One PLOT to Show Them All: Visualization of Efficient Sets in Multi-Objective Landscapes

TL;DR

This paper introduces PLOT, a hybrid visualization method that combines local and global optimality information in multi-objective landscapes, enhancing understanding of solution quality and attraction basins.

Contribution

It proposes a novel visualization technique that integrates gradient field heatmaps with local efficiency approximation to depict the entire landscape of multi-objective problems.

Findings

PLOT effectively visualizes local and global optimality in MOPs.

The method combines divergence of gradient fields with dominance relations.

It provides a comprehensive view of the landscape, including basins of attraction.

Abstract

Visualization techniques for the decision space of continuous multi-objective optimization problems (MOPs) are rather scarce in research. For long, all techniques focused on global optimality and even for the few available landscape visualizations, e.g., cost landscapes, globality is the main criterion. In contrast, the recently proposed gradient field heatmaps (GFHs) emphasize the location and attraction basins of local efficient sets, but ignore the relation of sets in terms of solution quality. In this paper, we propose a new and hybrid visualization technique, which combines the advantages of both approaches in order to represent local and global optimality together within a single visualization. Therefore, we build on the GFH approach but apply a new technique for approximating the location of locally efficient points and using the divergence of the multi-objective gradient vector field as a robust second-order condition. Then, the relative dominance relationship of the determined locally efficient points is used to visualize the complete landscape of the MOP. Augmented by information on the basins of attraction, this Plot of Landscapes with Optimal Trade-offs (PLOT) becomes one of the most informative multi-objective landscape visualization techniques available.