A road investment model for the Sudan

Efficiency has become a critical factor in overall development of the Sudanese road system because of increasing demand for transportation, spatially separated supply and demand points, and a low level of service offered by the road network. In this study, an investment programming model is developed and applied to the interregional road transportation system in the Sudan to determine improvements which maximize the network efficiency as measured by the minimization of ton-kilometers weighted by relative travel time. The improvements of road surfaces from earth to gravel, gravel to pavement, and earth to pavement satisfy specified budget and demand requirements over a short planning horizon.

A heuristic procedure which involves solving an alternative series of shortest path and optimization problems is employed to solve the model for various budget levels. The result is a number of alternative investment program sets each of which results in substantial reduction of the total sum of weighted ton-kilometers used by the system. Based on a cost-effectiveness measure, the best investment program set is identified for possible implementation.

The degree to which the model results deviate from optimality is evaluated by comparing the heuristic solutions to the lower bounds generated by reformulating an equivalent integer model and solving it by linear programming. The results of the original model are found to deviate no more than 7-5% from optimality.

In its application to the Sudanese situation, the model is capable of handling aspects of project planning that have not been treated in this way before. Chief among these aspects are the spatial distribution of investment projects, their priority ranking, and measures of network efficiency improvements. It is suggested that the model could also be used to determine optimal vehicle fleet requirements. The model handles the problem of partial link improvements effectively and allows for link improvements to be made in a continuous rather than a discrete fashion.