Abstract
Particle Swarm Optimization (PSO) is an effective and efficient metaheuristic algorithm that has been widely applied to valuable real-world applications. Since PSO was proposed nearly three decades, many PSO-based variants have been developed. In recent years, research on ensemble complementary operators and variants aimed at enhancing the performance of PSO has maintained a certain level of enthusiasm. In this work, an ensemble velocity learning strategy is proposed and combined with two local search methods: the Improved Sine Cosine Algorithm (ISCA) and Sequential Quadratic Programming (SQP). Meanwhile, to improve the effectiveness of the ensemble velocity learning strategy, elite learning, modified inertia weight, modified acceleration coefficients, and mutation strategy are incorporated. The proposed metaheuristic algorithm, named the Ensemble Velocity Learning Strategy for Particle Swarm Optimization Integrating Multiple Local Search Mechanisms (EVLS-PSOIMLSM), is tested on the Institute of Electrical and Electronics Engineers (IEEE) Congress on Evolutionary Computation 2017 (CEC2017) benchmark functions and applied to real-world problems, including the Gear Train Design (GTD) and the Planetary Gear Train Design (PGTD) problem. Compared to 14 recently proposed PSO-based variants and 11 non-PSO algorithms, the experimental results demonstrate that EVLS-PSOIMLSM achieves superior performance. The source code of EVLS-PSOIMLSM is provided at https://github.com/microhard1999/CODES.
Original language | English |
---|---|
Article number | 111117 |
Journal | Engineering Applications of Artificial Intelligence |
Volume | 156 |
DOIs | |
Publication status | Published - 15 Sept 2025 |
Keywords
- Acceleration coefficients
- Elite learning
- Ensemble velocity strategy
- Inertia weight
- Local Search Mechanism
- Particle Swarm Optimization
ASJC Scopus subject areas
- Control and Systems Engineering
- Artificial Intelligence
- Electrical and Electronic Engineering