基于自适应遗传算法的虚拟企业伙伴选择求解

针对虚拟企业合作伙伴的选择问题,根据虚拟企业的市场目标,采用层次分析算法确定评价因子的权重,然后建立多目标的虚拟企业伙伴选择优化模型.为简化模型,以最短的完工时间、最低的成本和最高的信誉度为目标.在遗传过程中对各个单目标进行排序选择,通过对交叉、变异算子进行自适应改进,采用最优保存策略迭代求出最佳合作伙伴及分配的具体工作量.通过与普通遗传算法的对比,仿真算例验证了改进后的算法的优越性.

Partner selection solving based on a self-adaptive genetic algorithm for virtual enterprise

To solve the partner selection problem in virtual enterprises,the weights of evaluation factors were defined based on the market objective by Analytic Hierarchy Process(AHP).And a multi-objective optimization model was proposed and accomplished by a self-adaptive Genetic Algorithm(GA).The objectives of GA were to achieve shortest completion time,lowest costs and best credit so as to simplify the model.This algorithm sorted and selected several single object fitness using quick-sorting algorithm based on total fitness,crossovers and mutates by using the self-adaptive probability,and found the global optimal solution by elitist strategy.Compared to standard GA,superiority of the improved GA was revealed by simulation example.