Reconfiguration Strategies for Mitigating the Impacts of Port Disruptions

The use of variable speed limits (VSL) along highway lanes in an effort to control traffic flow is a technique that has been around for some time. It has been implemented in various parts of Europe and has been studied by several researchers during the recent years. The incentive of using VSL has been mainly safety from the application point of view but benefits such as improved traffic flow rates, lower travel times, smooth speed and density distribution and possibly lower pollution have been conjectured in literature and in some cases analyzed using mainly macroscopic traffic models. The use of aggregate flow or macroscopic models and optimization techniques to develop and test these dynamic VSL controllers (which can be activated during incidents or other disruptions) has led to optimistic results in terms of lowering travel times, sometimes by as much as 20%. However, a closer look at these results shows that using the same macroscopic model to design and test VSL controllers raises questions as to whether the simplicity of the models used for evaluation are responsible for these optimistic results given the fact that other studies using microscopic simulations fail to demonstrate improvements on travel times, albeit for different VSL strategies. The question of whether the VSL strategies, or the macroscopic models used to analyze them, or both are responsible for the large differences (not only quantitative but also qualitative) in traffic flow and travel time benefits reported, remains unanswered. Consequently, the problem of what are the most appropriate dynamic VSL controllers and what benefits can be guaranteed in a consistent manner under different traffic flow conditions is an open one. This research capitalizes on past results obtained from modeling and traffic flow analyses. It addresses the design, analysis, and evaluation of dynamic VSL controllers. A control engineering approach is used, in which the control strategies are designed based on simplified models (validated macroscopic traffic models, in this case), but applied and tested on validated microscopic traffic models under different traffic conditions. The dynamic VSL control design will be based on optimizing travel times, smoothness of traffic, environmental impacts, and, indirectly, safety. Macroscopic traffic models and microscopic simulation models based on VISSIM of large sections of I-110 and I-710 in Los Angeles County will be used as examples for design and evaluation. The I-110 has several bottlenecks where VSL control may help during peak hours and I-710 has a high volume of trucks serving two major ports which will allow us to examine the impact of truck traffic on VSL. Data from California's Performance Measurement System (PeMS) and other sources will be used to validate the macroscopic and microscopic simulation models. The evaluation of the dynamic VSL control strategies will be based on comparison of the benefits in terms of traffic flow, travel times, smoothness of traffic, number of lane changes, and environmental impacts with control and without control under different traffic conditions which will include bottlenecks, incidents, and high/low peak traffic situations. The modal emissions model developed by the research group at the University of California at Riverside for different classes of vehicles will be used to measure fuel consumption and pollution levels for all vehicles in the network in order to assess the environmental impacts of different VSL strategies. Monte Carlo simulations will be performed in order to reduce the effect of random errors and check the consistency of results.