Automated design of local search algorithms: Predicting algorithmic components with LSTM

Meng, Weiyao; Qu, Rong

Abstract

With a recently defined AutoGCOP framework, the design of local search algorithms has been defined as the composition of elementary algorithmic components. The effective compositions of the best algorithms thus retain useful knowledge of effective algorithm design. This paper investigates machine learning to learn and extract useful knowledge in effective algorithmic compositions. The process of forecasting algorithmic components in the design of effective local search algorithms is defined as a sequence classification task, and solved by a long short-term memory (LSTM) neural network to systematically analyse algorithmic compositions. Compared with other learning models, the results reveal the superior prediction performance of the proposed LSTM. Further analysis identifies some key features of algorithmic compositions and confirms their effectiveness for improving the prediction, thus supporting effective automated algorithm design.

Citation

Meng, W., & Qu, R. (2024). Automated design of local search algorithms: Predicting algorithmic components with LSTM. Expert Systems with Applications, 237(Part A), Article 121431. https://doi.org/10.1016/j.eswa.2023.121431

• Journal Article Type: Article
• Acceptance Date: Sep 1, 2023
• Online Publication Date: Sep 6, 2023
• Publication Date: Mar 1, 2024
• Deposit Date: Sep 11, 2023
• Publicly Available Date: Sep 7, 2024
• Journal: Expert Systems with Applications
• Print ISSN: 0957-4174
• Electronic ISSN: 0957-4174
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 237
• Issue: Part A
• Article Number: 121431
• DOI: https://doi.org/10.1016/j.eswa.2023.121431
• Keywords: Automated algorithm design; Search algorithms; LSTM; Imbalanced dataset; Data re-sampling
• Public URL: https://nottingham-repository.worktribe.com/output/25083827
• Publisher URL: https://www.sciencedirect.com/science/article/abs/pii/S0957417423019334?via%3Dihub