E-DPSO Algorithm Design and Demonstration about Dynamic Selection and Merging Process of ac-Service Flow Xiaona Xia School of Computer Science, Qufu Normal University, Rizhao Shandong, 276826, China Email: xiaxn@sina.com Baoxiang Cao and Jiguo Yu School of Computer Science, Qufu Normal University, Rizhao Shandong, 276826, China Email: bxcao@126.com, jiguoyu@sina.com Abstract— Based on the research and development about Web service, the business platform topology of ac-service flow is deepened from architecture-centric perspective. For Web service’s dynamic composition and self-adaptation meeting user requirement, architecture-centric merging among flows is put forward, in order to get this goal, optimizing previous flow granularities’ selection and composition. To satisfy flexible building and mapping of flows, the evolution algorithm E-DPSO based on discrete particle Swarm is designed. With dynamic character of service composition and the integrity of flow sequence, ac- service flow’s implementation strategy is given and finished demonstration. ac-service flow is viewed as the business organization clue, the platform framework structure of service implementation process is built for planning and decision-making. Index Terms— Web Service; ac-Service Flow; Dynamic Service Selection; Service Flow Merging; E-DPSO Algorithm; Discrete Particle Swam Algorithm; Architecture-centric I. WEB SERVICES AND SERVICE FLOW As a new type of Web application mode, Web service has seen significant development in recent years, its high interoperability, cross-platform and loose coupling is the basic technology of software architecture. How to composite all kinds of Web services ,and finish selecting and merging among service flows adaptively, in order to form new topology structure, satisfy the timely requirement for different users and achieve architecture goal with value-added difficult service structure, which is the development application requirement and R&D difficulty of platform[1]. Under normal circumstances, we call this model as service flow, that consists of multiple service granularities, every service granularity includes its own goal, implementation sequence, participation roles and collaboration services etc, but not establish specific service’s instances. In the Web environment, how to select the service instances satisfying architecture goal and implementation dynamically , form a executable composting service sequences, and adjust service flow topology adaptively, which is the key implementation issues of architecture-centric business[2, 3]. II. RELATED WORK BPEL4WS （ Business Process Execution Language for Web Services）[4] and EFLOW[5] are the static and dynamic technology supporting for Web service composition, but the dynamic semantic evolution of flow has not been effectively demonstrated, and this is an effective supporting factors of adaptive service composition. About standard workflow’s implementation, state calculus and formal characterization of dynamic semantics is the strong supplement about effective logical flow. Early, researchers construct service composition process by Petri[6] theory, with the help of corresponding composition mode, service granularities are associated and extended by Petri. In addition, the use of process algebra[7], Pi calculus[8] and OWL-S[9] describe the mapping relationship between service granularities and model topology rules to better achieve unambiguous understanding of flow’s services. On the autonomous Web service composition, reference[10] proposed ontology-based intention- behavior-realization mechanism, but, there is no demonstration for adaptive requirement logic about composition topology supported by semantic and dynamic description. Reference [11] demonstrated realizing action, execution, projection and planning for service composition sequences, and discussed and evolved the topology process for open architecture, this dynamic description logic provided effective way to the modeling, reasoning, service discovery and adaptive scheduling[12]. Meanwhile, reference [13] extended this kind of dynamic description, and combining the Corresponding author: Xiaona Xia JOURNAL OF COMPUTERS, VOL. 7, NO. 7, JULY 2012 1631 © 2012 ACADEMY PUBLISHER doi:10.4304/jcp.7.7.1631-1638