Planning Graph as a (Dynamic) CSP: Exploiting EBL, DDB and other CSP Search Techniques in Graphplan

TL;DR

This paper enhances the Graphplan planning algorithm by integrating advanced CSP search techniques like EBL and DDB, resulting in significant performance improvements demonstrated through empirical benchmarks.

Contribution

It adapts explanation-based learning, dependency directed backtracking, and other CSP techniques to Graphplan, significantly boosting its efficiency on benchmark problems.

Findings

Up to 1000x speedup on benchmark problems

EBL and DDB are most effective in improving performance

Empirical results validate the proposed augmentations

Abstract

This paper reviews the connections between Graphplan's planning-graph and the dynamic constraint satisfaction problem and motivates the need for adapting CSP search techniques to the Graphplan algorithm. It then describes how explanation based learning, dependency directed backtracking, dynamic variable ordering, forward checking, sticky values and random-restart search strategies can be adapted to Graphplan. Empirical results are provided to demonstrate that these augmentations improve Graphplan's performance significantly (up to 1000x speedups) on several benchmark problems. Special attention is paid to the explanation-based learning and dependency directed backtracking techniques as they are empirically found to be most useful in improving the performance of Graphplan.