双层多目标遗传算法及应用

Two-layer Multi-objective Genetic Algorithm and Its Application

为解决工程应用中的多目标优化问题,提出一种双层多目标遗传算法(two-layer multi-objective genetic algorithm,TLMOGA)。该算法根据个体间的支配关系将种群分成2层,并分别采用快速k最邻近算法和净强度函数法为这2层中的个体分配适应度。在此基础上,设计相应的个体排序和种群修剪策略,并确定了算法的整体流程。通过与传统多目标遗传算法进行比较,证明TLMOGA能够很好地保持解的收敛性和分布性,同时也具有较高的运算效率。最后,以ALSTOM气化炉基准控制器的参数优化整定为工程应用实例,进一步验证TLMOGA的有效性。仿真试验的结果表明,经优化后的控制系统,控制品质有了显著提高,达到了ALSTOM气化炉基准测试的要求。

In order to solve the multi-objective optimization problems in engineering application,a type of two-layer multi-objective genetic algorithm(TLMOGA) was proposed.The algorithm divided the population into two layers according to the dominance relationship,and the individuals in these two layers were assigned fitness by using fast k-nearest-neighbor algorithm and net strength function respectively.On this basis,relevant individual ranking and population pruning strategy were designed,and then the overall process of the algorithm was determined.Through comparing with traditional multi-objective genetic algorithms,it is proved that TLMOGA can well maintain the convergence and distribution of the solutions and also has relatively high computing efficiency.In the end,taking parameter optimization for the baseline controller of ALSTOM gasifier as an engineering application example,the effectiveness of TLMOGA was further verified.The simulation test results show that the optimized control system demonstrates significantly better control performance,which meets the requirements of ALSTOM gasifier benchmark tests.