An integer timetabling and vehicle scheduling problem for two-way point-to-point ground transportation system with considerations for headway and number of vehicles to be deployed

• Main Authors: Santos, Charlynne, Tee, Michelle, Trinidad, Feliza
• Format: text
• Language: English
• Published: Animo Repository 2019
• Subjects: Transportation; Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/9356
• Institution: De La Salle University

The use of two-stop or point-to-point (P2P) ground transportation systems is growing significantly in the recent years due to the convenience and efficiency that it provides compared to the usual mode of transportation, which is the multiple stop system. In planning systems like these, there are usually four stages: line planning, timetabling, vehicle scheduling, and crew scheduling. This study will be focusing on integrating timetabling and vehicle scheduling. Timetabling provides the set of schedules in the system while vehicle scheduling assigns which vehicles will service the trips that was provided in the timetabling. The two planning stages are often done separately or via sequential approaches, since timetabling focuses on the customer's perspective while vehicle scheduling focuses on the operator's perspective. In integrating both studies, the model was able to fully utilize the system. Furthermore, in incorporating the integration to a point-to-point transportation, the model was able to capture the trade-offs that are present in the system. The study considered multiple periods with uneven headways and varying travel times in a roundtrip point-to-point ground transportation system. In the model, the objective functions are to minimize the passenger waiting time and the total operating costs, wherein the waiting time is on the basis of the first person to arrive in each trip. The major constraints consist of period, dispatching, vehicle capacity, last trip, vehicle assignment, forward time progression and waiting time. The model was run with the GAMS software with a mixed integer nonlinear (MINLP) approach and an incorporation of a non traditional goal-programming technique to properly integrate the two objectives. Goal programming was chosen in order to generate the best value for both timetabling and vehicle scheduling simultaneously through a formula that takes into consideration the ratio between the resulting slack variables and the minimum amount of waiting time and cost. Two initial runs were done in order to compute for the minimum waiting time and cost which were taken as an input to the integrated model. The objective of the integrated model is to minimize the maximum between the two ratios of the slack variables with the minimum waiting time and with the minimum cost. There were a total of five scenarios tested in order to see its impact on the waiting times and the assignment of vehicles. Based on the results of the tests done, the inter-arrival times of passengers per period, uneven headways, varying travel time and the number of vehicles assigned in the system are significant to the resulting waiting time of passengers and the operating cost. In incorporating the peak and non-peak periods in these factors, the model was able to properly utilize the available vehicles in system, and provided appropriate values of headway depending on the demand and period, which ultimately minimized the number of trips that the system can accommodate, which is beneficial to both the passengers and operator. Overall, the results of the analysis showed that the number of vehicles in the system and the values of the headway together affect the resulting number of trips and the average number of passengers in the system. Since the number of passengers is the direct result of the values of the headway, it could be used as the basis in knowing whether to add another vehicle in the system. Lastly, the results of the sensitivity analysis showed that the model was able to capture the trade-offs that are present in point-to-point systems wherein solving both problems separately or sequentially cannot provide and ultimately, provides best result in the system given the parameters and constraints that was identified.