CAJ | 학술논문

针对混洗蛙跳算法在求解连续函数优化问题中出现的收敛速度慢、求解精度低的缺点, 提出了一种基于反向学习策略的改进算法, 在种群初始化和进化过程中分别加入反向操作, 产生更靠近优质解的种群, 从而提高了算法的全局寻优能力, 促进了算法收敛。实验仿真表明, 新算法在寻优效率、计算精度等方面均优于原算法。</sup>
침대혼세와도산법재구해련속함수우화문제중출현적수렴속도만、구해정도저적결점, 제출료일충기우반향학습책략적개진산법, 재충군초시화화진화과정중분별가입반향조작, 산생경고근우질해적충군, 종이제고료산법적전국심우능력, 촉진료산법수렴。실험방진표명, 신산법재심우효솔、계산정도등방면균우우원산법。</sup>
Classical shuffled frog leaping algorithm is slow in convergence, and has a low convergent precision to address continuous function optimization problems. To overcome such shortages, this paper presented an improved shuffled frog leaping algorithm which combined the OBL strategy. The proposed approach employed OBL for population initialization and generation jumping to produce populations closer to high-quality solutions. The experiments carried on classic benchmark functions show that it performs significantly better both in terms of convergence speed and solution precision.