This week's New Scientist throws down an interesting challenge to transport
modellers in its cover feature article "Well Connected", describing the
Space Syntax method.
The article implicitly criticises conventional engineers' techniques - based
on "hugely expensive surveys", gravity models and "fudge factors" - which
often lead to unreliable forecasts for traffic in urban areas.
In contrast, space syntax uses only network parameters - the spatial
arrangement of the network, sometimes in combination with road width - using
no OD type data at all. This method has been used to explain both traffic
'generation' - of vehicular and pedestrian movement - and degeneration. For
example, it is claimed that the method can account for "more than 80% of the
variation in traffic from street to street".
It would be interesting to learn of traffic modellers' opinions of this
method and the implications of its potential use. Is it fundamentally
irreconcilable with conventional approaches, or can the two approaches learn
something from each other?
stephen marshall
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|