Parallel Search of Strongly Ordered Game Trees

作者: T. A. Marsland , M. Campbell

DOI: 10.1145/356893.356895

关键词:

摘要: The alpha-beta algorithm forms the basis of many programs that search game trees. A number methods have been designed to improve utility sequential version this algorithm, especially for use in game-playing programs. These enhancements are based on observation alpha beta is most effective when best move each position considered early search. Trees so-called strong ordering property not only practical importance but possess characteristics can be exploited both and parallel environments. This paper draws upon experiences gained during development which chess Over past decade major developed by people building programs, these will surveyed compared here. balance contains a study contemporary searching trees parallel, using an arbitrary independent processors. To make efficient processors, one must clear understanding basic properties actually traversed cutoffs occur. provides such insights concludes with brief description amore » refinement standard problem. 33 references.« less

参考文章(27)
K. Thompson, COMPUTER CHESS STRENGTH Advances in Computer Chess. pp. 55- 56 ,(1982) , 10.1016/B978-0-08-026898-9.50008-5
Gerard Maurice Baudet, The design and analysis of algorithms for asynchronous multiprocessors. Carnegie Mellon University. ,(1978)
Monroe M. Newborn, Recent Progress in Computer Chess Advances in Computers. ,vol. 18, pp. 154- 205 ,(1988) , 10.1007/978-1-4613-8716-9_9
James J. Gillogly, Performance analysis of the technology chess program. Carnegie Mellon University. ,(1978) , 10.21236/ADA056072
Murray S. Campbell, T.A. Marsland, A comparison of minimax tree search algorithms Artificial Intelligence. ,vol. 20, pp. 347- 367 ,(1983) , 10.1016/0004-3702(83)90001-2
James J Gillogly, The technology chess program Artificial Intelligence. ,vol. 3, pp. 145- 163 ,(1972) , 10.1016/0004-3702(72)90045-8
R. S. Bird, Tabulation Techniques for Recursive Programs ACM Computing Surveys. ,vol. 12, pp. 403- 417 ,(1980) , 10.1145/356827.356831