Research Of Service Composition Strategy In Procurement Logistics System Based On SOA

Frequent occurrence of business process reengineering(BRP) in procurement logisticssystem is mostly caused by the complexity, dynamic and dispersion of the logistics industry,as well as the diversity of logistic management. In this paper, an integrated procurementlogistics service system is designed based on SOA to reduce the cost of the systemreengineering and to improve the efficiency. Nowadays, the research about how to improvethe quality of service is a hot spot in the field of dynamic service composition. At the sametime, the problem about how to design a more suitable algorithm is needed to be solved toimprove the efficiency of procurement logistics service composition on SOA. The paperstudied a SOA service composition strategy of procurement logistics system by a carefullyanalysis from domestic and international researches about service composition.In order to make up for the lack of service quality analysis of traditional SOA system, themulti-level graphical value model is introduced in the procurement logistics system based onSOA. The value model can be used not only to evaluate qualities of atomic services withstandard evaluation, but also to evaluate quality of service composition. In the model, it give aformal description of customer needs in both functional and non-functional aspects; theextended UML sequence diagram is applied to describe the service process; and further use Picalculus and its value state graph to verify that the target value can be realized or not; Also, agroup of value calculation functions are employed to quantitatively measure procurementlogistics system to the satisfaction of customer needs.In order to improve the service quality and efficiency of procurement logistics system,combined with the value model, an improved genetic algorithm is proposed by analyzingtraditional genetic algorithm using in services composition. Firstly, the coding scheme ofinteger array based on mapping is designed; Secondly, the roulette method based on elitistselection strategy is used in the selection operator, and the adaptive adjustment rule ofcrossover probability is used in the crossover operator; then the mutation operator based onPSO operator is designed to improve ability of the mutation operator to adapt to theenvironment, and also to improve the speed of convergence. After all, based on the valuemodel, the fitness function of global target value is designed. Global target value consideredthe evaluation indexes of combination services of purchasing logistics system, which can usethe value calculation functions to quantitatively calculate the fulfillment of global target valueof service combination.At the end of this paper, an experimental analysis of the improved service compositionstrategy in a raw material procurement logistics service system is given. Firstly, using thevalue model, the customer needs are described, including the target value definition, thedependence of the target values and the service process model; secondly, the database and the algorithm example of service combination module is designed; then the verification andanalysis of the service combination module are given through experiment. The results of theexperiment show that, the procurement logistics system based on SOA introduced the valuemodel can effectively verify the service composition in both functional and non-functionalaspects, make up for the lack of service quality analysis of traditional SOA system; theimproved algorithm of service combination ensured the service quality of service combination,improved the operation efficiency of service combination of procurement logistics.