DiscreteValueIteration.jl
DiscreteValueIteration.jl copied to clipboard
JuliaPOMDP
Value iteration solver for MDPs
Right now the max iterations are very low (100) and the algorithm may not converge. This can lead to confusing results.
The current parallel solver is on a separate branch and does not work with the newer releases of POMDPs.jl. It would be helpful to have a parallel solver for solving...
Currently this package has two solvers: ValueIterationSolver and SparseValueIterationSolver. The SparseValueIteration solver converts the problem to matrices and then uses linear algebra to solve. This is faster for almost all...
This PR adds a new solver: `ParallelValueIteration` which allows to benefits from several CPUs and speed VI. Notes: - it returns the same `ValueIterationPolicy` type - the parallelization relies on...
Currently, the solver only supports Gauss-Siedel value iteration. We should have an optional flag to do value iteration without the Gauss-Siedel method if the user wishes. There are a few...
I started using action_index() to order the Q matrix for consistency, but if the Q matrix isn't stored, we don't really need to make the problem writer implement it, so...