FORECASTING VARIABLE ROUTE COSTS IN THE TRANSPORTATION INDUSTRY

Larry Woodward
Brad Clark
University of Mary Hardin-Baylor

ABSTRACT

The US transportation industry moves roughly fifty-five million tons of raw materials and goods valued at more than forty-nine billion dollars every day. Transportation and wholesale firms alone deliver approximately 14 billion tons of products annually. Approximately 70% of these goods travel less than 250 miles from the point of origin to their final destination. For this reason, transportation firms segregate business operations into local or regional networks comprised of distribution hubs and delivery routes. The high fixed costs associated with maintaining a transportation network dictate that routing must be optimized for both logistical efficiency as well as cost efficiency. However, identifying the optimal combination of routes for a network is a complex process that must account for variances in customers, delivery windows, and order volume over time, among other factors.
In optimized networks, all routes form a cohesive system that is able to satisfy customer demands and meet business objectives in the most efficient manner. Choosing the optimal solution(s), then, require an awareness of route cost efficiency and knowledge of the relationship between total route cost and variables that impact route performance. The authors employ linear regression methods to forecast variable costs of fixed routes, and groups of routes, based upon historical route data and changing customer demands to facilitate network optimization decisions and processes.

Keywords: Transportation network, logistical efficiency, delivery routes, rout