CAJ | 학술논문

针对蚁群算法收敛速度慢,容易陷入局部最优解的缺陷,提出了一种基于自然选择策略的改进型蚁群算法,改进后的算法利用自然选择中“优胜劣汰”的进化策略,对每次迭代的随机进化因子大于进化漂变阈值的路径信息素进行二次更新,增强满足进化策略路径上的信息素浓度,以加快算法的收敛速度;而随机进化因子的随机性增强了算法跳出局部最优解的概率.将提出的改进型蚁群算法求解经典的 TSP 问题,并通过实验证明了改进后的蚁群算法在最优解精度和收敛速度等方面均有所提高.
침대의군산법수렴속도만,용역함입국부최우해적결함,제출료일충기우자연선택책략적개진형의군산법,개진후적산법이용자연선택중“우성렬태”적진화책략,대매차질대적수궤진화인자대우진화표변역치적로경신식소진행이차경신,증강만족진화책략로경상적신식소농도,이가쾌산법적수렴속도;이수궤진화인자적수궤성증강료산법도출국부최우해적개솔.장제출적개진형의군산법구해경전적 TSP 문제,병통과실험증명료개진후적의군산법재최우해정도화수렴속도등방면균유소제고.
To solve basic ant colony algorithm’s drawbacks of low convergence rate, easiness of trapping in local optimal solution, an improved ant colony algorithm based on natural selection was proposed. The improved algorithm employed evolution strategy of survival the fittest in natural selection to enhance pheromones in paths whose random evolution factor was bigger than threshold of evolution drift factor in each process of iteration. It could accelerate convergence rate effectively. Besides the introduction of random evolution factor reduced probability of trapping local optimal solution notably. The proposed algorithm was applied to classic TSP problem to find better solution for TSP. Simulation results depict the improved algorithm has better optimal solution and higher convergence rate.