Travel-time Optimization on I-285 with Improved Variable Speed Limit Algorithms and Coordination with Ramp Metering Operations
The effectiveness of ramp metering (RM) systems in improving capacity and safety on mainline freeways has been well-established through empirical research. Similarly, the use of variable speed limit (VSL) systems has been proposed as a means of enhancing traffic flow. This study aimed to develop effective VSL control algorithms to minimize total travel time on the EB/SB I-285 corridor that encircles Atlanta between GA-400 and US78. To achieve this objective, a stochastic simulation-based optimization framework was employed that combines a microsimulation model and a genetic algorithm-based optimization module. This approach allowed for the optimization of the coordinated operation of VSL control in conjunction with existing RM control and the estimation of travel times to enhance the efficiency of VSL control. The combined variable speed limit and ramp metering (VSL-RM) algorithm aimed to maximize freeway capacity by avoiding the capacity drop phenomenon at merge bottlenecks. This study focused on the onset period of evening peak congestion and used 52 of Georgia DOT’s Vehicle Detection System (VDS) detectors that collected 20-second interval volume, speed, and occupancy data, during a one-month period in April 2016.
This study proposed a combined VSL-RM algorithm designed to maximize freeway capacity by mitigating the capacity drop phenomenon at merge bottlenecks. The optimal speed of the VSL-RM algorithm was derived from shock wave theory and evaluated through experimentation using a traffic microsimulation software. A stochastic simulation-based optimization framework comprising of the microsimulation model and a genetic algorithm-based optimization module was utilized to determine the optimal parameter values of the VSL-RM system that minimize total vehicle travel time. The performance of the VSL-RM algorithm was compared with no control, RM only and current GDOT VSL control.
This project produced results of the simulation-optimization for four cases (no control, the RM control only, current GDOT VSL control, and the VSL-RM control):
- RM control alone resulted in 4.7 percent reduction in total system travel times and 5.1 percent reduction in total freeway travel times (measured in vehicle hours), compared to no control. Travel times on the ramp increased by 10.9 percent.
- VSL-RM control was estimated to reduce the total system and freeway travel times by 8 percent and 8.5 percent, respectively, compared to no control. Travel times on the ramp increased by 9.7 percent.
- VSL-RM control was estimated to reduce the total system travel times by 15 percent, compared to GDOT’s current VSL algorithm.
- The main benefits of the controls were that they delay the bottleneck formation time and lessen the severity of the bottleneck (passing speed).