Solving difficult SAT problems by using OBDDs and greedy clique decomposition
FAW-AAIM'12 Proceedings of the 6th international Frontiers in Algorithmics, and Proceedings of the 8th international conference on Algorithmic Aspects in Information and Management
| Hi-index | 0.00 |
Recently, it has been suggested that BDD-based RePlanning A* (BDDRPA*), a BDD-based incremental version of A*, might be an efficient search method for solving path-planning problems in artificial intelligence. BDDRPA* combines ideas of BDD-based search and incremental search to repeatedly find shortest paths from a start vertex to a goal vertex while the topology of the graph changes. However, BDDRPA* only works well when vertices are added or deleted but does't consider the weighted edges. When the edge costs are changed, it doesn't work, and moreover, in BDDRPA*, the heuristic function $h$ is set to 0, so BDDRPA* is degenerated to BDD-based incremental breadth-first search. In this article, we consider BDD-based weighted and heuristic search methods and generalize BDDRPA* to be a real BDD-based incremental heuristic search algorithm (GBDDRPA*). We then show experimentally that GBDDRPA* indeed speeds BDDRPA* up on gridworlds and thus promises to provide a good foundation for building incremental heuristic BDD-search-based replanners.