Many real-world optimization problems involve both multiple objectives and constraints. Although constraint handling in multiobjective optimization has been considered in the literature, there is still a high demand for more advanced and versatile constraint handling techniques (CHTs) in real-world applications. For this reason, we propose a general approach to combine multiple CHTs into an ensemble-based method, providing a framework to easily construct new CHTs from existing ones. The approach is evaluated on nine test problems from the literature using an ensemble of four widely-used CHTs. The experimental results show that the ensemble is more robust than single CHTs and performs at least as well as the best single CHTs on all the test problems. Moreover, a positive synergistic effect of the ensemble is demonstrated on three problems.
Paper
Full text
Ensemble-based constraint handling in multiobjective optimization
Semantic Scholar · Computer Science · 2019
Abstract
Many real-world optimization problems involve both multiple objectives and constraints. Although constraint handling in multiobjective optimization has been considered in the literature, there is still a high demand for more advanced and versatile constraint handling techniques (CHTs) in real-world applications. For this reason, we propose a general approach to combine multiple CHTs into an ensemble-based method, providing a framework to easily construct new CHTs from existing ones. The approach is evaluated on nine test problems from the literature using an ensemble of four widely-used CHTs. The experimental results show that the ensemble is more robust than single CHTs and performs at least as well as the best single CHTs on all the test problems. Moreover, a positive synergistic effect of the ensemble is demonstrated on three problems.