# A little more on public transport

There are three factors that affect the choice of mode of transport – cost, time and convenience. You usually make the decision regarding what mode to take by evaluating how much you are willing to “spend” in terms of each of these. For example, if you live in Mumbai, taking the local train would be the most effective in terms of cost and time, but not really so in terms of convenience. Hence, you may sometimes give up time and cost and invest in a taxi so that you travel in better convenience.

This needs to be kept in mind when trying to design an efficient public transport system – the combination of cost, time and convenience needs to be attractive for the user to abandon private transport in favour of public. One mistake that is commonly made is to forget the convenience part, which leads to a large number of people to stick to their private transport. In related stuff, in my first ever post on the IEB, I had argued for segmented public transport so that people can find their own combinations within the public transport system.

This post is about the time aspect. The thing with a system based on buses is that you can do little about the time factor – unless you decide to create specific bus lanes, which aren’t too feasible currently in a place like Bangalore, you need to use the general road space. However, this is only as far as traveling time goes. What can, and should, be improved is waiting time.

Last week, while I was on my way to the M G Road area to meet sudheernarayan for the unofficial photowalk, I purchased a daily pass (for Rs. 30 you can travel on any bus from anywhere to anywhere else on that day). Though there exist direct buses from my place to MG Road, I figured out that it would be faster if I changed buses. Basic assumption was monotonicity – if bus A and bus B both go from point X to point Y, and if A reaches X before B, then A will reach Y before B. And that every time I went forward a small distance, that would increase my chances of getting a bus to the final destination.

So I take a random bus from my place to Banashankari – which improves the chances of getting a bus to MG Road by a factor of three. Then another 2 buses to fourth block – increases probability of getting bus by about 50%. By probability I mean the expected value of getting a bus to your destination in the next one second (I mean, assume Poisson distribution, etc.). I took four buses to get to M G Road but spent a total of five minutes waiting.

Now that the story is out of the way, the key to a good bus system would be to decrease waiting time. Right now, it’s quite horrible in Bangalore on a number of routes, as frequency of direct buses is low and changing is also quite difficult. And as for changing buses, there should be a system which makes interchanges easier – typically you should be able to switch buses by standing at the same bus stop.

Right now I need to go down to watch football, so I’ll write the second part of this post tomorrow. In that I’ll talk about how the network can be redesigned so as to decrease average waiting time for commuters. Of course, given the volumes and the fleet in operation today, one will need to introduce several more buses before the “convenience” bit becomes better.