calcRPO {MDP}R Documentation

Calculate the rentention payoff (RPO) for some states.

Description

Calculate the rentention payoff (RPO) for some states.

Usage

calcRPO(mdp, iW, iA, sId=1:mdp$states - 1, criterion="expected",
    iDur=0, rate=0.1, rateBase=365, g=0)

Arguments

mdp The MDP loaded using loadMDP.
iW Weight index we want to calculate RPO for.
iA The action index we calculate the RPO with respect to.
sId Vector of id's of the states we want to retrive.
criterion The criterion used. If expected used expected reward, if discount used discounted rewards, if average use average rewards.
iDur Index of duration such that discount rates can be calculated.
rate The interest rate.
rateBase The time-horizon the rate is valid over.
g The optimal gain (g) calculated (used if criterion = "average").

Details

The RPO is defined as the difference between the weight of the state when using action iA and the maximum weight of the node when using another predecessor different from iA.

Value

The rpo (matrix/data frame).

Author(s)

Lars Relund lars@relund.dk


[Package MDP version 1.0 Index]