Abstract:

The composition of quality of service (QoS) is a challenge to the system with service-oriented architecture (SOA). The classical optimal criteria for the SOA of a manufacturing network (MN) are the time cost of production and the response time of the system. In this paper , a new method is proposed to optimize the QoS of a MN based on the genetic algorithm (GA). In the proposed method , first , the material and information flows in a MN are modeled respectively with the colored Petri net (CPN) and the queuing theory.Then ,the models are scheduled based on the GA , and a nearly optimal solution to the QoS is thus obtained. In the algorithm , each chromosome is made up of two gene-coding sections that respectively correspond to the priority rules and the weight coefficients and denotes the schedule solution to the material and information flows in the MN , and the choice , the crossing and the variation are all involved in the genetic operation. Moreover , the performance index values of eveIγchromosome in each generation are gained by simulation. The QoS fitness of the MN is then computed by means of fuzzy comprehensive evaluation. Simulation results show that the proposed method can effectively optimize the QoS of a manufacturing network

Key words: manufacturing network, quality of service, genetic algorithm, colored Petri net, queuing theory, fuzzycomprehensive evaluation