An uncertainty theory and genetic algorithm-based performance optimization method for service-oriented enterprise networks

Abstract

Digital- Technologies based service-oriented enterprise networks (SOEN) are emerging. The performance optimization for SOEN is very complex. It is needed: first, to try to reduce an abstract service's waiting-number of its physical services; second, to optimize the service-orchestration-schemes of the physical services waited before. This paper introduces an uncertainty theory and genetic algorithm-based performance optimization method (UGPO) to solve these problems. The uncertainty theory of UGPO is used to determine an abstract service's waiting-number for its physical services. The genetic algorithm of UGPO is used to search for the best service-orchestration-scheme from a large number of potential combinatorial schemes by the physical services waited before. UGPO has made use of the information of the physical services' arriving rules and performance distribution functions thoroughly, which will improve the computational efficiency for the scheme design and performance optimization of the SOEN in digital enterprise environments. © Springer-Verlag Berlin Heidelberg 2010.