Understanding the Difficulties of Many-Objective Optimization

Evolutionary algorithms have gained wide recognition and acceptance in the domain of multiobjective optimization over the past decade. However there is widespread agreement that their efficiency at dealing with problems of more than 3 objectives is at best variable. There is hotbed of activity in attempting to understand the cause of the problem, with claims and counter-claims being made at frequent intervals. An easy conclusion seems beyond our grasp, but there is much exciting and important work to be done in this space. This project will take both theoretical and practical approaches to understanding the problem in an organized fashion.

PhD candidates with an interest in machine learning and optimization should apply. A strong mathematical background is desirable for this project.