examples.el_farol_bar_model.agents package

Submodules

examples.el_farol_bar_model.agents.agents module

Agents for the el farol bar problem

class examples.el_farol_bar_model.agents.agents.LikeCrowdedPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect

select_game()

The agent select a play from a strategy

class examples.el_farol_bar_model.agents.agents.LikeSixtyPercentPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect

class examples.el_farol_bar_model.agents.agents.Player(simulation, model, agent_number, agent_def)

Bases: sphinx.ext.autodoc.importer._MockObject

A basic player in the El Farol Bar Problem

game_payoff()

Get the game payoff

get_frequency(frequency)

“Get the frequency in the bar

play()

The agent plays a strategy

select_game()

The agent select a play from a strategy

step()

agent step

class examples.el_farol_bar_model.agents.agents.RandomPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect

select_game()

Random selection of the game

examples.el_farol_bar_model.agents.el_farol_bar_action_set module

Basic Strategy Class implementation

class examples.el_farol_bar_model.agents.el_farol_bar_action_set.LikeCrowded(agent, recall)

Bases: examples.el_farol_bar_model.agents.el_farol_bar_action_set.Strategy

Strategy Like Crownded

payoff(agent_play)

Specialized by subclass

select_strategy()

Like Crownded Strategy

class examples.el_farol_bar_model.agents.el_farol_bar_action_set.LikeSixtyPercent(agent, recall)

Bases: examples.el_farol_bar_model.agents.el_farol_bar_action_set.Strategy

Strategy under sixty percent

class examples.el_farol_bar_model.agents.el_farol_bar_action_set.RandomPlay(agent)

Bases: examples.el_farol_bar_model.agents.el_farol_bar_action_set.Strategy

Strategy under random

select_strategy()

Random Strategy

class examples.el_farol_bar_model.agents.el_farol_bar_action_set.Strategy(agent, recall)

Bases: object

Implementation of the strategy class

STRATEGY = ['GOING', 'NOT GOING']

get_frequency(frequency)

The strategy gets the current frequency in the bar

payoff(agent_play)

Specialized by subclass

select_game()

select_strategy()

selected_predictor()

Returns the selected predictor and its fitness

set_all_predictors()

set all predictors in the model

set_predictors()

Set the predictors for this strategy

update_fitness()

Updates the payoffs of the strategies

examples.el_farol_bar_model.agents.el_farol_bar_prediction_model module

Prediction Model for the El Farol Problem

class examples.el_farol_bar_model.agents.el_farol_bar_prediction_model.FitnessFunction(owner)

Bases: object

Fitness Function for the el-farol bar model following: Rand, W., & Stonedahl, F. (2007). The El Farol bar problem and computational effort:

Why people fail to use bars efficiently. Northwestern University, Evanston, IL.

f(S,t) = sum_(i=t-L)^(t-1) abs(p(S,i) - a(i)) Where: p(S,i) is the prediction of strategy S on time i and a(i) is the observed attendance on time i and L is the memory reacall lenght.

fitness(observed, predicted)

Calculate Strategy fitness

class examples.el_farol_bar_model.agents.el_farol_bar_prediction_model.PredictionModel(agent, recall)

Bases: object

The prediction model that the agent uses to choose the strategy Arthur’s paper (1994) strategies 1. the same as last week’s 2. a mirror image around 50 of last week’s 3. a (rounded) average of the last four weeks 4. the trend in last 8 weeks, bounded by [0,100] 5. the same as 2 weeks ago (2-period cycle detector) 6. the same as 5 weeks ago (5-period cycle detector)

agent_memory()

Returns the agent memory

get_best_prediction()

Get the prediction from the fittest predictor

get_best_predictor(predictors)

Get the prediction from the fittest predictor

get_fitness(predictor_name)

Get the fitness of one predictor

get_fitness_of_predictor(predictor_name, going, target_no)

Updates the fitness of one predictor

make_prediction(selected_predictors)

The prediction based on the model

no_of_agents()

Returns the number of agents

st_mirror_50_lw()

Strategy 2. a mirror image around 50 of last week’s

st_round_avg_4w()

Strategy 3. a (rounded) average of the last four weeks

st_same_2w()

Strategy 5. the same as 2 weeks ago (2-period cycle detector)

st_same_5w()

Strategy 6. the same as 5 weeks ago (5-period cycle detector)

st_same_lw()

Strategy 1. the same as last week’s

st_trend_8w()

Strategy 4. the trend in last 8 weeks, bounded by [0,100]

step()

get agent step

update_fitness(predictors, going, target_no)

Updates the fitness of the predictors

Module contents

class examples.el_farol_bar_model.agents.Player(simulation, model, agent_number, agent_def)

Bases: sphinx.ext.autodoc.importer._MockObject

A basic player in the El Farol Bar Problem

game_payoff()

Get the game payoff

get_frequency(frequency)

“Get the frequency in the bar

play()

The agent plays a strategy

select_game()

The agent select a play from a strategy

step()

agent step

class examples.el_farol_bar_model.agents.RandomPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect

select_game()

Random selection of the game

class examples.el_farol_bar_model.agents.LikeCrowdedPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect

select_game()

The agent select a play from a strategy

class examples.el_farol_bar_model.agents.LikeSixtyPercentPlayer(simulation, model, agent_number, agent_def)

Bases: examples.el_farol_bar_model.agents.agents.Player

A player that always defect