hxi903 / xpomdpy Goto Github PK

This work was a suplementary material of a research paper submitted to ACM TAAS, contains an extended POMPDy framework (https://github.com/pemami4911/POMDPy) for implementing discrete action/state POMDPs and MOMDPs in Python using single/scalar reward and multiple/vector reward. The current version is unstable and we are still working on the documentation.

• POMCP
• Value Iteration
• Linear Value Function Approximation

Please refer to https://github.com/pemami4911/POMDPy on how to use those originally supporter solvers.

• numpy >= 1.11
• matplotlib >= 1.4.3
• scipy >= 0.15.1
• future >= 0.16
• tensorflow >= 0.12