Public Transport Optimization

Public transport systems may suffer from two kinds of inefficiencies. They may be overloaded, leading to loss of income; or they may be underutilized, leading to loss of efficiency. Ironically, many-a-times, public transport systems are both, simultaneously.

If we could match unmet need to unutilized capacity, we may overcome both types of inefficiencies for free!

Discovering Unmet Need

Public transport systems generate large amounts of data. We can utilize this data to understand and predict demand for public transport. This is a hard data science problem, for many reasons:

Demand for public transport goes up and down depending on factors like trip time, cost, time of day, day of week, tourist season, etc.

Schedules keep changing in planned and unplanned ways.

People take connections or “break journeys”. Unless we have identity of each person, it is not easy to understand connection patterns. On the other hand, it is important to understand this, because the actual registered demand is for the end-to-end journey, not for the connection points.

Many people who are not currently using public transport, or using a different system, may switch over to using our system if certain new routes are started.

Various data science techniques may be used to build demand models despite the above problems. We have built such public transport demand models using statistical and data science expertise.

Rerouting Unutilized Capacity

Once we have a model for public transport demand, we can apply combinatorial optimization to achieve substantial efficiencies in public transport: we may add completely new routes, remove old ones, change schedules, change the mix of vehicles on road (big bus, little bus) etc. Though this is a hard problem, we have the technology expertise to do optimizations in reasonably short periods of time, allowing the transportation authority to make rapid course corrections.