CAJ | 학술논문

针对现有动态优化算法易陷入局部极值和多样性差等问题，提出了一种动态演化膜算法。依据膜计算理论，所提算法引入膜结构、多重集和反应规则来求解动态优化问题。为了增强在动态环境下的适应能力，所提算法使用了网格策略对搜索空间进行划分，同时设计了4个反应规则来保持算法在动态寻优过程中解的多样性。仿真实验采用标准移动峰测试问题验证算法的求解性能，并分别与3种动态优化算法的求解结果进行比较。仿真结果表明：所提算法提高了搜索过程中解的多样性，且求得的近似最优解更接近于问题的全局最优解，说明所提算法求解动态优化问题是可行的和有效的。
침대현유동태우화산법역함입국부겁치화다양성차등문제，제출료일충동태연화막산법。의거막계산이론，소제산법인입막결구、다중집화반응규칙래구해동태우화문제。위료증강재동태배경하적괄응능력，소제산법사용료망격책략대수색공간진행화분，동시설계료4개반응규칙래보지산법재동태심우과정중해적다양성。방진실험채용표준이동봉측시문제험증산법적구해성능，병분별여3충동태우화산법적구해결과진행비교。방진결과표명：소제산법제고료수색과정중해적다양성，차구득적근사최우해경접근우문제적전국최우해，설명소제산법구해동태우화문제시가행적화유효적。
Considering that the existing dynamic optimization algorithms can easily fall into local minima and have poor diversity, a novel dynamic evolutionary membrane algorithm is proposed. The proposed algorithm introduces three elements of membrane computing, including membrane structure, multiset and reaction rules, to solve dynamic optimization problems. To enhance the adaptive ability of the proposed algorithm under dynamic environments, the al-gorithm employs the grid to divide the search space. Furthermore, the four kinds of reaction rules are introduced to maintain the diversity of solutions found by the algorithm during a dynamic optimization process. In simulation experi-ments, the standard moving peaks benchmark was used to validate the solving performance of the algorithm. Moreo-ver, the performance of the proposed algorithm was compared with three state-of-the-art dynamic optimization algo-rithms. The simulation results indicate that the proposed algorithm improves the diversity of the candidate solutions, and the approximate optimal solution found by the algorithm is closer to the global optimal solution. Therefore, the proposed algorithm is feasible and effective in solving dynamic optimization problems.