A MIP-based optimal partition approach for service-oriented product line architecture

Software Product Line (SPL) offers developers a low-cost means to produce and maintain software product family with similar functions through reusing the domain-specific core assets. Service-Oriented Architecture (SOA) offers us an opportunity to extend this flexibility by creating dynamic product lines. Service-Oriented Product Line (SOPL), which combines SOA and SPL concepts and technologies, has attained an increasingly interest in software engineering community recent years, both in industry and academia. In this paper, an efficient architecture-centric development framework is proposed for SOPL to achieve mass customization, systematic planned reuse and high business flexibility As one of the most important phases of this framework mapping application architecture to BPEL attracts our more attention. In order to achieve the entire function of application architecture with minimum number of molecular services, we develop an optimal partition approach for application architecture based on the mapping relationship between abstract service components and concrete service candidates by using mixed integer programming (MIP) technique.