A stochastic programming approach for heterogeneous variable message sign location problem for freeway networks

Variable message signs (VMS) are electronic signage systems that display real-time traffic information to drivers to mitigate congestion and reduce travel time. We propose a heterogeneous VMS location problem based on a stochastic model of accidents on a freeway network. We consider both gantry and cantilever mounted VMS that displays both passive and active real-time messages. The problem is formulated as a two-stage stochastic programming model. The first-stage model determines the location and type of VMS installation. The second stage evaluates the performance of VMS location solutions by minimizing travelers’ travel time and the penalty for misleading guidance. The model is formulated as a mixed-integer linear programming (MILP) problem that can be solved using the Benders decomposition (BD) algorithm. The Nguyen-Dupuis and Sioux-Fall networks are used to verify the effectiveness of the proposed models. We believe this study will provide practical guidance to freeway administrators.