DOMinant

Project description

Dominant was a research project supported by The Research Council of Norway, Program: eVITA.

Duration: August 2006 – December 2009

Project management:

	Professor Marielle Christiansen (Project manager, Department of Industrial Economics and Technology Management, NTNU)
	Professor Arne Løkketangen (Molde University College)
	Chief Scientist, Dr. Geir Hasle (SINTEF Applied Mathematics)

Objectives:

Improve methods for solving computationally hard discrete optimization problems in maritime and road-based transportation.

Including

Formulation of rich variants of the

	Inventory Routing Problem
	Fleet Size and Mix Vehicle Routing Problem

Solution Methods

	Exact methods
	Bounds
	Approximation methods
	Hybrid methods

Prototype Solvers

Computational experiments on instances from the literature and industry

Project Summary:

Efficient maritime and road-based transport is very important to economies, Norway in particular. In 2002, the cost of sea transport world-wide was 120 billion USD. Even small improvements in logistics efficiency through optimization based decision-support systems (DSS) will have huge economical and environmental effects. DOMinant will improve optimization and approximation methods that determine the performance of such DSS. Three groups in discrete optimization participate in the project:

	Department of Industrial Economics and Technology Management, NTNU
	The optimization group at Molde University College (MUC)
	SINTEF Applied Mathematics (SAM)

The three groups develop mathematical formulations of critical discrete optimization problems within maritime and road-based transport. In particular, the groups study the Inventory Routing Problem (IRP) and the Fleet Size and Mix Vehicle Routing Problem (FSMVRP). Classical variants will be extended with side constraints that are important in industrial applications.

NTNU focuses on exact solution methods, relaxations and bounds. MUC and SAM focus on approximation methods based on metaheuristics for industrial size problems. Together, we develop hybrid methods. In the development, extensive computational experiments on instances from industry and the literature is an integral part.