Genetic algorithms for urban network design

In this paper an urban network design procedure based on the use of genetic algorithms is presented. The mathematical model adopted uses discrete variables to define the links direction and links capacities and continue variables to define the signal setting parameters in the junctions. The objective function is a linear combination of user’s cost (total time spent in the network) and community cost (polluting emission, fuel consumption). The problem is solved by using genetic algorithms. The genetic algorithms quickly find the sub optimal solutions of combinatorial problem through a search procedure which is only partly casual. The signal setting in the junctions is calculated through an iterative procedure based on the procedure of flow assignment on the network. A numerical application in a test network is also presented. Traffic flow allocation algorithm.