Traffic simulation models have been widely used in supporting decision making process byassessing potential impacts of new operational strategies, physical investments, etc. Studies haveshown that simulation models need to be well calibrated to ensure the findings from thesimulation model are reliable.While the previously developed procedure by authors well demonstrated that microscopictraffic simulation models can be properly calibrated and validated using the systematicprocedure, the procedure required basic knowledge in statistical analyses and the use of geneticalgorithm-based optimization method. This was in part why the procedure was mostly adoptedby researchers and was not widely used by practitioners. In this paper, a simplified procedurewas developed and tested for more practical applications. The proposed procedure does notrequire the genetic algorithm (GA)-based optimization.Two case studies dealing with the calibration of an urban signalized corridor and afreeway merge section in VISSIM simulation software showed that the proposed simplifiedprocedure outperforms the previous GA-based calibration procedure. Based on the fitness valuesindicating the quality of calibrated parameter set, the proposed procedure produced morepromising parameter sets having better fitness value than that obtained by the GA-basedprocedure. In addition, while the GA-based procedure produced a single optimal solution, theproposed procedure was able to generate multiple optimal solutions outperforming the GA-basedsolution.
展开▼