HAPRC ver. 2.0

Dimitris Vrakas, Grigorios Tsoumakas, Nick Bassiliades and Ioannis Vlahavas

HAPRC is an adaptive planning system that automatically fine-tunes its planning parameters, based on the morphology of the problem in hand. HAPRC is an extended version of the HAP planner which can be fine tuned by the user. In order to do that the system is equipped with a rule-based system that has acquired its knowledge using Machine Learning techniques on data from a large number of experiments. The system uses 35 measurable characteristics in order to capture the morphology of each planning problem and 58 rules that associate these characteristics with specific setups for the planner. The planning system is able to change its planning direction, the heuristic function, the size of the internal memory structures and it can select among a number of add-on techniques that can speedup the planning process.

Download the system

source code in C++ (Gzipped tar file)

For inquiries, bugs, comments and suggestions please send an e-mail to Dimitris Vrakas.

Publications

  • "Rule Induction for Automatic Configuration of Planning Systems". D. Vrakas, G. Tsoumakas, N. Bassiliades and I. Vlahavas.Technical Report
  • "Learning Rules for Adaptive Planning". D. Vrakas, G. Tsoumakas, N. Bassiliades and I. Vlahavas.In Proceedings of the 13th International Conference on Automated Planning and Scheduling (ICAPS-03). pp. 82-91. Trento, Italy 2003. [pdf]
  • "Towards adaptive Heuristic Planning through Machine Learning". D. Vrakas, G. Tsoumakas and I. Vlahavas. In the Proceedings of the 21st Workshop of the UK Planning and Scheduling SIG. pp. 12-21. Delft, Nederlands 2002.[pdf]