基于遗传算法和满意度函数的武器系统维修性分配规划问题研究

Study on Programming Problem for Maintainability-Distribution of Weapon System Based on Genetic Algorithm and Satisfactory Level Function

武器系统维修性分配是一个复杂的多目标规划问题.常规求解方法得到的结果难以满足其约束条件,并且难以体现不同目标间重要性的差别.据此,通过引入遗传算法(GA)和满意度函数解决了上述问题.首先利用GA求解单目标的最优解,建立各目标的满意度函数并综合为一个总满意度函数,最后利用GA求总满意度的最大值,即可获得模型的满意解.仿真算例表明,GA算法可适用于解决武器系统维修性分配问题,同时多目标规划中采用满意度函数法比线性加权法更可行.

The maintainability-distribution of weapon system is a complicated multi-objective programming. The result of the traditional method is hard to satisfy the corresponding re- striction and exhibit the importance differences between objectives. So genetic algorithm （GA） and satisfactory level function were cited to solve the problems. Firstly, GA was used to calculate the optimal result of each objective. Secondly, satisfactory level function of each objective was built and synthesized into a general one. Finally, the satisfactory solution of the model was drawn from calculation of the maximum of the general satisfactory level func- tion. A simulation example indicates that GA algorithm is more applicable to the problem for maintainability-distribution of weapon system, and satisfactory level function is more feasible than weighed linearity method in multi-objective optimization.