Thus said email@example.com (Charlie Thorne): >The problem with the current system of scheduling buses is that it >does not allow for any changes in traffic or ridership. The >assumption is that there will be a certain number of people waiting at >stops along the route and that traffic will impede buses and thus a >certain number (such as one every five minutes) is required to get the >passengers to their destination along the route. > >Obviously, as we know, traffic and numbers of people waiting to board >disrupt the schedule. To correct the problem, there needs to be >flexibility in the schedule. One lack of flexibility is the usual >refusal to account for the fact that buses arriving in the rear which >are relatively empty, and thus contribute nothing to the >transportation of those waiting. > >Having three buses coming along, with only the first one carrying >passengers if an obvious inefficiency. To correct for this (which >happens quite often) one needs to allow the rearmost (empty buses) to >pass the fuller buses in front. Transportation logistics (or what is >thought to be logistics) doesn't permit this for a number of reasons. >The ability to pass the loaded buses is an obvious correction of the >lack of flexibility--but is apparently used almost in no city.
This happens often in many cities, as other posters to this thread have observed. Here in Seattle it happens all the time. Generally it can only work on a bus that is in "accumulation phase", i.e. one that is having many boardings but few alightings. In the reverse case, bunching is almost irrelevant.
><KdmlivesREMOVE@hotmail.com> wrote: >>Marc VanHeyningen wrote: >>> It means instead of having to worry that >>> heavy traffic might make your bus ride take forever, you will know for >>> sure that your bus ride will take forever waiting at timepoints no >>> matter what the traffic is like. >>> >>> Is that really an improvement? >> >>Well, the timing would be adjusted so for maximum >>efficiency. The MTA already has people out there >>observing the buses and recording the times they hit certain >>stops. Simply feed this info into a computer, and -Bing!- >>out pops a schedule that wastes as little time as possible.
So, what happens when something unusual happens? For example, what if it rains? That can change traffic patterns, and also may increase the number of riders. I'm sure estimation of traffic delays can be done more accurately than many agencies do it, but it will still fall across a distribution. Where in that distribution do you put the schedule? In the middle? One standard deviation before the middle?
>>Even if I am simplifying things a bit, there simply HAS to >>be a better way to schedule buses than they way they do now.