Configurational modelling of urban movement networks

Transportation research has usually seen road networks as inert systems to be navigated and eventually filled up by traffic. A new type of 'configurational' road network modelling, coupled to detailed studies of vehicular and pedestrian flows, has shown that road networks have a much more constructive role. They strongly influence the pattern of flows through quantifiable properties of the network 'configuration'. Recent research results are presented showing that rates of vehicular movement in road segments are to a greater extent than previously realised the direct outcome of the location of those segments in the network configuration as a whole and that this is the case especially in the fine structure of the urban grid. A supply and demand model of urban movement is proposed in which the degree to which a street alignment is on simplest routes between all other pairs of alignments in the system determines the demand side of the equation, and the effective road width available to traffic determines the supply side. Regression analysis shows that these two factors alone account for the majority of the variance in flows from street to street (r(2) similar to 0.8) A model is then proposed of the evolution of the city in allocation of land uses to land parcels, and the allocation of capacity in the road network, where each reinforces the underlaying configurational logic through a feedback 'multiplier' effect. These findings suggest the possibility of using urban design parameters, such as the plan configuration of the street grid, building height, and street width, to arrive at a better controlled relationship between vehicles and pedestrians in urban areas. As these design parameters are under the direct control of the urban master-planner, the new techniques lend themselves to application in design decision support. A case example of the application of these techniques in the master-planning of the redevelopment of London's South Bank cultural centre is presented.