AETSA: Adaptive Selection and Dynamic Diversity Enhancement of Evolutionary Tree Species Algorithm
Jiayi Liu, Jiarui Fan, Zhili Wu, Yuzhu Ji · 2024
Since the original TSA has problems such as easy to fall into local optimum and limited population diversity, this study proposes an optimized Tree-Seed Algorithm(TSA), which significantly improves the search efficiency and global optimization ability of the algorithm by introducing an adaptive selection mechanism(ST), dynamic step-size adjustment and population diversity enhancement strategy. The adaptive selection mechanism dynamically adjusts the search trend based on the changes of the current iteration number and the global optimum, effectively balancing the local search and global search. In addition, the algorithm further optimizes the balance between exploration and exploitation by dynamically adjusting the step size (d). The population diversity enhancement mechanism is fused by randomly selecting individuals, and this fusion strategy is based on the weighted differential variance of three random individuals, which helps the algorithm to jump out of the local optimal solution. Combining these innovations, the improved tree species algorithm (AETSA) solves the limitation problem of the original TSA. Experimental results show that AETSA outperforms benchmark algorithms such as TSA, EST-TSA, DE, BOA, and GA in terms of accuracy and convergence speed, and the algorithm in this study demonstrates excellent performance and adaptability in solving complex optimization problems.