Dynamic Testing of an Analytical Traffic Model

The research goal of the presented experiment was to test if the current control plan can be improved through analytical modeling of the traffic by defining a bi-level problem and dynamically simulating the existing control plan and solving the analytical model. The comparison is done by queue lengths in an urban network of crossroads. Two models were developed for this purpose - a python-based package, named PAO and a solver were used for definition and solution of a bi-level optimization model and a microsimulation model was developed to test the results in a simulation environment named Aimsun. The results showed that the bilevel optimization led to better settings for the duration of green light but it also revealed that the originally measured settings provide good performance of the investigated traffic network.