TIME/PLACE: April 11th at 16:00 in FAMNIT-1-VP
LECTURER: Zorica Stanimirovič, izredna profesorica z
Oddelka za numerično matematiko in optimizacijo na
Fakulteti za matematiko Univerze v Beogradu.
TITLE: Variable neighborhood search for a vehicle scheduling problem in agriculture
ABSTRACT:
A Vehicle Scheduling Problem (VSP) that arises from sugar beet transportation within minimum working time under the set of constrains reflecting a real-life situation is considered. The problem is first modelled as Mixed Integer Quadratically Constrained Program (MIQCP) and then reformulated to Mixed Integer Linear Program (MILP). MILP formulation was used within the framework of Lingo 17 solver, producing optimal solutions only for small-size problem instances. Two variants of Variable Neighborhood Search metaheuristic – Basic VNS (BVNS) and Skewed VNS (SVNS) are designed to efficiently deal with large size problem instances. The proposed VNS approaches are evaluated and compared against Lingo 17 and each other on the set of real-life and generated problem instances. Both BVNS and SVNS reach all known optimal solutions on small-size instances and show similar performance on medium and large size instances regarding solution quality. In general, SVNS significantly outperforms BVNS in the sense of running times.
The content presented on the seminar is joint work with Ana ANOKIč, Tatjana DAVIDOVIč and ÄorÄe STAKIč.
Welcome!
