An evolutionary algorithm for constrained multi-objective optimization

Jiménez, Fernando; Gómez‐Skarmeta, Antonio F.; Sánchez, Gracia; Deb, Kalyanmoy

doi:10.1109/cec.2002.1004402

Cited by 60 publications

(42 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper we apply a wrapper feature selection mechanism via the adaptation of a multi-objective evolutionary algorithm known as ENORA [9], [10], and we compare its performance against the classical multi-objective evolutionary algorithm NSGA-II [11] along two directives: (i) performance of the multi-objective search strategy via measuring the hypervolume, and (ii) quality of the classifiers that have been built over the selected features. The goal is to build a classifier for the session data of a multi-skill contact center, which allows us to predict the outcome of a communication based on a limited set of features.…”

Section: Introductionmentioning

confidence: 99%

Attribute Selection Via Multi-Objective Evolutionary Computation Applied to Multi-Skill Contact Center Data Classification

Jiménez

Marzano²,

Sánchez

et al. 2015

2015 IEEE Symposium Series on Computational Intelligence

View full text Add to dashboard Cite

Abstract-Attribute or feature selection is one of the basic strategies to improve the performances of data classification tasks, and, at the same time to reduce the complexity of classifiers, and it is a particularly fundamental one when the number of attributes is relatively high. Evolutionary computation has already proven itself to be a very effective choice to consistently reduce the number of attributes towards a better classification rate and a simpler semantic interpretation of the inferred classifiers. We propose the application of the multi-objective evolutionary algorithm ENORA to the task of feature selection for multi-class classification of data extracted from an integrated multi-channel multi-skill contact center, which include technical, service and central data for each session. Additionally, we propose a methodology to integrate feature selection for classification, model evaluation, and decision making to choose the most satisfactory model according to a a posteriori process in a multi-objective context. We check out our results by comparing the performance and the classification rate against the well-known multi-objective evolutionary algorithm NSGA-II. Finally, the best obtained solution is validated by a data expert's semantic interpretation of the classifier.

show abstract

Section: Introductionmentioning

confidence: 99%

Attribute Selection Via Multi-Objective Evolutionary Computation Applied to Multi-Skill Contact Center Data Classification

Jiménez

Marzano²,

Sánchez

et al. 2015

2015 IEEE Symposium Series on Computational Intelligence

View full text Add to dashboard Cite

show abstract

“…Unfortunately, in described above approach there are no mechanisms responsible for respecting constraints defined in MOOP. Meanwhile in almost all real-life problems constraints are a crucial part of MOOP definition and it is nothing strange that among (evolutionary) algorithms for multi-objective optimization special attention is drown to techniques and algorithms for constrained multi-objective optimization and a varietymore or less effective-algorithms have been proposed (it is enough to mention here Jimenez-Verdegay-GomezSkarmeta's method [6], constrained tournament method or Ray-Tai-Seow's method [9]) 1 . In the course of this paper the idea as well as preliminary results of Constrained Evolutionary Multi-Agent System (conEMAS) for Multi-objective Optimization are presented.…”

Section: Evolutionary Multi-agent Systems In (Constrained) Multi-mentioning

confidence: 99%

Efficient Constrained Evolutionary Multi-Agent System for Multi-objective Optimization

Siwik

Sikorski

2008

2008 IEEE Congress on Evolutionary Computation (IEEE World Congress on Computational Intelligence)

View full text Add to dashboard Cite

Evolutionary Multi-agent System approach for optimization (especially for multi-objective optimization) is a very promising computational model. Its computational as well as implemental simplicity causes that approaches based on EMAS model can be widely used for solving optimization tasks. It turns out that introducing some additional mechanisms into basic EMAS-causes that EMAS-based system can be successfully applied for solving constrained multi-objective optimization tasks-and what is important results obtained by proposed approach are better/not worse than results obtained by NSGA-II or SPEA2 algorithms. In the course of this paper some extensions that can be introduced into EMAS in order to constrained multi-objective optimization are presented. What is important-any new additional mechanisms do not have to be introduced into EMAS to solve constrained optimization tasks-the only extensions causing that EMAS-based model becomes an efficient and simple both in conception as well as in implementation-is an appropriate strategy regarding the flow among agents crucial non-renewable resource which is usually called life energy. In this paper, both the idea as well as preliminary results of Constrained Evolutionary MultiAgent System (conEMAS) for Multi-objective Optimization are presented.

show abstract

“…Alternatively, one could use a more Pareto set like method whereby a population of valid networks is evolved and then those networks are explored for novelty via an altering objective function such as VEGA [19] or some kind of Pareto-reckoning approach such as SPEA [14]. In particular, if valid networks are clustered, one could evolve valid networks via a method of "niched-selection" [20].…”

Section: Weighting Of the Fitness Functions And Alternative Multi-objmentioning

confidence: 99%

Using Novelty-Biased GA to Sample Diversity in Graphs Satisfying Constraints

Overbury

Berthouze

2015

Proceedings of the Companion Publication of the 2015 Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

The structure of the network underlying many complex systems, whether artificial or natural, plays a significant role in how these systems operate. As a result, much emphasis has been placed on accurately describing networks using network theoretic metrics. When it comes to generating networks with similar properties, however, the set of available techniques and properties that can be controlled for remains limited. Further, whilst it is becoming clear that some of the metrics currently used to control the generation of such networks are not very prescriptive so that networks could potentially exhibit very different higher-order structure within those constraints, network generating algorithms typically produce fairly contrived networks and lack mechanisms by which to systematically explore the space of network solutions. In this paper, we explore the potential of a multi-objective novelty-biased GA to provide a viable alternative to these algorithms. We believe our results provide the first proof of principle that (i) it is possible to use GAs to generate graphs satisfying set levels of key classical graph theoretic properties and (ii) it is possible to generate diverse solutions within these constraints. The paper is only a preliminary step, however, and we identify key avenues for further development.

show abstract

An evolutionary algorithm for constrained multi-objective optimization

Cited by 60 publications

References 7 publications

Attribute Selection Via Multi-Objective Evolutionary Computation Applied to Multi-Skill Contact Center Data Classification

Attribute Selection Via Multi-Objective Evolutionary Computation Applied to Multi-Skill Contact Center Data Classification

Efficient Constrained Evolutionary Multi-Agent System for Multi-objective Optimization

Using Novelty-Biased GA to Sample Diversity in Graphs Satisfying Constraints

Contact Info

Product

Resources

About