A Many-Objective Evolutionary Algorithm With Two Interacting Processes: Cascade Clustering and Reference Point Incremental Learning

Ge, Hongwei; Zhao, Mingde; Sun, Liang; Wang, Zhen; Tan, Guangming; Zhang, Qiang; Chen, C. L. Philip

doi:10.1109/tevc.2018.2874465

Cited by 70 publications

(24 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [8], the algorithm generates a set of reference vectors to decompose the original MOPs into single-objective sub-problems and to elucidate user preferences to target a preferred subset of the whole PF(pareto fronts). In [9], the algorithm uses the reference vectors provided by the process based on incremental learning to select the potential solutions. One of the limitations is how to generate a set of uniformly distributed reference points or vectors.…”

Section: In Addition Moeas Can Easily Solve Complex Optimizationmentioning

confidence: 99%

A Novel Nonlinear Expanded Dominance Relation Based Evolutionary Algorithm for Many-Objective Optimization Problems

Wei

Liu

2021

IEEE Access

View full text Add to dashboard Cite

Multi-objective optimization problems exist widely in scientific research and engineering applications. With the number of objectives increasing, the proportion of non-dominated individuals in the population of many-objective optimization problems increases sharply, resulting in a reduction of convergence pressure of the traditional multi-objective optimization algorithms. In some cases, the optimal solutions may be located in the special regions, such as many discrete regions and the regions with very few feasible solutions. In this case, the existing nonlinear expanded evolutionary algorithm can not find the true Pareto fronts. To address the limitation, a novel nonlinear expanded dominance relation based manyobjective evolutionary algorithm is proposed to handle many-objective optimization problems. Experimental results show that compared with the state of art algorithms, the proposed algorithm is effective for DTLZs, in terms of IGD, PD and GD metrics.

show abstract

Section: In Addition Moeas Can Easily Solve Complex Optimizationmentioning

confidence: 99%

A Novel Nonlinear Expanded Dominance Relation Based Evolutionary Algorithm for Many-Objective Optimization Problems

Wei

Liu

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Associating K reference points with K solutions has a complexity Algorithm 2: Update reference set Input: population (P ), archive (A), last updated reference set (R), population size (N )); Output: Output R, B and P ; 1 Determine the number reference of points to add: K = min ( √ N , |A|); // Add K reference points 2 while |R| − N < K do 3 Compute the furthest memberÂ i of A to P by the max-min distance [32]; 4 AddÂ i to P ; 5 Add into R the projection ofÂ i on the reference plane; // Remove K reference points 6 Calculate reference score ζ(r i ) using Eq. (4) for all r i ∈ R; 7 while |R| > N and there exists a reference score larger than 0 do // pickî (randomly if a tie exists) 8 Identify the reference with the largest score:î = argmax 1≤i≤|R| ζ(r i ); 9 Remove the reference point rî from R and the solution associating with it from P ;…”

Section: Reference Set Selectionmentioning

confidence: 99%

“…Wu et al introduced a Gaussian process regression model to aid reference sampling [37]. Similarly, an incremental learning model was employed for reference adaptation [9]. It should be noted that, however, the use of models for reference setting possibly induces additional computational costs, as there is no free lunch for improvements [36].…”

Section: Introductionmentioning

confidence: 99%

AREA: An adaptive reference-set based evolutionary algorithm for multiobjective optimisation

Jiang

Guo

et al. 2020

Information Sciences

View full text Add to dashboard Cite

Population-based evolutionary algorithms have great potential to handle multiobjective optimisation problems. However, these algorithms depends largely on problem characteristics, and there is a need to improve their performance for a wider range of problems. References, which are often specified by the decision maker's preference in different forms, are a very effective method to improve the performance of algorithms but have not been fully explored in literature. This paper proposes a novel framework for effective use of references to strengthen algorithms. This framework considers references as search targets which can be adjusted based on the information collected during the search. The proposed framework is combined with new strategies, such as reference adaptation and adaptive local mating, to solve different types of problems. The proposed algorithm is compared with state of the arts on a wide range of problems with diverse characteristics. The comparison and extensive sensitivity analysis demonstrate that the proposed algorithm is competitive and robust across different types of problems studied in this paper.

show abstract

“…To address the above challenges, it is desirable to learn the location of reference vectors based on the knowledge acquired in the whole search history instead of based on the solutions in the current generation only. In [27], the active and inactive reference vectors after cascade clustering selection are used to train a classifier using incremental support vector machine to estimate the regions where the reference vectors should be located. In [28], self-organizing map (SOM) is adopted to learn the distribution of weight vectors using the neighbour information in the locations of the obtained solutions.…”

Section: Introductionmentioning

confidence: 99%

An Adaptive Reference Vector-Guided Evolutionary Algorithm Using Growing Neural Gas for Many-Objective Optimization of Irregular Problems

Liu

Jin

Heiderich

et al. 2022

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Most reference vector based decomposition algorithms for solving multi-objective optimization problems may not be well suited for solving problems with irregular Pareto fronts because the distribution of predefined reference vectors may not match well with the distribution of the Pareto optimal solutions. Thus, adaptation of the reference vectors is an intuitive way for decomposition based algorithms to deal with irregular Pareto fronts. However, most existing methods frequently change the reference vectors based on the activeness of the reference vectors within specific generations, slowing down the convergence of the search process. To address this issue, we propose a new method to learn the distribution of the reference vectors using the growing neural gas network to achieve automatic yet stable adaptation. To this end, an improved growing neural gas is designed for learning the topology of the Pareto fronts with the solutions generated during a period of the search process as the training data. We use the individuals in the current population as well as those in previous generations to train the growing neural gas to strike a balance between exploration and exploitation. Comparative studies conducted on popular benchmark problems and a realworld hybrid vehicle controller design problem with complex and irregular Pareto fronts show that the proposed method is very competitive.

show abstract

A Many-Objective Evolutionary Algorithm With Two Interacting Processes: Cascade Clustering and Reference Point Incremental Learning

Cited by 70 publications

References 56 publications

A Novel Nonlinear Expanded Dominance Relation Based Evolutionary Algorithm for Many-Objective Optimization Problems

A Novel Nonlinear Expanded Dominance Relation Based Evolutionary Algorithm for Many-Objective Optimization Problems

AREA: An adaptive reference-set based evolutionary algorithm for multiobjective optimisation

An Adaptive Reference Vector-Guided Evolutionary Algorithm Using Growing Neural Gas for Many-Objective Optimization of Irregular Problems

Contact Info

Product

Resources

About