A Multi-Time Period Asphalt Distribution Model was developed for the United Refining Company
(URC). Like the gasoline distribution model, it was designed so that individuals with no experience in operations research could use the model to develop seasonal distribution plans, examine the manufacturing process, examine the time value and cost of inventories, develop appropriate exchange agreements, examine buy/sell options, and several other options. The use of the model results in significant reductions in operating costs.

The system models shipments of different grades of asphalt from refineries to customers over a year. In addition to direct shipments, the system permits more complicated shipping procedures including exchange contracts (with regrades being taken into account) and restrictions on shipments of groups of products within single shipments. An important part of the model is a front-end manufacturing model that permits a quick analysis of the manufacturing options. Among several other features, the model also allows the analysis of buy/sell options and the options of incremental sales.

An accessible, menu-driven user interface permits easy modification of the components in the model without requiring any knowledge about the associated optimization problem. The model to be solved is typically a mixed integer program. It is solved using a version of MINOS modified to include a branch and bound integer programming algorithm in addition to the linear and nonlinear programming procedures it already contains. Output appears in terms of the data components familiar to the user instead of variables from an optimization problem, so interpretation of the model’s results is as easy as specifying its characteristics.