Proceedings of the 3rd International Conference on Operations Research and Enterprise Systems pages:328-333
International Conference on Operations Research and Enterprise Systems (ICORES 2014) location:Angers, France date:06-08 March 2014
In the single-objective automated algorithm configuration problem, given an algorithm with a set of parameters that need to be configured and a distribution of problem instances, the automated algorithm configurator will try to search for a good parameter configuration based on a pre-defined performance measure. In this paper, we point out two motivations for the development of a multi-objective algorithm configurator, in which more than one performance measure are considered at the same time. The first motivation is a parameter configuration case study for a deterministic single machine scheduling algorithm with two performance measures: minimization of the average running time and maximization of the total number of optimal solutions. The second one is the configuration problem for non-exact multi-objective optimization algorithms. In addition, a discussion of solving approach for the first motivating problem is also presented.