Conference article

Evaluation aspects of alternative traf?c signal control strategies for an arterial network are studied. The traf?c evolution of a signalized road network is modelled as a Store and Forward (SF) network of queues. The system state is the vector of all queue lengths at all intersections. The signal control at any time permits certain simultaneous turn movements at each intersection at pre-speci?ed saturation rates. Two control categories, open loop and traf?c-responsive policies are compared under ?xed and time-varying demand. The behaviour of the underlying queuing network model manifesting asynchronous nature over time while involving concurrence is modelled according to an event-driven approach virtually reproduced by discrete event simulations. Exploration of the implementation outputs results a pertinent mathematical framework for traf?c movement, analysis and signal control design. Subsequently, various metric measurements such as queue bounds, delays, trajectory travel times quantify the actual policy. Moreover, aggregate behaviour as in a macroscopic queuing model is also prompted. Experiments are performed using real data for a section of the Huntington-Colorado arterial adjacent to the I-210 freeway in Los Angeles. Lastly, the meso-micro simulation issues resulting from the employed decision tool, PointQ, are compared with microsimulation and mesosimulation forms of other traf?c simulation programs.

traf?c responsive signal, adaptive control, pre-timed control, max-pressure practical policy, discrete event simulation