Arriving at optimal signal timing parameters to improve the efficiency of traffic flow has been one of the major challenges faced by traffic engineers. The choice of a robust optimization framework and an accurate traffic model plays a significant role in determining optimal signal timing parameters. Though traffic flow is intuitively stochastic, few studies incorporate stochasticity in their optimization framework for traffic signal design. This study proposes two simulation-based stochastic optimization algorithms—an evolutionary algorithm-based framework and a simultaneous perturbation stochastic approximation (SPSA) algorithm-based framework for the optimal signal design of an isolated intersection using a calibrated microsimulation environment with reasonable accuracy. A software-in-the loop approach is used to control the traffic signals in the microsimulation environment. SPSA is a gradient descent algorithm with a powerful approach for approximating the gradient with just two function evaluations per gradient approximation. To evaluate the performance of the two frameworks, the study optimizes the signal timings for a case study on an isolated intersection in an urban arterial in Chennai. On comparing the two algorithms, it is found that SPSA performed better and took 100 function evaluations less than that taken by GA. A better (near optimal) initial solution is found to yield a faster rate of convergence for both algorithms. As the proposed optimization framework incorporates the stochastic nature of traffic in the optimization algorithm, it can accommodate the temporal variations in traffic and thereby provide traffic engineers a robust signal control strategy for improving the efficiency of traffic flow.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    A Microsimulation-Based Stochastic Optimization Approach for Optimal Traffic Signal Design


    Weitere Titelangaben:

    Transp. in Dev. Econ.


    Beteiligte:


    Erscheinungsdatum :

    2020-10-06




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Assessment of Stochastic Signal Optimization Method Using Microsimulation

    Park, Byungkyu “Brian” / Rouphail, Nagui M. / Sacks, Jerome | Transportation Research Record | 2001


    Assessment of Stochastic Signal Optimization Method Using Microsimulation

    Park, B. / Rouphail, N. M. / Sacks, J. et al. | British Library Conference Proceedings | 2001



    Optimal Length of Transit Network with Traffic Performance Microsimulation

    Estrada, Miquel / Robusté, Francesc / Amat, Jordi et al. | Transportation Research Record | 2012