Statistically assessing distributed computing approaches

This paper investigates the solution of manufacturing scheduling problems u sing different distributed computing schemes. The focus of this paper is on determining the statistically significant effects of a distributed compute r model with respect to elapsed solution time and the amount of problem spa ce investigated. Statistical inferences are drawn with respect to the numbe r of processors and how the problem is partitioned. The best combinations o f the number of processors and partitioning schemes are presented with resp ect to problem scenario and size. Additionally, comparisons are made with r espect to the performance of different configurations of the distributed co mputer model. (C) 2000 Elsevier Science Ltd. All rights reserved.