infoMat {MDP}R Documentation

Info matrices or data frames of the state and action space.

Description

Info matrices or data frames of the state and action space. Functions that generate matrices with info about the HMDP model under consideration.

Usage

stateIdxMat(file=stateIdx.bin)\cr
stateIdxDf(file="stateIdx.bin", labels="stateIdxLbl.bin")\cr
transProbMat(file="transProb.bin")\cr
actionIdxMat(file="actionIdx.bin")\cr
actionIdxDf(file="actionIdx.bin", labels="actionIdxLbl.bin")\cr
actionInfo(file="actionIdx.bin" , weightFile="actionWeight.bin",\cr
transPrFile="transProb.bin", labels="actionIdxLbl.bin")\cr
actionWeightMat(file="actionWeight.bin",labels="actionWeightLbl.bin")\cr

Arguments

prefix A character string with the prefix added to til file(s).
file The HMDP binary file containing the description under consideration.
labels The HMDP binary file containing the labels under consideration.
costFile The HMDP binary file containing the action costs.
transPrFile The HMDP binary file containing the transition probabilities.

Details

stateIdxMat return a matrix with columns (sId, d0, s0, a0, ...) where sId is the state row id, d0 the index of the stage at level 0, s0 the index of the state and a0 the index of the action. Moreover, if the HMDP has more than one level columns (d1, s1, a1, ...) are added.

stateIdxDf return a data frame with the same columns as in stateIdxMat plus another column containing the labels.

actionIdxMat return a matrix with columns (aId, ...) where aId is the action row id and ... are alternating pairs (scp, idx), one for each possible transition where scp is the scope that can be 4 values: 2 - A transition to a child process (stage zero in the child process), 1 - A transition to next stage in the current process, 0 - A transition to the next stage in the father process. the idx in the pair denote the index of the state at the stage considered. Finally, if scope = 3 then a transition to the state with sId = idx is considered.

actionIdxDf return a data frame with the same columns as in actionIdxMat plus another column containing the labels.

actionWeightMat return a matrix with columns (aId, ...) where aId is the action row id and ... are the weights of the action.

transProbMat return a matrix with columns (aId, ...) where aId is the action row id and ... are the probabilities of the action.

actionInfo return a matrix with columns from actionIdxMat, actionCostMat and transProbMat if labels is NULL. If labels not are NULL then a data frame are returned with a label column too.

Value

Matrix or data frame.

Author(s)

Lars Relund lars@relund.dk

Examples

stateIdxMat()
stateIdxDf()
actionIdxMat()
actionIdxDf()
actionWeightMat()
transProbMat()
a<-actionInfo()
a[order(a$sId),]

[Package MDP version 1.0 Index]