Package jmarkov.jmdp.solvers

This package contins the framwork of solvers used by jMDP to solve Markov Decision Processes.

See:
Description

Interface Summary

LPSolver<S extends State,A extends Action>
MpsLpSolver<S extends State,A extends Action>	This interface define the minimium elements for creating a MPS file.

Class Summary

AbstractAverageSolver<S extends State,A extends Action>	Structural class for average cost solvers to extend.
AbstractDiscountedSolver<S extends State,A extends Action>	This is a structural class that must be extended by classes solving the dicounted cost minimization problem.
AbstractFiniteSolver<S extends State,A extends Action>	Structural class for solvers to extend in order to solve finite horizon problems.
AbstractInfiniteSolver<S extends State,A extends Action>	Structural class to be extended by solvers in order to solve infinite horizon problems
AbstractTotalSolver<S extends State,A extends Action>	Structural class to be extended by solvers in order to solve the total cost criteria for an infinite horizon problem
FiniteSolver<S extends State,A extends Action>	This class belongs to the set of default solvers included in the jmdp package.
LPBCLAverageSolver<S extends State,A extends Action>	This solver solves a average-cost infinite horizon MDP by building and solving a linear problem using as interface Xpress BCL.
LPBCLDiscountedSolver<S extends State,A extends Action>	This solver solves a discounted infinite horizon MDP by building and solving a linear problem using as interface Xpress BCL.
MpsLpAverageSolver<S extends State,A extends Action>	This class builds the Dual Linear Program for an average infinite horizon MDP in a MPS file.
MpsLpDiscountedSolver<S extends State,A extends Action>	This class builds a Linear Program for a discounted infinite horizon MDP in a MPS file.
PolicyIterationSolver<S extends State,A extends Action>	This class solves infinite horizon discounted problems using the policy iteration algorithm.
ProbabilitySolver<S extends State,A extends Action>	This class is designed to calculate the long run probabilities of infinite horizon problem.
RelativeValueIterationSolver<S extends State,A extends Action>	This class solves the average cost criteria for infinite horizon problems
Solver<S extends State,A extends Action>	Structural class for every solver.
StochasticShortestPathSolver<S extends StateC,A extends Action>	This solver gives a solution for the minimization of the total cost criterion for an infinite horizon MDP.
ValueIterationSolver<S extends State,A extends Action>	This class belongs to the set of default solvers included in the jmdp package.

Package jmarkov.jmdp.solvers Description

This package contins the framwork of solvers used by jMDP to solve Markov Decision Processes. See the jMDP manual for details.