The development of a mathematical programming technique as a design tool for traffic management
In urban areas, competition for road space at junctions is one of the major causes of congestion and accidents. Routes chosen to avoid conflict at junctions have a mutually beneficial effect which should improve circulation and reduce accidents. A prototype design tool has been developed to provide for traffic management based on such routes. The mathematical model behind the design tool works with a given road network and a given O-D demand matrix to produce feasible routes for all drivers in such a way that the weighted sum of potential conflicts is minimised. The result is a route selection in which all journeys from origin i to destination j follow the same route. The method which works best splits the problem into single commodity problems and solves these repeatedly by the Out-of-Kilter algorithm. Good locally optimal solutions can be produced by this method, even though global optimality cannot be guaranteed. Software for a microcomputer presented here as part of the design tool is capable of solving problems on realistic networks in a reasonable time. This method is embedded in a suite of computer programs which makes the input and output straightforward. Used as a design tool in the early stages of network design it gives a network-wide view of the possibilities for reducing conflict and indicates a coherent set of traffic management measures. The ideal measure would be automatic route guidance, such as the pilot scheme currently being developed for London. Other measures include a set of one-way streets and banned turns. The resulting turning flows could be used as input to the signal optimiser TRANSYT to determine signal settings favouring the routeing pattern. The project was funded by the S. E. R. C. and carried out at Middlesex Polytechnic in collaboration with MVA Systematica.