解空间编码遗传算法在三维布局中的应用  

Genetic Algorithm Basedon Solution-Vector Coding and Its Application to Three-Dimensional Layout Design

作　　者：张刚[1] 殷国富[1] 邓克文[2] 成尔京[1] 

机构地区：[1]四川大学,成都610065 [2]中国工程物理研究院结构力学研究所,绵阳621900

出　　处：《中国机械工程》2006年第1期79-83,共5页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(50275100);中国工程物理研究院结构力学研究所资助项目

摘　　要：采用改进遗传算法———解空间编码的遗传算法求解带性能约束的复杂三维布局设计问题。算法对三维空间布局问题进行了预处理,直接将问题的求解变量作为染色体基因进行编码,降低了算法的编码和解码难度。在求解过程中,对不同变量类型采用不同的交叉和变异算子进行计算,并且在变异前对复制或交叉所得个体的某些变量随机加一个扰动量,以控制其搜索范围,算法以计算结果满足设计要求为算法终止准则。通过对二维和三维带性能约束布局算例的求解,验证了该算法的可行性和有效性。To solve three-dimensional layout problems with behavior constraints, a genetic algorithm based on solution-vector coding was proposed for solution spatial layout optimization problems. In the algorithm, the variables of optimization function were regarded directly as genes of chromosomes and were converted into code. By this means, the coding and decoding difficulties of the algorithm were reduced and the efficiency of the algorithm was improved. In the course of running, the algorithm was not terminated until the final results can meet the desired layout design. For the different gene-variables, there were different cross and mutating operators. Moreover, to control the searching range of the algorithm, it was necessary to add a fluctuant scale to an individual before mutating. In addition, some disposal methods of math-models of spatial layout problems and the algo rithm＇s solution process were described. To test the feasibility and validity of the algorithm, a two and a three-dimensional layout problems with behavior constraints were presented and solved by this algorithm. Their layout results indicate solution-vector coding genetic algorithm is feasible and valid.

关 键 词：遗传算法 解空间编码 三维布局 性能约束 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]