Research On Service Composition For Crowdsourcing Distribution

The rapid development of logistics industry and the growth of the quantity and quality of consumer demand,Terminal distribution of logistics efficiency has become the key research problems,and the emergence of crowdsourcing distribution provides a new way to solve the problem,which integrates public resources to provide distribution services.But when many existing platform addresses the problem of orders allocation,they usually make delivery staff on the platform scramble alone for real-time,This way makes orders allocation appear per capita imbalances,reduce the efficiency of distribution,long waiting time for orders,which indirectly causes users loyalty to the platform be lower.How to extract and use large amount of information resource in the platform that makes orders can be reasonable allocated is the one of key problems that platform to maintain and attracket users.So this paper put forward that all orders on the platform should be subpackaged before delivered.Inspired by the Web service combination,orders allocation can be assimilated to a service.For orders can be reasonable allocated,which can be assimilate with Web service composition.For which,the following work has been done:(1)The service description for crowdsourcing distribution.In order to In order to understand which information elements in the distribution service can be used as a combination and make the crowdsourcing distribution platform more rapid identify,store and call information.This article used the method of Web service descripition for reference to analysis the requirement of crowdsourcing distribution.Through the method of modeling ontology to describe the service of crowdsourcing distribution,and used Semantic Works tools complete interface implementation.(2)The service composition for crowdsourcing distribution.Analyzing crowdsourcing distribution requirements and extracting the basics of service composition,combined with the actual that use obstacle distance to replace the euclidean distance as the main factors in the process of combination,and to improve the efficiency of distance calculation so we refer to the BSP(Binary Space Partitioning)criteria.In the obstacle distance in position of supply and demand both sides for difference measures with the improved k-medoids clustering algorithm to solve the problem of service composition.As a result,the differences inside of various combination has been reduced partly in quantity of service,the distribution of all these service position is concentrated and concentrated.This way reduced the occurrence of timeout,cross-regional distribution,long distance distribution and so on.(3)Application case study.D company is selected to collect the real order data to carry out the case study of crowdsourcing distribution service.Working out the results of service composition,contrasting and analyzing the result of before and after through the evaluation model to prove that the proposed k-medoids clustering algorithm can get effective service combination,considering the supply and demand both sides location and capacity constraints of the crowdsourcing distribution service on the overtime rate,the per capita amount of orders,the per capita both trip distance and the consumption of time such as is better than not combined.