TrafficInfraTech Magazine Linking People Places & Progress
Increasing Revenues for an Indian bus operator
In the next case study, we look at the case of an Indian city which has a bus network that connects various parts of the city. It has around 5 routes covering the city and around 10 buses are available to run on a route. The ticketing is done through a ticketing machine that is carried on buses with the conductors during a trip.
Objective: On any given day, the transit agency has only a few buses of the 10 allocated, running on a route. The reason for this is that buses could be under repair or the crew could be absent from work. The schedules need to adapt to the number of buses available so as to maximize the revenue collected by running the trips on any particular day. The constraints on these schedules also include the conditions on the crew including an equal distribution of driving hours between the crew and time window for having lunch.
Data: The data available was the ticket swipes recorded for commuters using the transit network. There was three months ticketing data available through the ticketing solution. Apart from this, the existing schedules, information on number of buses and the capacity of bus were also provided.
Methodology: The methodology used for this case was also the same as the previous case study. The only difference between the two was that in the current case, the objective to maximize was the revenue. There were constraints also added to the scheduling engine, which accounted for the crew scheduling as mentioned before. Using the demand patterns learned from the ticketing data, the schedules are produced for different cases of vehicle availability. There would be a schedule when there are 10 buses available, another schedule when nine buses are available and so on. The schedules are created such that they would meet the most amount of demand using just the number of buses available on a particular day.
Results: The evaluation was done following the same methodology as in the previous case. The first month of demand data was used to build the set of schedules and the data from the later months was used to simulate the schedule and measure the revenue earned. This was compared to the revenue earned from the current schedule. The results of the comparison is given in Figure 1.
The comparisons show the demand satisfied by the current approach and the schedules produced by our system labeled as recommended. The revenue comparison is done across cases where the number of vehicles available are different. We can see in each of these cases, the recommended set of schedules is able to earn more revenue than the current schedules. This is to be expected since the current methodology of scheduling does not adapt to the number of vehicles available. On the other hand, using the methodology explained here, different schedules are used when different number of vehicles is available and results in meeting more of the demand.
Reliability and punctuality of public transit are key drivers encouraging customer satisfaction and patronage. As we can see above, demand sensitive schedules can significantly improve bus utilization and revenues, while maintaining costs. Combined with pricing strategies that can be dynamic balanced with commuter satisfaction, public transit agencies can improve their overall service quality and meet customer demand.
