Parallel A* Algorithms and Their Performance on Hypercube Multiprocessors

In this paper we develop parallel A* algorithms suitable for distributed-memory machines. In parallel A* algorithms , ineeciencies grow with the number of processors P used, causing performance to drop signiicantly at lower and intermediate work densities (the ratio of the problem size to P). To alleviate this eeect, we propose a novel parallel startup phase and eecient dynamic work distribution strategies, and thus improve the scalability of parallel A* search. We also tackle the problem of duplicate searching by diierent processors, by using work transfer as a means to partial duplicate pruning. The parallel startup scheme proposed requires only (logP) time compared to (P) time for sequential startup methods used in the past. Using the Traveling Salesman Problem (TSP) as our test case, we see that our work distribution strategies yield speedup improvements of more than 30% and 15% at lower and intermediate work densities, respectively, while requiring 20% to 45% less memory, compared to previous approaches. Moreover, our simple duplicate pruning scheme provides an average reduction of 20% in execution time for up to 64 processors, compared to previous approaches that do not prune any duplicates.