Optimization of Subway Trains Service Schedule Using NSGA-II in order to Decrease Passengers Travel Time, System Economic Loss and Energy Consumption

Scheduling vehicles’ runtime and dwell time is one important phase of designing and developing public transportation systems. Apart from trains’ headway time in subway systems, parameters such as time and rate of acceleration when trains startup or brake and also their dwell time at stations influence scheduling design criteria including passengers travel time, economic efficiency of the service and the amount of energy consumed. Since these criteria are not positively correlated, Tehran subway line no. 1 service schedule has been optimized using NSGA-II in this study. To achieve this goal, the problem was modelled first and objective functions of optimization procedure were formulated based on scheduling criteria. Then, by performing the multi-objection algorithm, solutions occurred on the Pareto Front was obtained. Since the number of solutions obtained by NSGA-II are numerous, K-Means analysis was used to cluster solutions. Centroids of each cluster could be considered as candidate solutions that one of them will be chosen by decision makers as the preferred solution. Results of the study show that using the suggested approach would output a wide variety of solutions and provide possibility of choosing the final solution based on decision preferences for the decision maker. Calculated objective functions for candidate solutions were compared with those for current subway schedule that resulted in the preferred solution to be chosen which leads to passengers travel time and economy loss of the system decreasing by 7.16 and 5.13 percent respectively.