Visualising the search process for multi-objective optimisation
Mardé Helbig · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2018
This paper proposes different visualisation techniques to understand the behaviour of an algorithm's entities during the search process when solving multi-objective optimisation problems. A scatter plot is used to highlight the Pareto-ranking of the entities during the search. Different fronts of the entities are indicated through the ball sizes of the scatter plot. Using a parallel coordinate plot for visualising the effect of control parameter values on the performance of an algorithm is also proposed. Possible extensions for dynamic multi-objective optimisation are also discussed.