Multi-objective Memetic Algorithm Based on NSGA-II and Simulated Annealing for Calibrating CORSIM Micro-Simulation Models of Vehicular Traffic Flow

This paper proposes a multi-objective memetic algorithm based on NSGA-II and Simulated Annealing (SA), NSGA-II-SA, for calibration of microscopic vehicular traffic flow simulation models. The NSGA-II algorithm performs a scan in the search space and obtains the Pareto front which is optimized locally with SA. The best solution of the obtained front is selected. Two CORSIM models were calibrated with the proposed NSGA-II-SA whose performance is compared with two alternative state-of-the-art algorithms, a singleobjective genetic algorithm which uses simulated annealing (GASA) and a simultaneous perturbation stochastic approximation algorithm (SPSA). The results illustrate the superiority of the NSGA-II-SA algorithm in terms of both runtime