Lakmali Weerasena presents: A Local-search Algorithm and a Tolerance Function for Multi-objective Combinatorial Optimization Problems.
If you go
What: Math Colloquium: Lakmali Weerasena
When: Friday, Oct. 20 at 2 p.m.
Where: EMCS 422
Admission: This talk may be appropriate for undergraduate and graduate students with an interest in operations research and applications.
A multi-objective combinatorial optimization (MOCO) problem consists a solution set, known as Pareto set, due to the conflicting nature of the objective functions. In the literature variety of methods have been developed to find representations for Pareto sets of NP-hard MOCO problems and many of them do not provide error-bounds for the representation. In this study, a local-search algorithm is proposed to find representations for the class of 0-1 MOCO problems. In addition, a tolerance function is defined to identify the maximum error of the representation and is derived for the proposed algorithm. It is shown that the tolerance function depends on the characteristics of the objective functions of the MOCO problem and the initial solutions. Further, the performance of the algorithm is examined using some well-known MOCO problems.