基于自适应并行遗传算法的结构模糊动力优化
The Fuzzy Dynamic Optimization of Structure Based on Adaptive Parallel Genetic Algorithm
作者：张智畯(中冶赛迪工程技术股份有限公司 民用建筑设计院, 重庆 400013)；于建华(四川大学 建筑与环境学院,四川 成都 610065)；邱清水(西华大学 建筑与土木工程学院,四川 成都 610039)
Author：(CISDI Eng. CO, LTD, Chongqing 400013, China)；(School of Architecture and Environment,Sichuan Univ, Chengdu 610065,China)；(Architecture and Civil Eng. College of Xihua Univ., Chengdu 610039,China)
收稿日期：2007-09-18 年卷（期）页码：2009,41(1):68-73
期刊名称：工程科学与技术
Journal Name：Advanced Engineering Sciences
关键字：自适应;并行遗传算法;动力优化;模糊;最优水平截集因子
Key words：adaptation;paralle genetic algorithm;dynamic optimization;fuzzy method;optimal fuzzy level value λ*
基金项目：国家自然科学基金资助项目(102760307)
中文摘要
考虑到普通串行遗传算法处理结构动力优化问题时效率不高，引入并行遗传算法进行求解以提高效率，提出了一种新的自适应迁移算子,将其与串行遗传算法的自适应交叉,以及变异算子相结合，在4台微机组成的COW集群上实现了双向链式自适应迁移粗粒度并行遗传算法。同时为更加准确的描述实际工程问题，建立了结构模糊优化模型，提出了模糊水平截集因子的遗传编码确定方法，方便了模糊优化模型的求解。在最后给出的两个算例中，显示了自适应并行遗传算法的高效性以及模糊优化模型较之普通优化模型的合理性和有效性。
英文摘要
See to the fact that normal serial genetic algorithm handling dynamic optimization problem is not efficient enough, parallel genetic algorithm is suggested to solve this problem, and a new adaptive migration operator is proposed. Combining this adaptive migration operator with adaptive crossover operator and mutation operator of serial genetic algorithm, the adaptive bi oriented migration coarse grained parallel genetic algorithm which uses ring topology is implemented within a 4 processor COW environment. The fuzzy model of normal dynamic optimization is established for describing the real engineering problems more practicaly, the genetic coding method of calculating the optimal fuzzy level value λ* is presented, which can solve fuzzy model more conveniently. The two examples of the last are given to demonstrate the efficiency of the adaptive parallel genetic algorithm and certificated that the fuzzy optimization model is better than the normal optimization model, which is more practical and effective.
【关闭】
