.png){kind=icon}
.png){kind=icon}
Return
Evolutionary Dynamic Multiobjective Optimisation Assisted by Inverse Regression Tree Predictor
.png){kind=icon}
OA
AI
DOI：10.1049/cit2.70099.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Dynamic multiobjective optimisation problems (DMOPs) are optimisation problems with multiple conflicting objectives that can change over time. Most dynamic multiobjective optimisation evolutionary algorithms (DMOEAs) attempt to estimate Pareto-optimal sets (PS) directly in the decision space. However, the objectives of the obtained solutions may not meet the decision-maker's expectations. The inverse model can obtain a high-quality population by mapping the desired objectives back to the decision space. However, the existing ones suffer from the difficulty of the nonlinear relationship between objectives and decision variables. In addition, the construction of the existing inverse model uses less historical information, which means the algorithm has limited ability to capture the dynamic nature. This paper proposes a DMOEA based on an inverse regression tree (IRT), called IRT-MOEA/D. A predictor is constructed by regression training of the mapping from objective space to decision space. The expected solutions in the objective space are sampled by the difference method, and a high-quality initial population can be predicted by the trained predictor. In order to demonstrate the performance of the proposed algorithm, six state-of-the-art DMOEAs are designed to compare on the well-known comprehensive benchmark DF series test suits. The statistical results indicate that IRT-MOEA/D is promising for addressing complex DMOPs.
Keywords:
dynamic multiobjective optimisation problem (DMOP)
inverse model
multiobjective evolutionary algorithm (MOEA)
objective space
prediction
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
7.3
Papers:
649
Citations:
2.4K