Adaptive Tasmanian Devil Optimizer for Global Optimization and Application in Wireless Sensor Network Deployment

The Tasmanian Devil Optimization (TDO) algorithm, a recently popular metaheuristic algorithm, exhibits issues such as slow convergence, low precision, and susceptibility to getting stuck in local optima when applied to practical problems. In order to address these drawbacks, we propose an improved version called the Adaptive Tasmanian Devil Optimizer (ATDO). In the enhancements, we introduce a best point set population initialization to enhance ATDO’s exploration capabilities. The adoption of an adaptive Levy flight strategy improves convergence precision and speed. The introduction of a Brownian random walk strategy effectively prevents the problem of being trapped in local optimal solutions. Additionally, the use of a cosine-type change-based differential evolution algorithm’s crossover strategy enhances the overall population quality. To validate the performance of ATDO, a detailed analysis is conducted in terms of population diversity, exploration-exploitation balance, and convergence behavior. Furthermore, we apply ATDO to CEC-2017 for global optimization performance evaluation. Results from Wilcoxon rank-sum and Friedman average rank tests demonstrate that, compared to TDO and other optimization algorithms, ATDO exhibits superior comprehensive performance. Finally, we apply ATDO to wireless sensor network layout optimization and two engineering optimization problems, experimental results show that ATDO outperforms other comparative algorithms, further validating its scalability and excellence.