Optimization Models for a Delivery Truck Environment
Team: Robert Walters, Donnet Phillips
Faculty advisor: Dr. Barr Year: 1986
Documents: Final report (PDF)
There are many optimization procedures and algorithms that can be implemented for delivery operations. The thrust of this paper concerns the following topics and their relavance to delivery vehicles: equipment replacement, truck utilization and optimization, and efficient routing practices.
An important decision facing all businesses is when to replace equipment. Equipment can be industrial machinery, office machines, or vehicles. In our consideration here, we look at the gains in recent years in the area of equipment replacement because of the availability of interactive financial planning languages such as IFPS. The development of an IFPS model for an optimal equipment replacement strategy for a relatively small delivery fleet is presented and examined here. Questions such as the lease versus buy decision are investigated.
Next, a linear programming model is developed which determines an organization’s optimal number of vehicles to maintain in order to meet current demands while staying within specified constraints. Questions are answered concerning which routes should be run on which days in order to meet organizational constraints and not violate current organizational structure. The resulting model that was developed was solved using LINDO. A sensitivity analysis was performed by the authors to present management with two different operation strategies.
Finally, the routing of the trucks was investigated. We investigated the problem of the m routes n stops routing problem. The model was developed as a ‘traveling-salesman problem’ because the trucks started and ended at a central location. The algorithm we present can be applied to any travelling salesman problem with n stops on a route with a relatively small n. The unique structure of our problem aillowed us to develop a general policy to follow in the routing of the trucks. The number of stops in each route of our problem was small and we were able to develop a ‘try-all-possibilities’ algorithm rather than utilize heuristics. The code for our algorithm as well as the output is presented.
The results of our study can be applicable not only to small delivery fleet operations but any type of organization that maintains fleet vehicles.