An Interactive Framework for Offline Data-Driven Multiobjective Optimization

Mazumdar, Atanu; Chugh, Tinkle; Hakanen, Jussi; Miettinen, Kaisa

doi:10.1007/978-3-030-63710-1_8

Cited by 7 publications

(7 citation statements)

References 18 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, a hybrid of the two approaches can provide the benefits of both and produce a set of solutions with a wider range of uncertainties and objective values. This is especially advantageous for decision making due to the wider choices it provides [26]. A flowchart of the hybrid approach is shown in Figure 7.…”

Section: Hybrid Of Probabilistic and Generic Approachesmentioning

confidence: 99%

Probabilistic Selection Approaches in Decomposition-Based Evolutionary Algorithms for Offline Data-Driven Multiobjective Optimization

Mazumdar

Chugh

Hakanen

et al. 2022

IEEE Trans. Evol. Computat.

Self Cite

View full text Add to dashboard Cite

In offline data-driven multiobjective optimization, no new data is available during the optimization process. Approximation models, also known as surrogates, are built using the provided offline data. A multiobjective evolutionary algorithm can be utilized to find solutions by using these surrogates. The accuracy of the approximated solutions depends on the surrogates and approximations typically involve uncertainties. In this paper, we propose probabilistic selection approaches that utilize the uncertainty information of the Kriging models (as surrogates) to improve the solution process in offline data-driven multiobjective optimization. These approaches are designed for decomposition-based multiobjective evolutionary algorithms and can, thus, handle a large number of objectives. The proposed approaches were tested on distance-based visualizable test problems and the DTLZ suite. The proposed approaches produced solutions with a greater hypervolume, and a lower root mean squared error compared to generic approaches and a transfer learning approach that do not use uncertainty information.

show abstract

Section: Hybrid Of Probabilistic and Generic Approachesmentioning

confidence: 99%

Probabilistic Selection Approaches in Decomposition-Based Evolutionary Algorithms for Offline Data-Driven Multiobjective Optimization

Mazumdar

Chugh

Hakanen

et al. 2022

IEEE Trans. Evol. Computat.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Each iteration uses the preferences to update the decomposition and guide the search toward the region of interest. The most common modification to the decomposition involves rearranging the reference vectors according to the preference information [3,12,15,19]. Some other methods utilize the preference information to modify the approximation of the ideal point required by the decomposition-based MOEA [23,24].…”

Section: Related Workmentioning

confidence: 99%

A General Architecture for Generating Interactive Decomposition-Based MOEAs

Lárraga

Miettinen

2022

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Evolutionary algorithms have been widely applied for solving multiobjective optimization problems. Such methods can approximate many Pareto optimal solutions in a population. However, when solving real-world problems, a decision maker is usually involved, who may only be interested in a subset of solutions that meet their preferences. Several methods have been proposed to consider preference information during the solution process. Among them, interactive methods support the decision maker in learning about the trade-offs among objectives and the feasibility of solutions. Also, such methods allow the decision maker to provide preference information iteratively. Typically, interactive multiobjective evolutionary algorithms are modifications of existing a priori or a posteriori algorithms. However, they mainly focus on finding a region of interest and do not support the decision maker finding the most preferred solution. In addition, the cognitive load imposed on the decision maker is usually not considered. This article proposes an architecture for developing interactive decomposition-based evolutionary algorithms that can support the decision maker during the solution process. The proposed architecture aims to improve the applicability of interactive methods in solving real-world problems by considering the needs of a decision maker. We apply our proposal to generate an interactive decompositionbased algorithm utilizing a reference vector re-arrangement procedure and MOEA/D. We demonstrate the performance of our proposal with a real-world problem and multiple benchmark problems.

show abstract

“…In some studies, the performance of interactive methods has been assessed by considering each interaction as a distinct a priori step [23,24,26], and indicators for a priori methods have been used to assess the median performance of interactions of each phase. This allows the use of existing (a priori) indicators in the absence of those designed for interactive methods.…”

Section: Gp9mentioning

confidence: 99%

Desirable properties of performance indicators for assessing interactive evolutionary multiobjective optimization methods

Pour

Bandaru

Afşar

et al. 2022

Proceedings of the Genetic and Evolutionary Computation Conference Companion

View full text Add to dashboard Cite

Interactive methods support decision makers in finding the most preferred solution in multiobjective optimization problems. They iteratively incorporate the decision maker's preference information to find the best balance among conflicting objectives. Several interactive methods have been developed in the literature. However, choosing the most suitable interactive method for a given problem can prove challenging and appropriate indicators are needed to compare interactive methods. Some indicators exist for a priori methods, where preferences are provided at the beginning of the solution process. We present some numerical experiments that illustrate why these indicators are not suitable for interactive methods. As the main contribution of this paper, we propose a set of desirable properties of indicators for assessing interactive methods as the first step of filling a gap in the literature. We discuss each property in detail and provide simple examples to illustrate their behavior.

show abstract

An Interactive Framework for Offline Data-Driven Multiobjective Optimization

Cited by 7 publications

References 18 publications

Probabilistic Selection Approaches in Decomposition-Based Evolutionary Algorithms for Offline Data-Driven Multiobjective Optimization

Probabilistic Selection Approaches in Decomposition-Based Evolutionary Algorithms for Offline Data-Driven Multiobjective Optimization

A General Architecture for Generating Interactive Decomposition-Based MOEAs

Desirable properties of performance indicators for assessing interactive evolutionary multiobjective optimization methods

Contact Info

Product

Resources

About