EAPO: A Multi-Strategy-Enhanced Artificial Protozoa Optimizer and Its Application to 3D UAV Path Planning

Three-dimensional unmanned aerial vehicle (UAV) path planning presents a challenging optimization problem characterized by high dimensionality, strong nonlinearity, and multiple constraints. To address these complexities, this study proposes an Enhanced Protozoan Optimizer (EAPO) by refining the initialization, behavioral decision-making, environmental perception, and population diversity preservation mechanisms of the original Protozoan Optimizer. Specifically: Latin hypercube sampling enriches initial population diversity; a behavior adaptation mechanism based on historical success dynamically adjusts the exploration-exploitation balance; environmental structure modeling using perception fields enhances local exploitation capabilities; an adaptive hibernation-reconstruction strategy boosts global escape ability. Ablation experiment validates the effectiveness of each enhancement module, while exploration-exploitation analysis demonstrates EAPO maintains an optimal balance throughout the optimization process. Comprehensive evaluations using CEC2022 and CEC2020 benchmark datasets, ten real-world engineering design problems, and four drone path planning scenarios of varying scales and complexities further validate its excellent performance. Experimental results demonstrate that EAPO outperforms the baseline APO and twelve advanced optimizers in convergence accuracy, stability, and robustness. In UAV path planning applications, paths generated by EAPO satisfy all constraints and outperform APO-generated paths across multiple path quality evaluation metrics concerning safety, smoothness, and energy consumption. Compared to APO, EAPO achieved average fitness improvements of 14.0%, 4.5%, 8.7%, and 31.42% across the four maps, respectively, fully demonstrating its practical value and formidable capability in tackling complex engineering optimization problems.