This Julia package implements the incremental pruning solver for partially observable Markov decision processes.
using Pkg # Pkg.Registry.add("https://github.com/JuliaPOMDP/Registry) # for julia 1.1+ # for julia 1.0 add the registry throught the POMDP package # Pkg.add("POMDPs") # using POMDPs # POMDPs.add_registry() Pkg.add("IncrementalPruning")
using IncrementalPruning using POMDPModels pomdp = TigerPOMDP() # initialize POMDP solver = PruneSolver() # set the solver policy = solve(solver, pomdp) # solve the POMDP
The result of
solve is a
Policy that contains the alpha vectors of the solution.
This solver implements the incremental pruning algorithm as described in Zhang and Liu (1996) and Cassandra et al. (1997). This solution method is exact (ϵ-optimal) but is much slower than modern approximate solution techniques. As such, it is only computationally feasible for small problems.
Cassandra, A., Littman, M., & Zhang, N. (1997). Incremental pruning: A simple, fast, exact method for partially observable Markov decision processes. Proceedings of the Thirteenth Annual Conference on Uncertainty in Artificial Intelligence (UAI-97), 54–61.
Zhang N. L., Liu W. (1996). Planning in stochastic domains: Problem characteristics and approximation. Technical Report HKUST-CS96-31, Hong Kong University of Science and Technology.
8 months ago