CAJ | 학술논문

针对现有算法中初始种群随机性强、局部搜索能力差、移动公式效率低等问题，提出了一种改进的类电磁机制算法。结合反向学习理论，引入带扰动因子的反向学习机制构造初始种群；提出了一种双混沌优化机制用于局部搜索；运用改进后的公式计算粒子之间的合力；设计了一种自适应移动算子来更新粒子。实验结果表明，改进后的算法具有更好的收敛效果和更高的求解精度。
침대현유산법중초시충군수궤성강、국부수색능력차、이동공식효솔저등문제，제출료일충개진적류전자궤제산법。결합반향학습이론，인입대우동인자적반향학습궤제구조초시충군；제출료일충쌍혼돈우화궤제용우국부수색；운용개진후적공식계산입자지간적합력；설계료일충자괄응이동산자래경신입자。실험결과표명，개진후적산법구유경호적수렴효과화경고적구해정도。
An improved Electromagnetism-like mechanism algorithm is proposed to overcome the drawbacks of the original EM algorithm,such as strong randomness of the initial population,low ability for local search and low efficiency in the movement according to the total force.The new algorithm generates the initial population with the disturbance factor and opposite learning mechanism,improves the local search algorithm with the double chaotic search method,and calculates the total force between particles with the modified equation.Besides,the algorithm is used to design an adaptive move operator for updating the locations of those particles. Experimental results indicate that the proposed algorithm has a better convergence effect and a higher solution accuracy.