Minimizing the traffic impact caused by infrastructure maintenance using ant colony optimization

Maintenance measures at infrastructural buildings such as bridges or tunnels in urban areas always have influence on the traffic flow. Finding a schedule that keeps this impact as low as possible by considering the mutual interaction between several road closures is a highly complex optimization problem. In a current research project we have developed an optimization tool to solve this problem based on ant colony optimization. Over a number of iterations teams of ants construct maintenance schedules for the next few years based on information about the quality of the schedules found in the last iteration steps as well as general information about the condition of the different bridges. The quality of the found schedules is evaluated in the external traffic simulator VISUM. For each year of the schedule a disturbed road network (with reduced capacity at the roads containing bridges under maintenance) is created from the original network and evaluated. As traffic demand and its temporal distribution stay the same for the disturbed network as for the undisturbed, the total net number of vehicle hours during rush hour is a good measurement of the impact of the maintenance activities on traffic. The optimization goal is to minimize this number for the worst year in the schedule.