期刊名称：International Journal of Software Engineering and Its Applications
印刷版ISSN：1738-9984
出版年度：2012
卷号：6
期号：4
出版社：SERSC
摘要：Developing a system architecture that satisfies the functional requirements and keeping the quality attributes high is still an open challenge.The task of selecting the optimal combination and redundancy levels of components (which is known as RAP problem) that meet system-level constraints is hard since each one of these alternatives must be modeled in order to obtain the global optimal solution. This problem is classified as NP-complete and many researches have suggested use of metaheuristic to encode this challenge as an optimization problem for improving architectures. The search for optimal solution in metaheuristics techniques is guided by objective function/functions. Majority of studies in RAP area were focused on reliability and mostly simple cost functions were used. An adopted cost-with-discount function is proposed in this paper to increase the efficiency of the approach within especial conditions. Single Objective - Particle Swarm Optimization (SO-PSO) to automatically generate and to evaluate architecture is implemented in order to support design decision. The Anti-lock Breaking System (ABS) has been used as a case study to motivate, demonstrate the applicability, and to validate the study. And statistical t-test is performed to evaluate the approach. The evaluation of the approach has proven that; it is significantly impact on providing an efficient solution and it can be used to assist architectures to easily achieve their objectives since they can decide on the automatically provided design options according to technical and financial inclinations.
