optimise.sequential.eu {bdpopt}R Documentation

Optimise Sequential Expected Utility

Description

Optimise the expected utility for a sequential decision problem. The optimisation proceeds by backward induction, computing the optimal decision and corresponding expected utility at each stage over a grid for the current state summarising the posterior distribution of the unknown parameter.

Usage

optimise.sequential.eu(dp, mins, maxs, steps, n.sims, state.start = NA)

Arguments

dp

A decision problem object constructed by calling sequential.dp.

mins

A numeric, atomic vector specifying the lower boundary of the grid points for each dimension of the state vector.

maxs

A numeric, atomic vector specifying the upper boundary of the grid points for each dimension of the state vector.

steps

A numeric, atomic vector specifying the step size between grid points for each dimension of the state vector.

n.sims

A numeric, atomic vector of length equal to dp$n.stages specifying the number of simulation draws to perform at each stage.

state.start

An optional start value for the state. If provided, this value effectively determines a fixed prior for the parameter of the decision problem (before the first stage). The first stage computations will then only be performed for this particular value. If the default value NA is used, computations for the first stage will be performed for all grid points, as for the subsequent stages.

Value

If state.start equals NA, then the value returned is a list with components

opt.decision

Function taking a stage and a state into the optimal decision corresponding to the closest grid point at that stage.

opt.utility

Function taking a stage and a state into the optimal expected utility corresponding to the closest grid point at that stage.

If an explicit value for state.start is provided, the value returned is a list with components

opt.stage1.decision

Optimal stage 1 decision.

opt.stage1.utility

Optimal stage 1 utility.

opt.decision

Function taking a stage (greater than 1) and a state into the optimal decision corresponding to the closest grid point at that stage.

opt.utility

Function taking a stage (greater than 1) and a state into the optimal expected utility corresponding to the closest grid point at that stage.

Author(s)

Sebastian Jobjörnsson jobjorns@chalmers.se

See Also

optimise.sequential.normal.eu


[Package bdpopt version 1.0-1 Index]