模拟退火教学式优化算法

Simulated annealing teaching-learning-based optimization algorithm

针对教学式优化算法在求解组合优化问题时易陷入局部最优问题进行了研究,提出模拟退火教学式优化算法。利用模拟退火方法,在"教"与"学"两个阶段按照模拟退火计算的概率,随机接受个体中某一位较差解作为新解的一部分。通过增加群体多样性的方法,增强教学式优化算法逃离局部最优解的能力。分别对单模、多模和旋转函数进行仿真,并与其他算法进行了对比实验。结果表明,提出的方法在收敛速度和收敛精度上具有较好的性能。

This paper studied the problem that standard teaching-learning-based optimization algorithm（ TLBO） easily converges to local optima when solving combinatorial optimization,and proposed a simulated annealing TLBO（ SATLBO） algorithm. In the method,it used the simulated annealing algorithm. Randomly selected a bit of the bad individuals according to a calculated possibility of simulated annealing algorithm to the new population in the teacher phase and learner phase. It increased the ability of running away from local optima of TLBO by increasing the diversity of the population. It simulated the unmultimodal functions,multimodal functions and rotation functions,and compared the results with some other evolutionary computation algorithms. The results indicate that the improved algorithm has good performance in terms of convergence speed and accuracy.