Title: Characterization of neighborhood behaviours in a multi-neighborhood local search algorithm
Authors: Dang, Nguyen Thi Thanh ×
De Causmaecker, Patrick #
Issue Date: 29-May-2016
Publisher: Springer
Series Title: Lecture Notes In Computer Science
Conference: Learning and Intelligent OptimizatioN Conference edition:10 location:Ischia Island (Napoli), Italy date:29/5 - 1/6 2016
Abstract: We consider a multi-neighborhood local search algorithm with a large number of possible neighborhoods. Each neighborhood is accompanied by a weight value which represents the probability of being chosen at each iteration. These weights are fixed before the algorithm runs, and are considered as parameters of the algorithm. Given a set of instances, off-line tuning of the algorithm's parameters can be done by automated algorithm configuration tools (e.g., SMAC). However, the large number of neighborhoods can make the tuning expensive and difficult even when the number of parameters has been reduced by some intuition. In this work, we propose a systematic method to characterize each neighborhood's behaviours, representing them as a feature vector, and using cluster analysis to form similar groups of neighborhoods. The novelty of our characterization method is the ability of reflecting changes of behaviours according to hardness of different solution quality regions. We show that using neighborhood clusters instead of individual neighborhoods helps to reduce the parameter configuration space without misleading the search of the tuning procedure. Moreover, this method is problem-independent and potentially can be applied in similar contexts.
Publication status: accepted
KU Leuven publication type: IC
Appears in Collections:Computer Science, Campus Kulak Kortrijk
Computer Science - miscellaneous
× corresponding author
# (joint) last author

Files in This Item:

There are no files associated with this item.

Request a copy


All items in Lirias are protected by copyright, with all rights reserved.