Optimization of traffic signal timings in road networks has been a matter of concern for traffic researchers, and various simulators have been developed. TRANSYT (Traffic Network Study Tool), especially, which was developed at the Transport Research Laboratory in the UK, is a representative simulator. TRANSYT deals with a combinatorial optimization problem in which a solution is found by searching for combinations of distributed variables. Characteristic of the problem is that the larger the road network becomes, the more the number of combinations increase, and therefore takes extra time to calculate. This paper proposes to introduce such search methods as Genetic Algorithm (GA), Simulated Annealing (SA) and Random Search (RS) to obtain the optimum solution. To verify the validity of the approach, several experiments were conducted. As a result, it was shown that these methods are superior to the conventional method.
Abstract