CAJ | 학술논문

针对差分进化算法在多Agent系统协作时容易过早的收敛到局部极小点，从而导致寻优性能不佳的问题，本文提出了一种基于参数自适应调整差分进化算法的多Agent系统协作模型，首先对标准差分进化算法的参数进行自适应调整，在选择阶段如果子代的适应度值差于相应的父代个体，调整相应的差分进化控制参数，然后将改进算法应用于多Agent系统的协作中，以提高其协作的性能。算法仿真试验结果表明，本文提出的基于参数自适应调整差分进化算法的多Agent系统协作模型相比较传统算法而言，具有较好的寻优能力和协作性能。
침대차분진화산법재다Agent계통협작시용역과조적수렴도국부겁소점，종이도치심우성능불가적문제，본문제출료일충기우삼수자괄응조정차분진화산법적다Agent계통협작모형，수선대표준차분진화산법적삼수진행자괄응조정，재선택계단여과자대적괄응도치차우상응적부대개체，조정상응적차분진화공제삼수，연후장개진산법응용우다Agent계통적협작중，이제고기협작적성능。산법방진시험결과표명，본문제출적기우삼수자괄응조정차분진화산법적다Agent계통협작모형상비교전통산법이언，구유교호적심우능력화협작성능。
According to differential evolution algorithm easy premature convergence to the local minimum points in multi Agent system coordination, thus leading to poor performance of optimization, this paper proposes a multi Agent system coordination model based on parameter adaptive adjustment of differential evolution algorithm. First, the parameters of the standard differential evolution algorithm for adaptive adjustment, in the selection phase if there is difference between offspring fitness value and its corresponding parent individuals, adjust the corresponding differential evolution control parameters. Then the improved algorithm is applied to multi Agent system, to improve the performance of the coordination. Simulation test results show that, compared with the traditional algorithm, the algorithm the proposed multi Agent system coordination model based on the parameter adaptive adjustment of differential evolution algorithm has a good performance of optimization and cooperation.