mistat (version 2.0.3)

simOAB: Bayesian One-Armed Bernoulli Bandits process

Description

Simulate the expected number of trials on arm B before switching to the known arm A, and the expected reward.

Usage

simOAB(N, p, al, k, gam, Ns)

Arguments

N

number of trials.

p

the probability of reward on arm B (unknown).

al

the known probability of reward on arm A.

k

the initial sample size on arm B.

gam

Bayesian confidence level.

Ns

number of runs in the simulation.

Value

MeanValueStoppingTime

mean value at the stopping time

StandardDeviationST

standard deviation of the value at the stopping time

MeanValueExpectedReward

mean value of the expected reward

StandardDeviationST

standard deviation of the expected reward

See Also

dynOAB

Examples

# NOT RUN {
set.seed(123)
simOAB(N = 50, p = 0.6, al = 0.5, k = 10, gam = 0.95, Ns = 1000)
# }