Escaping Local Optima with Penalties in Distributed Iterative Improvement Search

Abstract

The advantages offered by iterative improvement search make it a popular technique for solving problems in centralised settings. However, the key challenge with this approach is finding effective strategies for dealing with local optima. Such strategies must push the algorithm away from the plateaux in the objective landscape and prevent it from returning… (More)

Topics

10 Figures and Tables

Slides referencing similar topics