Issue No. 04 - April (1993 vol. 4)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/71.219757
<p>Analytical models and experimental results concerning the average case behavior ofparallel backtracking are presented. Two types of backtrack search algorithms areconsidered: simple backtracking, which does not use heuristics to order and prunesearch, and heuristic backtracking, which does. Analytical models are used to comparethe average number of nodes visited in sequential and parallel search for each case. Forsimple backtracking, it is shown that the average speedup obtained is linear when thedistribution of solutions is uniform and superlinear when the distribution of solutions isnonuniform. For heuristic backtracking, the average speedup obtained is at least linear,and the speedup obtained on a subset of instances is superlinear. Experimental results formany synthetic and practical problems run on various parallel machines that validate thetheoretical analysis are presented.</p>
Index Termsparallel backtracking; backtrack search algorithms; simple backtracking; heuristicbacktracking; speedup; parallel algorithms; search problems
V. Kumar and V. Rao, "On the Efficiency of Parallel Backtracking," in IEEE Transactions on Parallel & Distributed Systems, vol. 4, no. , pp. 427-437, 1993.