MultiAO: A Novel Optimizer Inspired by Collaborative Intelligence in Complex Systems for Delivery Path Planning in UAV Clusters

Within the complex domain of cluster optimization, high-dimensional decision-making environments have persistently confronted formidable optimization bottlenecks. To address these challenges, this paper proposes a novel multi-agent optimization algorithm predicated upon a state-dependent dynamic evolution mechanism, designated as the Multi-Agent Optimizer (MultiAO). The primary contributions of this work are threefold. First, we introduce an innovative algorithmic framework characterized by three distinct evolutionary phases: gravitational accretion and centroid rebellion, orthogonal singularity decoupling, and a hysteresis effect coupled with a memory potential field, which accurately capture agent state changes across different search stages. Second, we provide rigorous empirical validation across two dimensions of the CEC2022 benchmark suite. MultiAO ranked first in comparative analyses against five highly-cited and five recently proposed algorithms, yielding average Friedman rankings of 2 and 2.25. Finally, we successfully deploy the algorithm to solve complex real-world engineering challenges, specifically unmanned aerial vehicle (UAV) path planning across urban, mountainous, and composite terrains. In both single-UAV and multi-UAV contexts, MultiAO consistently exhibited optimal convergence, achieving a minimum optimal value as low as 324.332 in multi-UAV scenarios. These contributions collectively demonstrate the superior cluster optimization capabilities of MultiAO, establishing it as a highly competitive metaheuristic algorithm and providing a significantly more robust application paradigm for the field.