Optimization of Large Transport Networks Using the Ant Colony Heuristic

Long-term transportation planning for larger regions must assess synergies and interference among sets of projects. The objective is the maximization of the overall benefit within specific budget restrictions by finding the most favorable bundle of projects, that is, solving the network design problem. For large numbers of projects, complete enumeration of all combinations is not feasible for detailed networks. The ant colony heuristic is suitable for this kind of problem. According to our knowledge, this article presents the above-mentioned heuristic's first application to a realistically sized network. A detailed multimodal network assignment of a substantial Swiss city provides the basis for calculations. First, each infrastructure project is assessed using a cost-benefit analysis. The ant colony heuristic is then successfully executed and the bundles are evaluated. The article provides new insights into applications of the heuristic in large networks and focuses on problematic calibration details as well as the choice of objective function. Suggestions are made for general applications and further research.