Simulation-based optimization for heterogeneous traffic control

Abstract

One of the major reasons for frequent bottlenecks at intersections is the operation of poorly designed signals with static timings, irrespective of the variations in traffic flow. The fixed signal timings are calculated using conventional design methodologies such as Webster’s and HCM signal design methodologies that are primarily valid for lane disciplined homogenous traffic conditions. Implementation of such design procedures under mixed traffic conditions may not yield the best performance. Therefore, the present study aims to develop a simulation-based optimization that optimizes the performance of the intersection. Spatial performance measure such as travel time is perceived by the users and operators alike and therefore considered for minimization. A simulation-based optimization is performed by implementing various derivative-free optimization algorithms such as Nelder–Mead simplex algorithm and COBYLA using a microsimulation software, VISSIM. The obtained results are better when compared with that of traditional Webster’s design.