Hybrid Strategy improves Crayfish Optimization Algorithm
Abstract
In response to the deficiencies of the Crayfish Optimization Algorithm (COA), such as slow convergence speed and susceptibility to local optima when solving complex optimization problems, an improved Crayfish Optimization Algorithm (ICOA) with a hybrid strategy is proposed. This improvement incorporates five strategies: an enhanced position update function for followers from the Sparrow Algorithm, a refined spiral position update function from the Whale Optimization Algorithm, modifications to the criteria and methods for assessing food size, the golden ratio coefficient, and an information accumulation function. Optimization tests conducted on 12 benchmark functions demonstrate that the improved algorithm shows enhancements in both convergence speed and optimization performance. Finally, experiments applying the improved optimization algorithm to engineering application problems further validate the superiority of the enhanced Crayfish Optimization Algorithm.
Keywords
Download Options
Introduction
Traditional optimization algorithms usually target structured problems, which have relatively clear descriptions in terms of the problems and conditions, such as linear programming [1], integer programming [2], mixed programming [3], and constrained or unconstrained conditions [4]. These algorithms possess clear structural information, stable structures and parameters, and unique and clear global optima. Traditional optimization algorithms can be theoretically analyzed in terms of computational complexity and convergence. While they are generally effective for single-extremum problems, they still exhibit many deficiencies when solving multi-extremum problems. In recent years, the increasing complexity of research problems and the ambiguity of computational results have rendered traditional mathematical models increasingly insufficient for these types of problems, creating a growing demand for superior optimization algorithms.
Conclusion
The improved crayfish optimization algorithm proposed in this paper balances its exploration and exploitation capabilities by filtering participants through sorting and controlling temperature, allowing ICOA to enter different phases. The cooling phase is the exploration phase of ICOA, while the participant and foraging phases represent the development stage of ICOA. The distinct characteristic of this algorithm lies in the exploration and development process based on participant updates and temperature control. During the exploration phase, when a cave is removed, the crayfish enter the cave to escape from high temperatures. The development phase is divided into the participant phase and the foraging phase. In the participant phase, ICOA further splits based on sorting, adopting different cruising strategies. In the foraging phase, ICOA adjusts food size criteria to decide whether to break it up, while foraging is controlled by the amount of food available, allowing for different strategies.
The improved crayfish optimization algorithm proposed in this paper balances its exploration and exploitation capabilities by filtering participants through sorting and controlling temperature, allowing ICOA to enter different phases. The cooling phase is the exploration phase of ICOA, while the participant and foraging phases represent the development stage of ICOA. The distinct characteristic of this algorithm lies in the exploration and development process based on participant updates and temperature control. During the exploration phase, when a cave is removed, the crayfish enter the cave to escape from high temperatures. The development phase is divided into the participant phase and the foraging phase. In the participant phase, ICOA further splits based on sorting, adopting different cruising strategies. In the foraging phase, ICOA adjusts food size criteria to decide whether to break it up, while foraging is controlled by the amount of food available, allowing for different strategies.
This study has validated the effectiveness of ICOA in solving optimization problems using 12 standard benchmark functions and engineering problems. In future work, we will further enhance the exploration capability of ICOA and apply this algorithm to solve practical engineering problems such as three-dimensional path planning for drones, feature selection, image processing, and multi-threshold image segmentation.