A Bandit-Inspired Memetic Algorithm for Quadratic Assignment Problems
Year: 2012
Type: dissertation
Abstract: In this thesis a new metaheuristic for combinatorial optimization is proposed,\nwith focus on the Quadratic Assignment Problem as the hard-problem of choice -\na choice that is reflected in the name of the method, BIMA-QAP. The algorithm\nemploys a memetic structure and stores information on the single components\nalong the search. This information is used to guide the search, through an\noperator inspired by the solution approaches to the Multi-Armed Bandit model.\n\nOnce the algorithm has been laid out and its set of parameters defined, its\nimplementation has been extensively tested under a Naive-Bayesian assumption of\nindependence among the parameters. The results show that BIMA-QAP consistently\nperforms better than Multi-start Local Search, and the new operator perturb()\nalters the solutions better than a randomized approach.
Author: Francesco Puglierin
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W610424308