关键词: 二进制电鳗觅食优化算法, 转换函数, 复合型转换函数, 背包问题

Abstract: The electric eel foraging optimization (EEFO) algorithm is a recently proposed metaheuristic optimization algorithm used to solve continuous optimization problems and has been applied in various engineering problems. However, many real-world optimization problems are discrete, which necessitates a binary version of the algorithm. Researchers typically use transfer functions to convert continuous solutions into discrete ones for solving discrete optimization problems. Traditional S-shaped transfer functions tend to diverge and struggle with convergence, while V-shaped transfer functions are prone to getting trapped in local optima and are difficult to escape. To address these challenges, this paper designs an adaptive V-shaped transfer function and combines it with an S-shaped transfer function using the electric eel energy factor, proposing an adaptive compound transfer function to binarize the electric eel algorithm. However, due to the lack of local diversity during the resting and hunting phases, and premature convergence during the interaction and migration phases of the electric eel algorithm, further improvements are made. A weight control factor is added during the interaction phase to exploit the divergent properties of the S-shaped transfer function, enhancing global search capabilities. A boosting factor is introduced in the migration phase to regulate the behavior of the electric eel, preventing premature convergence into local optima. Additionally, random factors are incorporated during the resting and hunting phases to increase local diversity. Finally, experiments conducted on 35 knapsack problem instances, including convergence, mean performance, and ablation analysis, confirm the effectiveness of the proposed binary electric eel foraging optimization algorithm.

Key words: binary electric eel foraging optimization （EEFO）algorithm , transfer function, compound transfer function, knapsack problem