Fast recursive formulations for best-first search that allow controlled use of memory

作者: A. Bagchi , Anup K. Sen

DOI:

关键词:

摘要: MREC is a new recursive best-first search algorithm which combines the good features of A* and IDA*. It closer in operation to IDA*, does not use an OPEN list. In order execute, all needs sufficient memory for its implicit stack. But it can also be fed at runtime parameter M tells how much additional available use. this extra memory, stores as possible explicit graph. When = 0, identical when > make far fewer node expansions than This advantageous problems where time expand significant. Extensive runs on variety problems, involving graphs that may or trees, indicate with 0 IDA* such 15- puzzle suitable, while large fast expansion negligible.

参考文章(6)
A. Bagchi, K. V. Viswanathan, An exact best-first search procedure for the constrained rectangular guillotine knapsack problem national conference on artificial intelligence. pp. 145- 149 ,(1988)
A. Bagchi, A. Mahanti, Three approaches to heuristic search in networks Journal of the ACM. ,vol. 32, pp. 1- 27 ,(1985) , 10.1145/2455.2458
P.P. Chakrabarti, S. Ghose, A. Acharya, S.C. de Sarkar, Heuristic search in restricted memory (research note) Artificial Intelligence. ,vol. 41, pp. 197- 222 ,(1989) , 10.1016/0004-3702(89)90010-6
Dura W. Sweeney, Katta G. Murty, John D. C. Little, Caroline Karel, An Algorithm for the Traveling Salesman Problem ,(2019)
Richard E. Korf, Depth-first iterative-deepening: an optimal admissible tree search Artificial Intelligence. ,vol. 27, pp. 97- 109 ,(1985) , 10.1016/0004-3702(85)90084-0
Judea Pearl, Heuristics : Intelligent Search Strategies for Computer Problem Solving xvii, 382 p. : ill. California: Addison-Wesley Pub. Co., 1984. includes bibliography: p. 363-370 and index. -- (Artificial Intelligence series). ,(1984)