一种基于GA的多目标模糊决策方法研究  

Research on Multi-Objective Fuzzy Decision-Making Method Based on GA

作　　者：姚倡锋[1] 张定华[1] 彭文利[1] 

机构地区：[1]西北工业大学现代设计与集成制造教育部重点实验室,西安710072

出　　处：《系统仿真学报》2005年第10期2341-2344,2348,共5页Journal of System Simulation

基　　金：国家863/CIMS高科技计划重大专项(2001AA412020);西北工业大学博士论文创新基金(CX200412);航空科学基金(04H53064)

摘　　要：针对多目标决策中目标非标准化对决策的影响,提出一种基于相对隶属度的改进遗传算法。在求解过程中,首先计算出群体中所有个体的目标特征值,然后计算目标非标准化对决策的影响度和影响度因子;通过影响度因子将目标非标准化对决策的影响计算到相对隶属度中,将目标特征值矩阵转化为相对隶属度矩阵,构造基于相对隶属度的相对适应度函数,以减小目标非标准化对决策的影响;最后给出了应用该算法处理复杂零件协同制造中制造资源优化配置的一个实例,证明算法的可行性与有效性。To decrease the influence degree of objective＇s non-standardization in multi-objectives decision- making, an improved algorithm based on GA was proposed. During seeking the answer to multi-objectives decision- making, the objective eigenvalue of each chromosome was computed. Then, the influence degree and the influence degree coefficient caused by the objective＇s non-standardization were obtained in each generation. The relative membership degree of each chromosome to the optimized value was calculated based on the influence degree coefficient, and it included the influence caused by the objective＇s non-standardization. By changing the objective eigenvalue matrix into the relative membership degree matrix in each generation, the relative fitness function was set up and its value was calculated based on the relative membership degree to reduce the influence degree of objective＇s non-standardization to decision-making. At last, an example about resources ootimizing, deplovment was given to prove the algorithm＇s feasibilitv and validity.

关 键 词：遗传算法 影响度因子 相对隶属度 相对适应度 制造资源优化配置 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]