PSTSA: A Phase-Driven Exploration and Multi-Strategy Nelder-Mead Enhanced Tree Species Optimization Algorithm
Chenxi Li, Zhengwei Li, Xinzhe Li, Jiayi Liu, Zhenhao Yu · 2024
To address the limitations of traditional tree seeding algorithm (TSA), A Phase-Driven Exploration and Multi-Strategy Nelder-Mead Enhanced Tree Species Optimization Algorithm (PSTSA) is proposed.PSTSA enhances the diversity of population initialization by introducing Tent chaotic mapping, which solves the problem of uneven distribution of population caused by random initialization in TSA. In addition, the proposed dynamic phase switching mechanism effectively balances global exploration and local exploitation, enhancing the adaptability and optimization efficiency of the algorithm in different phases. In the local search phase, PSTSA incorporates an improved multi-strategy Nelder-Mead optimization method, which enhances the local exploitation capability and avoids the local optimum problem. These innovative improvements effectively solve the problems of premature convergence and local optimum sensitivity of TSA. Experimental results show that PSTSA outperforms other benchmark algorithms such as TSA, EST - TSA, PSO, GWO, SWO, and COA in terms of searchability, accuracy, and convergence speed, demonstrating its excellent performance and potential for wide application in solving complex optimization problems.