Public distribution for: 1. cp2fsc for generating finite-state controllers for partially observable control problems, and 2. k_replanner for partially observable problems. The parser is based on a parser by Patrik Haslum. Some benchmarks for k_replanner use maps form the Hierarchical Open Graph (HOG) framework developed by Nathan Sturtevant from University of Alberta. Developed by Blai Bonet (c) 2011. =============================================================================== cp2fsc =============================================================================== The executable cp2fsc constructs a finite-state controller for solving a given partically observable problem. The problem is specified through a domain and problem file in an extension of the PDDL language. cp2fsc receives different paramenters, the most important being: --fsc-states <n> that specifies the number of internal states for the finite-state controller. For a complete list of paramenters, run cp2fsc without parameters. [still under construction] [take a look at the examples in benchmarks/fsc. See the README and experiments.sh files therein] =============================================================================== k_replanner =============================================================================== The executable k_replanner implements a replanning algorithm based on a classical planner. The classical planner used can be specified with a parameter. [still under construction]