A Heuristic Approach for GPS-Based Routing

Abstract

Vehicle routing (such as for package delivery) presents challenges for operations planning and operations control. Planning ensures that vehicles are assigned to “good” routes, and control enables routes to be changed in real time in response to changes in destination requirements. Both planning and control can be accomplished using web-based, intelligent geographic information system tools to rapidly generate a heuristic solution using an embedded algorithm, rather than the established approach of using explicit customized optimization models. The authors contrast the established approach of using customized integer optimization models to a heuristic that integrates human judgment with Google Maps travel time data to solve vehicle routing problems. This paper discusses the data requirements, simplifying assumptions, and practical performance of both approaches. The advantage of the heuristic approach is that genuine, useful access to much of the power of highly sophisticated OR network models can be provided to large numbers of analytically unsophisticated managers, along with enhanced operational control.