Comparing Decomposition-Based and Automatically Component-Wise Designed Multi-Objective Evolutionary Algorithms
Leonardo C. T. Bezerra,Manuel López-Ibáñez,Thomas Stützle +2 more
- 29 Mar 2015
- Vol. 9018, pp 396-410
TL;DR: A systematic analysis that compares algorithms produced using the MOEA/D decomposition-based framework and the AutoMOEA component-wise design framework identifies a version of MOEE/D that outperforms the best known MOEA /D algorithm for several scenarios and confirms the effectiveness of decomposition on problems with three objectives.
read more
Abstract: A main focus of current research on evolutionary multi-objective optimization (EMO) is the study of the effectiveness of EMO algorithms for problems with many objectives. Among the several techniques that have led to the development of more effective algorithms, decomposition and component-wise design have presented particularly good results. But how do they compare? In this work, we conduct a systematic analysis that compares algorithms produced using the MOEA/D decomposition-based framework and the AutoMOEA component-wise design framework. In particular, we identify a version of MOEA/D that outperforms the best known MOEA/D algorithm for several scenarios and confirms the effectiveness of decomposition on problems with three objectives. However, when we consider problems with five objectives, we show that MOEA/D is unable to outperform SMS-EMOA, being often outperformed by it. Conversely, automatically designed AutoMOEAs display competitive performance on three-objective problems, and the best and most robust performance among all algorithms considered for problems with five objectives.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Figures
Fig. 1. Boxplots of the relative hypervolume achieved by MOEA/DDRA-DE using default or tuned parameter settings on selected 3-objective 40-variable WFG problems. 
Table 2. Parameter space for tuning all MOEA/D algorithms. 
Fig. 4. Boxplots of the relative hypervolume achieved by all algorithms on WFG problems with 40 variables and 5 objectives. 
Fig. 3. Boxplots of the relative hypervolume achieved by all algorithms on WFG problems with 40 variables and 3 objectives. 
Table 3. Rank sum analysis depicting overall performance on all scenarios. The best ranked algorithms are shown on top. Algorithms in boldface present rank sums not significantly worse than the best ranked algorithm. Algorithms within the same block are not significantly different, in terms of ranking, to the first algorithm of the same block. 
Table 1. Algorithm components of the AutoMOEAs used in this work. From top to bottom, AutoMOEAD3, AutoMOEAD5, AutoMOEAW3, and AutoMOEAW5.
Citations
A Large-Scale Experimental Evaluation of High-Performing Multi- and Many-Objective Evolutionary Algorithms.
TL;DR: A systematic, comprehensive evaluation of a large number of MOEAs that covers a wide range of experimental scenarios and confirms some of the assumed knowledge in the field, while at the same time providing new insights on the relative performance ofMOEAs for many-objective problems.
Markov Chain methods for the Bipartite Boolean Quadratic Programming Problem
TL;DR: In this article, a new metaheuristic schema called conditional Markov chain search (CMCS) was proposed for the bipartite Boolean quadratic programming problem (BBQP), which is an extension of the well-known Boolean Quadratic Programming Problem (BQP).
37
•Journal Article
Automatic design of evolutionary algorithms for multi-objective combinatorial optimization
TL;DR: This paper investigates the automatic design of MOEAs, extending previous work on other multi-objective metaheuristics, and shows that the automatically designed MOeAs are able to outperform six traditional MOEas, confirming the importance and efficiency of this design methodology.
18
A component-wise approach to multi-objective evolutionary algorithms: From flexible frameworks to automatic design
Leonardo C. T. Bezerra,Thomas Stützle +1 more
- 04 Jul 2016
TL;DR: This thesis empirically demonstrates the efficacy of a flexible and representative algorithmic framework that assembles components originally used by many different MOEAs from the literature, providing a way of seeing algorithms as instantiations of a unified template.
14
•Posted Content
Markov Chain methods for the bipartite Boolean quadratic programming problem
TL;DR: A new metaheuristic schema is designed which is flexible enough to model several standard metaheuristics; this flexibility is controlled by multiple numeric parameters, and so is convenient for automated generation.
13
References
A fast and elitist multiobjective genetic algorithm: NSGA-II
TL;DR: This paper suggests a non-dominated sorting-based MOEA, called NSGA-II (Non-dominated Sorting Genetic Algorithm II), which alleviates all of the above three difficulties, and modify the definition of dominance in order to solve constrained multi-objective problems efficiently.
MOEA/D: A Multiobjective Evolutionary Algorithm Based on Decomposition
Qingfu Zhang,Hui Li +1 more
TL;DR: Experimental results have demonstrated that MOEA/D with simple decomposition methods outperforms or performs similarly to MOGLS and NSGA-II on multiobjective 0-1 knapsack problems and continuous multiobjectives optimization problems.
Comparison of Multiobjective Evolutionary Algorithms: Empirical Results
TL;DR: This paper provides a systematic comparison of various evolutionary approaches to multiobjective optimization using six carefully chosen test functions and shows that elitism is shown to be an important factor for improving evolutionary multiobjectives search.
Comparison of multiobjective evolutionary algorithms: empirical results
Eckart Zitzler,Kalyanmoy Deb,Lothar Thiele +2 more
- 01 Jan 1999
TL;DR: In this article, the authors provide a systematic comparison of various evolutionary approaches to multiobjective optimization using six carefully chosen test functions, each test function involves a particular feature that is known to cause difficulty in the evolutionary optimization process, mainly in converging to the Pareto-optimal front.
3.9K
Multiobjective Optimization Problems With Complicated Pareto Sets, MOEA/D and NSGA-II
Hui Li,Qingfu Zhang +1 more
TL;DR: The experimental results indicate that MOEA/D could significantly outperform NSGA-II on these test instances, and suggests that decomposition based multiobjective evolutionary algorithms are very promising in dealing with complicated PS shapes.
2.4K