Fleet configuration subject to stochastic demand: an application in the distribution of liquefied petroleum gas

• Published in: The Journal of the Operational Research Society Vol. 53; no. 9; pp. 961 - 971
• Main Authors: Singer, M, Donoso, P, Jara, S
• Format: Journal Article
• Language: English
• Published: London Taylor & Francis 01.09.2002; Palgrave Macmillan Press; Palgrave Macmillan UK; Palgrave; Taylor & Francis Ltd
• Subjects: Analytical models; Applied sciences; Business and Management; Case-oriented Paper; Case-Oriented Papers; Cylinders; Delivery vehicles; Distribution; Exact sciences and technology; Flows in networks. Combinatorial problems; Linear programming; Liquefied petroleum gas; Logistics; Management; Markov analysis; Markov processes; Mathematics; Modeling; Motor vehicle fleets; Operational research and scientific management; Operational research. Management science; Operations Research/Decision Theory; Performance metrics; Probability and statistics; Probability theory and stochastic processes; Queuing; Queuing theory. Traffic theory; Routing; Sciences and techniques of general use; Simulation; Simulations; Studies; Truck routes; Trucks; Vehicle fleets; vehicle routing; Markov processes; queuing; distribution; vehicle routing; simulation; Markov process; LPG; Vehicle routing problem; Non linear programming; Buffer system; Waiting time; Delivery (good); Simulation; Queueing system; Time average; Distribution; Analytical method; Objective function
• ISSN: 0160-5682; 1476-9360
• DOI: 10.1057/palgrave.jors.2601376

We study the problem of configuring a fleet, in which vehicles receive information on-line about the demand that they should fulfil while they are on the road. In each district it must be decided the number of vehicles and their capacity. The objective function is to minimise the operational cost subject to constraints for the minimum delivery capacity, the maximum vehicle size and the average waiting time for customers. The last constraint is modelled as a queuing system that is adjusted according to the simulation of the delivery process of a Chilean company that distributes liquefied petroleum gas in portable cylinders. We provide the analytical form of all the components of the model, so it can be solved using a standard non-linear programming package. We show that the fleet may increase its sales by 3% and reduce the waiting time of customers 10% by allowing a set of vehicles to share the buffer of orders rather than having vehicles to exclusively serve smaller sectors.