MEIAO: A Multi-Strategy Enhanced Information Acquisition Optimizer for Global Optimization and UAV Path Planning.

With the expansion of unmanned aerial vehicles (UAVs) into complex three-dimensional (3D) terrains for reconnaissance, rescue, and related missions, traditional path planning methods struggle to meet multi-constraint and multi-objective requirements. Existing swarm intelligence algorithms, limited by the "no free lunch" theorem, also face challenges when the standard Information Acquisition Optimizer (IAO) is applied to such tasks, including low exploration efficiency in high-dimensional search spaces, rapid loss of population diversity, and improper boundary handling. To address these issues, this study proposes a Multi-Strategy Enhanced Information Acquisition Optimizer (MEIAO). First, a Levy Flight-based information collection strategy is introduced to leverage its combination of short-range local searches and long-distance jumps, thereby broadening global exploration. Second, an adaptive differential evolution operator is designed to dynamically balance exploration and exploitation via a variable mutation factor, while crossover and greedy selection mechanisms help maintain population diversity. Third, a globally guided boundary handling strategy adjusts out-of-bound dimensions to feasible regions, preventing the generation of low-quality paths. Performance was evaluated on the CEC2017 (dim = 30/50/100) and CEC2022 (dim = 10/20) benchmark suites by comparing MEIAO with eight algorithms, including VPPSO and IAO. Based on the mean, standard deviation, Friedman mean rank, and Wilcoxon rank-sum tests, MEIAO demonstrated superior performance in local exploitation of unimodal functions, global exploration of multimodal functions, and complex adaptation on composite functions while exhibiting stronger robustness. Finally, MEIAO was applied to 3D mountainous UAV path planning, where a cost model considering path length, altitude standard deviation, and turning smoothness was established. The experimental results show that MEIAO achieved an average path cost of 253.9190, a 25.7% reduction compared to IAO (341.9324), with the lowest standard deviation (60.6960) among all algorithms. The generated paths were smoother, collision-free, and achieved faster convergence, offering an efficient and reliable solution for UAV operations in complex environments.