This paper presents a dynamic simulation based on a combination of Adaptive Neuro-Fuzzy Inference System (ANFIS) and Dijkstra's algorithm (named ANFIS-Dijkstra's model), to design and simulate a dynamic suburban public transportation system in Iran. To evaluate the proposed suburban public transportation system, two simulation methods related to the current system and the proposed ANFIS-Dijkstra's model are designed and implemented. To make the simulator closer to the reality, a machine learning technique based on ANFIS is used for route traffic estimation, while and Fuzzy C-Means (FCM) clustering algorithm is utilized to determine the number of passengers. Then, Dijkstra's algorithm is applied for the optimal routing of passengers from the source to the destination. Finally, the results of the two simulation models are compared to justify the performance of the proposed ANFIS-Dijkstra's model. Based on the results, the rate of using bus compared to taxi increased by 8.2%, the rate of using bus capacity increased by 12.66%, and the total fare paid by passengers decreased.
No takes yet. Share an insight, caveat, or question.
A 2021 study studied this question.