Implementations of serveral solvers for probabilistic planning problems such as MDP using Hindsight Optimization (HOP)
-
An implementation of the paper: Hindsight optimization for probabilistic planning with factored actions. The solver reduces the HOP problem into a integer linear program via satistfiability clauses and employ a generic ILP solver. This also includes a custom parser for the Symbolic Sperseus (SPUDD) file format.
- Dependencies: Gurobi, ANTLR v4 runtime
-
An currently researched HOP solver which reduces the HOP problem into a Maximum A Posteriori problem in an Markov Random Field.
- Dependency: David Sontag's MPLP solver