Exploring the Applicability of Chaotic Maps in Artificial Hummingbird Algorithm: From Algorithm Construction to Case Verification

A research gap exists concerning how different chaotic mappings influence the applicability of metaheuristic algorithms, along with inherent limitations of the traditional artificial hummingbird algorithm (AHA). Specifically, blind spots in population coverage and vulnerability to local optima stemming from random initialization. To address these issues, this study proposes a suite of chaotic artificial hummingbird algorithms (CAHAs). By integrating multidimensional analysis with engineering practice, the research systematically explores the performance, applicability boundaries, and practical application value of the proposed CAHA. First, 10 chaotic maps with distinct characteristics are incorporated into AHA, resulting in 10 corresponding CAHA variants. Leveraging the pseudorandom and ergodic properties of chaotic maps, the spatial distribution of the initial population is optimized, thereby mitigating the risk of local optima at the source. Concurrently, the “butterfly effect” inherent to chaotic mapping is integrated throughout the entire iterative process of the algorithm, establishing a full‐chain enhanced optimization mechanism that spans initialization, iteration, and memory phases. Second is to overcome the limitations of traditional research, which relies solely on numerical results from test functions to assess algorithm performance. This study systematically categorizes the types of real‐world optimization tasks represented by 23 classic test functions. Based on these task categories, a set of multidimensional evaluation metrics is designed to conduct quantitative analyses of the applicability of the 10 CAHA variants. To comprehensively evaluate CAHA performance, a parameter sensitivity analysis is also performed to clarify how key parameters affect the algorithm’s optimization effectiveness. Finally, to verify the practical application capability of CAHA, the 10 CAHA variants are applied to four typical engineering optimization scenarios, with application‐specific analyses conducted in accordance with the unique constraints of each scenario. The results demonstrate that integrating chaotic mapping significantly enhances AHA’s ability to solve complex engineering problems, while also revealing clear applicability boundaries for CAHA variants corresponding to different chaotic maps.