A Comparative Visual Analytics Framework for Evaluating Evolutionary Processes in Multi-Objective Optimization

Evolutionary multi-objective optimization (EMO) algorithms have been demonstrated to be effective in solving multi-criteria decision-making problems. In real-world applications, analysts often employ several algorithms concurrently and compare their solution sets to gain insight into the characteristics of different algorithms and explore a broader range of feasible solutions. However, EMO algorithms are typically treated as black boxes, leading to difficulties in performing detailed analysis and comparisons between the internal evolutionary processes. Inspired by the successful application of visual analytics tools in explainable AI, we argue that interactive visualization can significantly enhance the comparative analysis between multiple EMO algorithms. In this paper, we present a visual analytics framework that enables the exploration and comparison of evolutionary processes in EMO algorithms. Guided by a literature review and expert interviews, the proposed framework addresses various analytical tasks and establishes a multi-faceted visualization design to support the comparative analysis of intermediate generations in the evolution as well as solution sets. We demonstrate the effectiveness of our framework through case studies on benchmarking and real-world multi-objective optimization problems to elucidate how analysts can leverage our framework to inspect and compare diverse algorithms.

1 Introduction

Decision making under more than one criteria often appears in real-world optimization problems. Unlike single-aspect decision tasks where only one target will be satisfied, multiple aspects should be considered simultaneously to obtain optimal solutions with trade-offs between different yet often conflicting objectives. Considering water reservoir systems as an example [49], the operators are intended to design various operation policies that maximize power production. However, blindly optimizing the factors for power production may result in negative impact on irrigation or even increase the risk of flooding. To solve the multi-criteria decision making problems, various types of multi-objective optimization algorithms have been developed in the recent decades. The underlying mechanism of these algorithms is to search for a solution set where each solution cannot dominate the others, i.e., better than another solution on all objectives. Owing to the non-dominance nature in the solution set, decision makers are able to choose between a variety of feasible solutions to meet the incoming requirements.