The fishing cat's unique hunting strategies, including ambush, detection, diving and trapping, inspired the development of a novel metaheuristic optimization algorithm named the Fishing Cat Optimizer (FCO). The purpose of this paper is to introduce FCO, offering a fresh perspective on metaheuristic optimization and demonstrating its potential for solving complex problems.
The FCO algorithm structures the optimization process into four distinct phases. Each phase incorporates a tailored search strategy to enrich the diversity of the search population and attain an optimal balance between extensive global exploration and focused local exploitation.
To assess the efficacy of the FCO algorithm, we conducted a comparative analysis with state-of-the-art algorithms, including COA, WOA, HHO, SMA, DO and ARO, using a test suite comprising 75 benchmark functions. The findings indicate that the FCO algorithm achieved optimal results on 88% of the test functions, whereas the SMA algorithm, which ranked second, excelled on only 21% of the functions. Furthermore, FCO secured an average ranking of 1.2 across the four benchmark sets of CEC2005, CEC2017, CEC2019 and CEC2022, demonstrating its superior convergence capability and robustness compared to other comparable algorithms.
Although the FCO algorithm performs excellently in solving single-objective optimization problems and constrained optimization problems, it also has some shortcomings and defects. First, the structure of the FCO algorithm is relatively complex and there are many parameters. The value of parameters has a certain impact on solving optimization problems. Second, the computational complexity of the FCO algorithm is relatively high. When solving high-dimensional optimization problems, it takes more time than algorithms such as GWO and WOA. Third, although the FCO algorithm performs excellently in solving multimodal functions, it rarely obtains the theoretical optimal solution when solving combinatorial optimization problems.
The FCO algorithm is applied to the solution process of five common engineering design optimization problems.
This paper innovatively proposes the FCO algorithm, which mimics the unique hunting mechanisms of fishing cats, including strategies such as lurking, perceiving, rapid diving and precise trapping. These mechanisms are abstracted into four closely connected iterative stages, corresponding to extensive and in-depth exploration, multi-dimensional fine detection, rapid and precise developmental search and localized refinement and contraction search. This enables efficient global optimization and local fine-tuning in complex environments, significantly enhancing the algorithm's adaptability and search efficiency.
