CTSA : Tree Seed Algorithm Improved with Cyclic Mutation Spiral Mixing Mechanism and New Crossover Strategy
Yingying Dong, Zhenya Wang · 2024
Based on the limitations of the Tree Seed Algorithm (TSA), such as its proneness to getting trapped in local optima and its relatively slow convergence speed, this study proposes an improved TSA optimization algorithm - the Tree Seed Algorithm improved by the periodic mutation spiral hybrid mechanism and a new crossover strategy (CTSA). The CTSA incorporates periodic mutation and helical mixing mechanisms, which significantly enhance the algorithm's local search capabilities and its ability to escape local optima, thus mitigating TSA's tendency to converge prematurely to suboptimal solutions. Additionally, we introduce a novel inter-species crossover strategy that markedly improves the algorithm's global search efficiency. Through comparative experiments, the CTSA has demonstrated superior optimization capabilities and faster convergence speeds compared to several benchmark algorithms, including Grey Wolf Optimizer (GWO), Adaptive Tree Seed Algorithm (ATSA), Modified Tree Seed Algorithm (MTSA), Particle Swarm Optimization (PSO), and Biogeography-Based Optimization (BOA).