CAJ | 학술논문

提出了一种基于网格支配的微型多目标遗传算法，该算法在求解较多目标函数的优化问题时具有较好的收敛性和较高的计算效率。该算法引入网格支配概念并结合微型多目标遗传算法，在每一代进化种群中计算各个个体的网格值、网格拥挤距离和网格坐标点距离，根据网格支配分级和网格选择机制策略选取精英个体，并对其进行交叉和变异操作，使其朝前沿面收敛以获得 Pareto 最优解。4个测试函数和2个工程实例验证了该算法的有效性。
제출료일충기우망격지배적미형다목표유전산법，해산법재구해교다목표함수적우화문제시구유교호적수렴성화교고적계산효솔。해산법인입망격지배개념병결합미형다목표유전산법，재매일대진화충군중계산각개개체적망격치、망격옹제거리화망격좌표점거리，근거망격지배분급화망격선택궤제책략선취정영개체，병대기진행교차화변이조작，사기조전연면수렴이획득 Pareto 최우해。4개측시함수화2개공정실례험증료해산법적유효성。
A micro multi-objective genetic algorithm was proposed herein based on grid domination to solve multi-objective optimization problems and it had good convergence and high computational ef-ficiency.The method combined with the concept of the grid dominance and micro multi-objective ge-netic algorithm.In each generation,the grid value,the grid crowding distance and grid coordinate point distance of every individual were calculated,respectively.Then elite individuals were selected to do crossover and mutation operators based on the grid domination sorting and grid selection strate-gies.The individuals were iterated toward the Pareto front and the Pareto optimal solutions were ob-tained.Finally,the proposed algorithm was verified effectively through four test functions and two practical engineering problems.