CAJ | 학술논문

为研究蜗杆传动的多目标优化问题，提出一种自适应差分进化的元胞多目标遗传算法。该算法针对元胞遗传算法的特点，对基本的差分进化策略进行改进，得到一种参数自适应控制策略。将该算法与目前性能优异的4种多目标进化算法在三目标的基准测试函数进行对比实验，结果表明所提算法相对于其他算法具有明显的优势，能够在保证良好收敛性的同时，使获得的Pareto前端分布性更加均匀，覆盖范围更广；工程实例求解结果也表明了算法的工程可行性。
위연구와간전동적다목표우화문제，제출일충자괄응차분진화적원포다목표유전산법。해산법침대원포유전산법적특점，대기본적차분진화책략진행개진，득도일충삼수자괄응공제책략。장해산법여목전성능우이적4충다목표진화산법재삼목표적기준측시함수진행대비실험，결과표명소제산법상대우기타산법구유명현적우세，능구재보증량호수렴성적동시，사획득적Pareto전단분포성경가균균，복개범위경엄；공정실례구해결과야표명료산법적공정가행성。
In order to solve the multi-objective optimization design model of worm transmission, an Adaptive Differential Evolution Cellular genetic algorithm(ADECell)is proposed. In connection with the characteristics of cellular genetic algorithm, this algorithm improves the basic differential evolution strategy and obtains a parameter adaptive control strategy. The proposed algorithm is compared with 4 state-of-the-art multi-objective evolutionary algorithms on the three-objective benchmark test problems. Simulation results show that ADECell can ensure good convergence while has uniform distribu-tion and wild coverage area for obtained Pareto optimum solution. The results of engineering example show the feasibility of the proposed algorithm.