To solve the problem of limited searching ability,easy falling into local optimum,and huge loss of population vari-ety,a novel squirrel optimization algorithm(TRWSSA)is developed,which combines adaptive t-distribution with random walk strategy.For population initialization,the method employs a refraction reverse learning technique,which increases the population's total variety.The chance of the method falling into a local optimum is lowered and the global optimization ability is boosted by incor-porating a non-linear search factor and adding an adaptive t-distribution perturbation site to each squirrel location update.A ran-dom walk approach is introduced to the algorithm's last position update to update the ideal squirrel location,which improves the al-gorithm's convergence accuracy and speed.The experimental findings and analysis reveal that TRWSSA has a considerable improve-ment in convergence speed and accuracy,and it can better tackle the problem of insufficient optimization,when compared to other intelligent algorithms and improved algorithms.
关键词
智能优化算法/松鼠算法/算法改进/融合策略/折射反向学习/自适应t分布/随机游走/基准函数
Key words
intelligent optimization algorithm/squirrel algorithm/algorithm improvement/integration strategy/refraction reverse learning/adaptive t distribution/random walk/benchmark function
维普
万方数据