jun_pl_56 wrote:I suspect your 1 second will be swallowed up by the network without leaving anything for the routing.
About Dijkstra, I will give it a try, but I'll have to modify it, as i have to take into account the waiting time between two changes, and also limit the number of changes
(because I'm not sure someone would like to change bus 5 times to go from A to B).
What do you mean exactly by pretty fast? For me it is important to do everything within a second (sending the request to server, processing the request, sending the K shortest paths back, and displaying them on the map), with a good internet connexion obviously.
About walking you're right, it might be smarter to differentiate it. I'll investigate it later and let you know the results.As I see it, there are at least three parts to the dynamics. First one has the slow moving dynamics caused by the way the timetable changes during the day/week, second one has the statistical dynamics caused by local traffic conditions and finally one has the dynamics associated with person making the journey (how fast he/she walks).
The problem is not as simple as it sounds, as it implies dynamics graphs, multi-criteria searchs and some extra issues proper to public transportation.