egttools.numerical.estimators.PairwiseComparisonEstimator

class PairwiseComparisonEstimator(game, pop_size, cache_size=100000, strategy_names=None, mutation_weights=None)

Bases: object

Wraps egttools.numerical.PairwiseComparisonNumerical and returns rich result objects with standard errors and optional progress bars.

Parameters:

• game (AbstractGame) – The game to associate with the estimator. nb_strategies is read from the game object.

• pop_size (int) – Population size.

• cache_size (int, optional) – LRU cache size for fitness computations (default 100 000).

• strategy_names (list[str], optional) – Human-readable strategy names used in result reprs.

• mutation_weights (array_like, optional) – Biases the mutation kernel away from uniform. Either a 1D array of length nb_strategies (target-strategy bias shared by all source strategies) or a 2D array of shape (nb_strategies, nb_strategies) (row i is the bias over target strategies when mutating away from strategy i). See PairwiseComparisonNumerical.set_mutation_weights for details. Defaults to uniform mutation when omitted.

Methods

estimate_absorption_probabilities
estimate_fixation_probability
estimate_mean_absorption_time
estimate_stationary_distribution
estimate_strategy_distribution

__getattr__(name)

__init__(game, pop_size, cache_size=100000, strategy_names=None, mutation_weights=None)

estimate_absorption_probabilities(beta, init_state, nb_runs, verbose=0, n_chunks=10)

Return type:

AbsorptionProbabilityResult

estimate_fixation_probability(invader, resident, nb_runs, nb_generations, beta, verbose=0, n_chunks=10)

Return type:

FixationResult

estimate_mean_absorption_time(beta, init_state, nb_runs, verbose=0, n_chunks=10)

Return type:

AbsorptionTimeResult

estimate_stationary_distribution(nb_runs, nb_generations, transitory, beta, mu, verbose=0, n_chunks=10, **kwargs)

Return type:

StationaryDistributionResult

estimate_strategy_distribution(nb_runs, nb_generations, transitory, beta, mu, verbose=0, n_chunks=10, **kwargs)

Return type:

StrategyDistributionResult

__annotations__ = {}