An Enhanced Polar Lights Optimization Algorithm with Symmetry Mechanisms for Global Optimization and Its Application to Wind Power Forecasting
Open Access
Abstract
Optimization problems in high-dimensional, nonlinear, and multimodal scenarios remain challenging for conventional methods. Although metaheuristic algorithms have shown strong adaptability, many still suffer from premature convergence, inefficient boundary handling, insufficient population diversity, and a lack of effective symmetry exploitation in the search space. To address these limitations, this paper proposes an Enhanced Polar Lights Optimization (EPLO) algorithm, which extends the recently developed Polar Lights Optimization (PLO) method by incorporating symmetry-aware optimization mechanisms. EPLO integrates three complementary strategies—an elite-guided strategy, a global-best-informed boundary control strategy, and a hybrid crossover strategy—forming a unified “Guidance–Constraint–Innovation” optimization mechanism that leverages symmetry to balance exploration and exploitation. Specifically, elite guidance improves search directionality and convergence speed, symmetry-informed boundary control enhances solution validity and stability, and hybrid crossover preserves population diversity by maintaining symmetric and diverse solution distributions. The performance of EPLO is comprehensively evaluated on the CEC2017 and CEC2022 benchmark suites and compared with nine representative metaheuristic algorithms. Statistical analyses using Wilcoxon rank-sum and Friedman tests confirm the superior convergence speed, robustness, and solution quality of EPLO. Furthermore, to demonstrate its practical applicability, EPLO is combined with a BP neural network and applied to a wind power forecasting task, where symmetry-enhanced optimization contributes to improved predictive accuracy. The results validate EPLO as an effective and robust optimization algorithm with strong potential for both benchmark problems and real-world applications.Keywords
Funding Information
- General Program of the Natural Science Foundation of Jiangxi Province (20252BAC240223)
- General Project of Science and Technology Research Project of Jiangxi Provincial Department of Education (GJJ2409107)
- Key Project of Science and Technology Research Project of Jiangxi Provincial Department of Education (GJJ2508904)
- Jiujiang Natural Science Foundation (S2024KXJJ0001)
This publication has 42 references indexed in Scilit:
- Kirchhoff’s law algorithm (KLA): a novel physics-inspired non-parametric metaheuristic algorithm for optimization problemsArtificial Intelligence Review, 2025
- Efficient multi-UAV path planning in dynamic and complex environments using hybrid polar lights optimizationJournal of King Saud University Computer and Information Sciences, 2025
- Augmented Harris hawks optimization and for engineering design problems and UAV path planningInternational Journal of Machine Learning and Cybernetics, 2025
- Fractional order calculus enhanced dung beetle optimizer for function global optimization and multilevel threshold medical image segmentationThe Journal of Supercomputing, 2024
- Secretary bird optimization algorithm: a new metaheuristic for solving global optimization problemsArtificial Intelligence Review, 2024
- Dung beetle optimizer: a new meta-heuristic algorithm for global optimizationThe Journal of Supercomputing, 2022
- Safety-enhanced UAV path planning with spherical vector-based particle swarm optimizationApplied Soft Computing, 2021
- A modification of I-SOS: performance analysis to large scale functionsApplied Intelligence, 2021
- The Whale Optimization AlgorithmAdvances in Engineering Software, 2016
- Grey Wolf OptimizerAdvances in Engineering Software, 2014