Hybrid Strategies Based Seagull Optimization Algorithm for Solving Engineering Design Problems

Pingjing Hou,Jiang Liu,Feng Ni,Leyi Zhang

DOI: https://doi.org/10.1007/s44196-024-00439-2

IF: 2.259

2024-03-27

International Journal of Computational Intelligence Systems

Abstract:Abstract The seagull optimization algorithm (SOA) is a meta-heuristic algorithm proposed in 2019. It has the advantages of structural simplicity, few parameters and easy implementation. However, it also has some defects including the three main drawbacks of slow convergence speed, simple search method and poor ability of balancing global exploration and local exploitation. Besides, most of the improved SOA algorithms in the literature have not considered the drawbacks of the SOA comprehensively enough. This paper proposes a hybrid strategies based algorithm (ISOA) to overcome the three main drawbacks of the SOA. Firstly, a hyperbolic tangent function is used to adjust the spiral radius. The spiral radius can change dynamically with the iteration of the algorithm, so that the algorithm can converge quickly. Secondly, an adaptive weight factor improves the position updating method by adjusting the proportion of the best individual to balance the global and local search abilities. Finally, to overcome the single search mode, an improved chaotic local search strategy is introduced for secondary search. A comprehensive comparison between the ISOA and other related algorithms is presented, considering twelve test functions and four engineering design problems. The comparison results indicate that the ISOA has an outstanding performance and a significant advantage in solving engineering problems, especially with an average improvement of 14.67% in solving welded beam design problem.

computer science, artificial intelligence, interdisciplinary applications

What problem does this paper attempt to address?

### Problems the paper attempts to solve

This paper aims to solve three main defects of the Seagull Optimization Algorithm (SOA) when solving engineering design problems:

1. **Slow convergence speed**: In the iterative process of the standard SOA algorithm, due to the fixed value of the spiral radius \( r \), the search range is too large, so it is prone to oscillation when approaching the optimal solution and cannot converge quickly.
2. **Poor balance between global exploration and local exploitation capabilities**: In the standard SOA, the influence weight of the current optimal individual remains unchanged throughout the iterative process, which causes the algorithm to lack global exploration ability in the early stage and is prone to fall into local optima in the later stage.
3. **Single search pattern**: The standard SOA has only one position update method, which reduces the search diversity and is especially prone to fall into local optima in multi - peak function optimization.

To overcome these defects, the paper proposes an improved Seagull Optimization Algorithm (ISOA) based on a hybrid strategy. Specific improvement measures include:

1. **Using the hyperbolic tangent function to adjust the spiral radius**: By introducing the hyperbolic tangent function (Tanh) to dynamically adjust the spiral radius \( r \), the algorithm has a larger search range in the early stage to enhance global exploration ability, and gradually reduces the search range in the later stage to accelerate convergence.
2. **Introducing an adaptive weight factor**: Adjust the position update method through the adaptive weight factor \( \omega(t) \) so that the algorithm pays more attention to global exploration in the early stage and more attention to local exploitation in the later stage.
3. **Introducing a chaotic local search strategy**: Increase the search diversity through the chaotic local search strategy and prevent the algorithm from falling into local optima.

Through comparative experiments with the standard SOA and other related optimization algorithms, the paper verifies the superior performance of ISOA in solving engineering design problems, especially improving the solution effect by an average of 14.67% in the welded beam design problem.