configuration_multiplayers module

Configuration for the simulations, for the multi-players case.

configuration_multiplayers.HORIZON = 10000

HORIZON : number of time steps of the experiments. Warning Should be >= 10000 to be interesting “asymptotically”.

configuration_multiplayers.REPETITIONS = 200

REPETITIONS : number of repetitions of the experiments. Warning: Should be >= 10 to be statistically trustworthy.

configuration_multiplayers.DO_PARALLEL = True

To profile the code, turn down parallel computing

configuration_multiplayers.N_JOBS = -1

Number of jobs to use for the parallel computations. -1 means all the CPU cores, 1 means no parallelization.

configuration_multiplayers.NB_PLAYERS = 3

NB_PLAYERS : number of players for the game. Should be >= 2 and <= number of arms.

configuration_multiplayers.collisionModel(t, arms, players, choices, rewards, pulls, collisions)

The best collision model: none of the colliding users get any reward

configuration_multiplayers.VARIANCE = 0.05

Variance of Gaussian arms

configuration_multiplayers.CACHE_REWARDS = False

Should we cache rewards? The random rewards will be the same for all the REPETITIONS simulations for each algorithms.

configuration_multiplayers.NB_ARMS = 6

Number of arms for non-hard-coded problems (Bayesian problems)

configuration_multiplayers.LOWER = 0.0

Default value for the lower value of means

configuration_multiplayers.AMPLITUDE = 1.0

Default value for the amplitude value of means

configuration_multiplayers.ARM_TYPE

alias of Arms.Bernoulli.Bernoulli

configuration_multiplayers.ENVIRONMENT_BAYESIAN = False

True to use bayesian problem

configuration_multiplayers.MEANS = [0.1, 0.26, 0.42000000000000004, 0.58, 0.74, 0.9]

Means of arms for non-hard-coded problems (non Bayesian)

configuration_multiplayers.configuration = {'averageOn': 0.001, 'collisionModel': <function onlyUniqUserGetsReward>, 'environment': [{'arm_type': <class 'Arms.Bernoulli.Bernoulli'>, 'params': [0.1, 0.26, 0.42000000000000004, 0.58, 0.74, 0.9]}], 'finalRanksOnAverage': True, 'horizon': 10000, 'n_jobs': -1, 'players': [<Policies.SIC_MMAB.SIC_MMAB object>, <Policies.SIC_MMAB.SIC_MMAB object>, <Policies.SIC_MMAB.SIC_MMAB object>], 'plot_lowerbounds': False, 'repetitions': 200, 'successive_players': [[CentralizedMultiplePlay(kl-UCB), CentralizedMultiplePlay(kl-UCB), CentralizedMultiplePlay(kl-UCB)], [Selfish(kl-UCB), Selfish(kl-UCB), Selfish(kl-UCB)], [rhoRand(kl-UCB), rhoRand(kl-UCB), rhoRand(kl-UCB)], [MCTopM(kl-UCB), MCTopM(kl-UCB), MCTopM(kl-UCB)]], 'verbosity': 6}

This dictionary configures the experiments

configuration_multiplayers.nbArms = 6

Number of arms in the first environment