Authors:
Imen Ben Mansour
;
Ines Alaya
and
Moncef Tagina
Affiliation:
National School of Computer Sciences, COSMOS Laboratory and University of Manouba, Tunisia
Keyword(s):
Multi-objective Multidimensional Knapsack Problem, Iterated Local Search, Scalarization Functions, Tchebycheff Functions.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Business Analytics
;
Cardiovascular Technologies
;
Computing and Telecommunications in Cardiology
;
Data Engineering
;
Decision Support Systems
;
Decision Support Systems, Remote Data Analysis
;
Enterprise Software Technologies
;
Health Engineering and Technology Applications
;
Intelligent Problem Solving
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Software Engineering
;
Symbolic Systems
Abstract:
The multi-objective multidimensional knapsack problem (MOMKP) which is one of the hardest multi-objective combinatorial optimization problems, presents a formal model for many real world problems. Its main goal consists in selecting a subset of items in order to maximize m objective functions with respect to q resource constraints. For that purpose, we present in this paper a resolution approach based on a Min-Max Tchebycheff iterated Local Search algorithm called Min-Max TLS. In this approach, we propose designing a neighborhood structure employing a permutation process to exploit the most promising regions of the search space while considering the diversity of the population. Therefore, Min-Max TLS uses Min-Max N(s) as a neighborhood structure, combining a Min-Extraction-Item algorithm and a Max-Insertion-Item algorithm. Moreover, in Min-Max TLS two Tchebycheff functions, used as a selection process, are studied: the weighted Tchebycheff (WT) and the augmented weighted Tchebycheff
(AugWT). Experimental results are carried out with nine well-known benchmark instances of MOMKP. Results have shown the efficiency of the proposed approach in comparison to other approaches.
(More)