Teaching-learning guided salp swarm algorithm for global optimization tasks and feature selection

Jun Li

doi:10.1007/s00500-023-09070-3

2023-08-18

PRE

DOI：10.1007/s00500-023-09070-3

原文链接

原文求助

摘要

En 中文

The basic salp swarm algorithm (SSA) is a novel nature-inspired swarm intelligence optimization algorithm based on the foraging behavior of salp individuals in the deep sea. Since its development, the salp swarm algorithm has attracted widespread interest from scholars both at home and abroad for solving complex real-world practical problems. With continuous research, the SSA algorithm has revealed some shortcomings such as slow convergence speed and low accuracy. To enhance the optimization capability of the algorithm, in this paper, we propose an improved hybrid algorithm called TLSSA based on two phases of the teaching-learning-based optimization method: the teaching phase and the learner phase. In the teaching phase, students' ability is improved by updating the difference between the teacher and the class average level, which helps to improve the overall learning ability of the salp population, resulting in higher quality solutions. In the learning phase, by simulating the discussion, statement, and communication between students, the average level of the individual is improved, and the global search speed of the algorithm is accelerated. To verify the effectiveness and competitiveness of the proposed method, it is first tested on 30 IEEE CEC 2017 benchmark functions. The test results demonstrate that the proposed TLSSA method obtains better overall performance compared to 8 mainstream meta-heuristics and 8 advanced algorithms. After that, we applied the proposed method to solve two classical real-world engineering design problems and feature selection. Again, the experimental results show that our method has significant advantages over the traditional methods in solving these practical problems. The remarkable performance of our proposed improved TLSSA algorithm in solving theoretical and practical complex optimization problems also provides potential possibilities for applying more intelligent optimization algorithms to solve complex optimization problems in real-life situations in the future.

Keyword:

Salp swarm algorithm

Teaching-learning-based optimization

Engineering design

Feature selection