Optimisation of variable speed limits at the freeway lane drop bottleneck

Abstract

The primary objectives of this study were to use variable speed limits (VSL) upstream of freeway lane drop to maintain capacity and reduce congestion. As driving behaviours are the main reasons leading to capacity drop and the microscopic simulation can reflect driving behaviours precisely, microscopic simulations were first used to test lane drop scenarios. The objective function and constraints determined according to traffic engineering practice were optimised using a modified genetic algorithm (GA) based on microscopic simulation to get the optimal speed limit combination. The modified GA can guarantee the solution diversity and optimal results. Then, the cell transmission model, a macroscopic flow model, was used to crosscheck the simulated results. Both microscopic and macroscopic analysis results demonstrated that VSL could only improve lane drop traffic efficiency if speed limits were set appropriately. This study provided a new process from microscopic to macroscopic aspects for analysing traffic problems.