Tools
Tools
Yazdani, D, Nguyen, TT and Branke, J (2018) Robust optimization over time by learning problem space characteristics. IEEE Transactions on Evolutionary Computation, 23 (1). pp. 143-155. ISSN 1089-778X
|
Text
manuscript_and_supplement.pdf - Accepted Version Download (4MB) | Preview |
Abstract
Robust optimization over time is a new way to tackle dynamic optimization problems where the goal is to find solutions that remain acceptable over an extended period of time. The state-of-the-art methods in this domain try to identify robust solutions based on their future predicted fitness values. However, predicting future fitness values is difficult and error prone. In this paper, we propose a new framework based on a multi-population method in which sub-populations are responsible for tracking peaks and also gathering characteristic information about them. When the quality of the current robust solution falls below the acceptance threshold, the algorithm chooses the next robust solution based on the collected information. We propose four different strategies to select the next solution. The experimental results on benchmark problems show that our newly proposed methods perform significantly better than existing algorithms.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | 0801 Artificial Intelligence And Image Processing, 0906 Electrical And Electronic Engineering, 0806 Information Systems |
| Subjects: | T Technology > T Technology (General) |
| Divisions: | Maritime & Mechanical Engineering (merged with Engineering 10 Aug 20) |
| Publisher: | IEEE |
| Date Deposited: | 16 May 2018 10:12 |
| Last Modified: | 04 Sep 2021 10:30 |
| DOI or ID number: | 10.1109/TEVC.2018.2843566 |
| URI: | https://researchonline.ljmu.ac.uk/id/eprint/8676 |
 |
View Item |