract in nulled_contracts:
victim = self.agents[contract.partner]
victim_factory = self.a2f.get(victim.id, None)
# calculate compensation (as money)
compensation_quantity = int(fraction * contract.q)
compensation = min(available, compensation_quantity * contract.u)
if compensation < 0:
self.nullify_contract(contract.contract, 0)
continue
if self.compensate_immediately:
# pay immediate compensation if indicated
victim_factory.pay(-compensation)
available -= compensation
else:
# add the required products/money to the internal compensation inventory/funds to be paid at the
# contract execution time.
if contract.is_seller:
self.compensation_records[contract.contract.id].append(
CompensationRecord(
contract.product,
int((compensation // contract.u) * contract.u),
0,
True,
victim_factory,
)
)
else:
self.compensation_records[contract.contract.id].append(
CompensationRecord(-1, 0, compensation, False, victim_factory)
)
victim.on_contract_nullified(
contract.contract,
compensation if self.compensate_immediately else 0,
compensation_quantity,
)
self.nullify_contract(contract.contract, compensation_quantity)
self.record_bankrupt(factory)
@property
def winners(self):
"""The winners of this world (factory managers with maximum wallet balance"""
if len(self.agents) < 1:
return []
if 0.0 in [self.a2f[aid].initial_balance for aid, agent in self.agents.items()]:
balances = sorted(
(
(self.a2f[aid].current_balance, agent)
for aid, agent in self.agents.items() if aid != "SYSTEM"
),
key=lambda x: x[0],
reverse=True,
)
else:
balances = sorted(
(
(
self.a2f[aid].current_balance / self.a2f[aid].initial_balance,
agent,
)
for aid, agent in self.agents.items() if aid != "SYSTEM"
),
key=lambda x: x[0],
reverse=True,
)
max_balance = balances[0][0]
return [_[1] for _ in balances if _[0] >= max_balance]