点击次数：

论文类型：期刊论文

发表时间：2015-01-01

发表刊物：Journal of software engineering

收录刊物：EI、Scopus

卷号：9

期号：1

页面范围：77-86

摘要：Complex spatial layout design problems contain mutual conflicting1 objectives and constraints. These types of problems can be divided into several layout sub-problems based on their structure or assemble components. The main difficulties in solving these problems lie in combinational explosion of computational complexity, engineering complexity and pragmatic approaches in engineering practice. Recently, Cooperative Co-Evolutionary Algorithms (CCEA) have proven to be an efficient way to solve complex engineering problems. The CCEAs decompose the problem into several sub-problems using a cooperative co-evolutionary multi-population architecture. Based on the framework of CCEA, this study presents a cooperative co-evolutionary method with consistency coordination mechanisms for a complex layout design problem. This method extract objectives from all technical requirements by analyzing the nonlinear coupling relations among the sub-problems. The method then coordinates consistency among the sub-populations during a co-evolution process and ensures all sub-populations converge on a global consistency objective. Compared with a traditional CCEA method, results show that the proposed method can improve the computational accuracy of solutions. The proposed consistency coordination mechanisms can better reduce the conflicting of the sub-populations and make all the sub-populations sustained coevolve within short runtime. ? 2015 Academic Journals Inc.