Research On Trust And Selection Of Virtual Service Resources In Service Ecosystems

With the development of network and software technology, service resource are moving away from a traditional entity service resources, being transferred to virtual service resources, being transferred to virtual service which support distributed services attributes, it constitute service ecosystem. Service ecosystem has some characters such as open, dynamic and resource sharing, facing the vast amounts of resources and services, service providers and service requesters must be make a decision making with reasonable and security under some uncertain factors without a predefined authorization relations and unknown for quality of service, Users increase the chance of selection while face how to identify and choose a service resources selection problem with safe, reliable and meet their individual preferences. Therefore, trust is a basic premise and necessary foundation for effective interaction of virtual service under the environment of virtual service.In service ecosystems, a large number of virtual service resources are distributed in the Internet resource pool, in the face of the user’s individualized demand, we need to select service resources quickly and efficiently to meet the needs of users, and build application service system. According to the above problem, our work includes the following aspects:1. Firstly, we discuss the architecture of services ecosystem, virtual service resources, service system, and concept intensions of the trust service.2. We gives the user’s trust in the trust evaluation index through the description of the characteristics of trust, Based on user behavior similarity evaluation, building evaluation model about trust (SOC-TPM) and trust mechanism of service resources.3. According to the disadvantage of the standard genetic algorithm (GA) into local optimum easily, particle swarm optimization (PSO) premature convergence caused loss of population diversity and parameter’s premature convergence caused loss of population diversity and in simulated annealing algorithm (SA) and too much still depends on some heuristic rules, put forward an improved genetic algorithm (AGA) which based on impact factor in order to realize the choice of service with multidimensional trust attribute.4. Through the experimental simulation verify the feasibility and effectiveness of AGA algorithm.