On Mon, 01 Sep 2003 21:39:02 GMT, "Richard J."
wrote:

The most elementary study of traffic jams will show that tiny events (such
as a car parking for 10 minutes in a bus lane, or a side road temporarily
blocked by a large truck reversing into a narrow entrance) can have
unforeseen effects on traffic flow. The idea that it is all predictable
apart from the occasional accident is not realistic. If it's all so simple,
why have we had bunching problems for the last 50+ years?

Because...

1) Off-bus ticketing, the majority holding passes, boarding at the
back while people pay at the front etc. are more or less unknown in
this country - and are only just appearing in London. This is what
causes the extensive delays at stops, day in, day out, which causes
day-to-day bunching - and is in particular what causes bunching to get
so bad after what seems to be a minor delay to start with.

2) Privatised bus companies are in it for the money, and produce
unrealistic, too-tight timetables which will always lead to bunching.
Admittedly this has only been the case since the mid-1980s, but it is
very relevant.

You'll not get 100% of bunching by addressing those two issues, but it
will certainly help. You're right in that not all traffic congestion
can be foreseen - but, especially on the approach to cities in the
morning peak - so much of it is, and is not adequately planned for.

I think it's not so much a case that solving bunching (in many, but
admittedly not all, cases) is an impossible task. It just costs
money, which is something that is scarce in the PT world - certainly
the bus world - with the possible exception of London where these (and
other) issues *are* being addressed to an extent.

The unforeseen bunching can be helped, as well - you'll likely not get
the offending bus back on time, but if you've got enough spare
vehicles in the correct places and a decent overall control setup
(again, these cost), an extra could be dropped in so that bus could
run more or less non-stop to its destination without delaying those in
front as was being suggested.

Neil

--
Neil Williams
is a valid email address, but is sent to /dev/null.
Try my first name at the above domain instead if you want to e-mail me.