Abstract
In this paper we discuss techniques for online generation
of macro-actions as part of the planning process and demonstrate their use in a forward chaining search planning framework. The macroactions learnt are specifically created at places in the search space where the heuristic is not informative. We present results to show that using macro-actions generated during planning can improve
planning performance.
Original language | English |
---|---|
Number of pages | 6 |
Publication status | Published - Dec 2005 |
Event | Twenty-fourth UK Planning and Scheduling SIG - London, UK Duration: 15 Dec 2005 → 16 Dec 2005 |
Conference
Conference | Twenty-fourth UK Planning and Scheduling SIG |
---|---|
City | London, UK |
Period | 15/12/05 → 16/12/05 |
Keywords
- macro-actions
- searching
- search algorithm
- forward-chaining planning