1992 Volume 1992 Issue 1 Pages 21-37
We consider a freight network design problem which determines optimal routes and shipment size of multiple commodities between various origin-destination pairs in a network to minimize the sum of the transportation and inventory costs. We formulate this problem as a nonlinear programming problem and develop a composite algorithm of lower and upper bounding procedures using Lagrangean relaxation technique.
Our numerical experiments with up to 40 nodes for layered network problems and 50 nodes for general network problems show that our algorithm generates good solutions within reasonable computational time.