Version 1
: Received: 30 August 2024 / Approved: 30 August 2024 / Online: 30 August 2024 (15:54:27 CEST)
How to cite:
Xie, D. Self-adaptive Differential Evolution for Multi-objective Optimization with Local Search and Indicator-based Selection. Preprints2024, 2024082273. https://doi.org/10.20944/preprints202408.2273.v1
Xie, D. Self-adaptive Differential Evolution for Multi-objective Optimization with Local Search and Indicator-based Selection. Preprints 2024, 2024082273. https://doi.org/10.20944/preprints202408.2273.v1
Xie, D. Self-adaptive Differential Evolution for Multi-objective Optimization with Local Search and Indicator-based Selection. Preprints2024, 2024082273. https://doi.org/10.20944/preprints202408.2273.v1
APA Style
Xie, D. (2024). Self-adaptive Differential Evolution for Multi-objective Optimization with Local Search and Indicator-based Selection. Preprints. https://doi.org/10.20944/preprints202408.2273.v1
Chicago/Turabian Style
Xie, D. 2024 "Self-adaptive Differential Evolution for Multi-objective Optimization with Local Search and Indicator-based Selection" Preprints. https://doi.org/10.20944/preprints202408.2273.v1
Abstract
Differential Evolution (DE) is an evolutionary algorithm that has become increasingly popular due to its effectiveness and efficiency. However, its adaptation for multi-objective optimization requires further investigation. To address this issue, a novel algorithm called ILSDEMO has been proposed, which incorporates indicator-based selection and local search into a self-adaptive DE. ILSDEMO employs an archive population to store non-dominated solutions, generating an initial population uniformly distributed over the feasible solution space using orthogonal design. Additionally, two variants of DE are used to expedite convergence, while q-Gaussian mutation is utilized to exploit better trial individuals. The k-nearest neighbor rule is used to eliminate crowded solutions, and indicator-based selection is employed to generate a new parent population without diversity preservation. The performance of ILSDEMO was investigated on the test instances from the ZDT series and DTLZ series in terms of the selected indicators. The results suggest that ILSDEMO accurately and evenly approximates the true Pareto front compared to NSGAII, IBEA, and DEMO.
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.