Research Repository

Bandit-based Random Mutation Hill-Climbing

Liu, J and Perez Liebana, D and Lucas, SM (2017) Bandit-based Random Mutation Hill-Climbing. In: IEEE Congress on Evolutionary Computation (CEC), 2017, 2017-06-05 - 2017-06-08, San Sebastian, Spain.

[img]
Preview
Text
1606.06041v1.pdf - Accepted Version

Download (413kB) | Preview

Abstract

The Random Mutation Hill-Climbing algorithm is a direct search technique mostly used in discrete domains. It repeats the process of randomly selecting a neighbour of a best-so-far solution and accepts the neighbour if it is better than or equal to it. In this work, we propose to use a novel method to select the neighbour solution using a set of independent multi-armed bandit-style selection units which results in a bandit-based Random Mutation Hill-Climbing algorithm. The new algorithm significantly outperforms Random Mutation Hill-Climbing in both OneMax (in noise-free and noisy cases) and Royal Road problems (in the noise-free case). The algorithm shows particular promise for discrete optimisation problems where each fitness evaluation is expensive.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Published proceedings: 2017 IEEE Congress on Evolutionary Computation, CEC 2017 - Proceedings
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Depositing User: Elements
Date Deposited: 08 Sep 2017 13:00
Last Modified: 28 Nov 2017 16:15
URI: http://repository.essex.ac.uk/id/eprint/20340

Actions (login required)

View Item View Item