Speculative computation with master agent belief revision and its algorithm for resource negotiation

Abstract

Speculative computation is a process which uses default hypothesis as tentative answer and continues computation when resource information cannot be obtained in time. In computing process, if the response is not consistent with belief, then master agent will revise its belief. The more information gotten, the more accurate the result of speculative computation is, and the smaller the risk of decision is, so the negotiation is a key mean in reducing risks of decision. In this paper, framework extended of speculative computation is presented, and the further negotiation framework and negotiation algorithm with belief revision are presented, the algorithm is embedded in speculative computation. Finally three experiments on goods transportation are given. The experiments have proved the algorithm can improve the accuracy of result of speculative computation, and reduce the risk of the result.