Sokoban: Enhancing general single-agent search methods using domain knowledge

Artificial Intelligence

Volumn 129, Issue 1-2, 2001, Pages 219-251

  Junghanns, Andreas ^a   Schaeffer, Jonathan ^a  

^a UNIVERSITY OF ALBERTA   (Canada)

Author keywords

IDA*; Pattern database; Pattern search; Rapid random restart; Single agent search; Sokoban; Transposition table

Indexed keywords

COMPUTATIONAL COMPLEXITY; DATA STRUCTURES; DATABASE SYSTEMS; KNOWLEDGE BASED SYSTEMS; PATTERN MATCHING; RANDOM PROCESSES; STATE SPACE METHODS;

RAPID RANDOM RESTART THEORY; SINGLE-AGENT SEARCH METHODS;

ARTIFICIAL INTELLIGENCE;

EID: 0035369907     PISSN: 00043702     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0004-3702(01)00109-6     Document Type: Article

References (23)

1. Some methods of controlling the tree search in chess programs
2. Sokoban is PSPACE-complete
3. Pattern databases
4. Boosting combinatorial search through randomization
5. Criticizing solutions to relaxed models yields powerful admissible heuristics
6. Relevance cuts: Localizing the search
7. Single-agent search in the presence of deadlock
8. Sokoban: Evaluating standard single-agent search techniques in the presence of deadlock
9. Domain-dependent single-agent search enhancements
10. Sokoban: Improving the search with relevance cuts
11. Pushing the envelope: Planning, propositional logic and stochastic search
12. A primal method for minimal cost flows
13. Macro-operators: A weak method for learning
14. Linear-space best-first search
15. Finding optimal solutions to Rubik's Cube using pattern databases
16. The Hungarian method for the assignment problem
17. BIDA*: An improved perimeter search algorithm
18. Quantitative results concerning the utility of explanation-based learning
19. Enhanced iterative-deepening search
20. Chess 4.5 - The Northwestern University chess program