Bus public transport algorithm

Question

I am working on an offline C# application that can find bus routes. I can extract the timetable/bus/route data. I am searching for the most simple solution that will work with

Answer 1

Just wanted to share my final approach on this problem. This was part of a university project, so it might not be fully suitable for real world use. It had to be reasonably fast to run on a Windows Mobile device.

I ended up using 4 tables (SQLite). One table keeps the list of buses, the second one keeps a list of stations. Another table keeps the combinations - what bus does stop at a specific station and where does it go from this station and how long (minutes) does it take. All combinations have to be stored. And the last table is a simple time-table. For every station it lists every bus that stops there and the time (I stored the time as integer value - 14:34 is 1434, to make it faster for comparing).

I used a bi-directional breadth first search algorithm. The neighbor stations (directly accessible) are retrieved for the start station and destination station. There is a path from station A to station X if these two "graphs" overlap on a station. For example, from station A you can get to stations B,C,D,E (by using specific buses). And from the destination station X you can get directly to N,C,Z. These two graphs overlap on station C. So a path A -> C -> X exists. If no paths are found in this first iteration, then the algorithm continues and spreads out the graphs (BFS) again.

Time is not taken into account in the first step - this makes it fast enough. You only get a list of possible paths with a list of buses that you have to use to take these paths. The times are evaluated in the last step, you go through the list of possible paths and check if the bus travels within the specific time (increasing the time every stop).

On a small city, with 250 stations and 100+ buses/railways, these approach works up to 3 changes (where you have to change the buses on the journey). It takes only seconds to calculate. But I had to load the whole database into memory (Dictionary) to speed up the queries, which were taking too long.

I don't think this would work for a large network though. But it works for a public transport of a single small-medium size city.