An ant colony optimisation model for traffic counting location problem

Abstract

Strategic location of traffic counting stations across transportation road networks is theoretically a bi-objective integer optimisation problem. The problem is formulated with the two objectives considered as: (a) a maximum number of origin–destination pairs being separated and (b) a minimum number of traffic counting stations. This research proposed an ant colony optimisation algorithm to solve the traffic counting location problem by explicitly generating the Pareto solutions. Numerical results from two case studies, a small nine-node grid network and a middle-sized modified Sioux Falls network, were provided to demonstrate the feasibility of the proposed model, although the initial procedure converges too slowly towards high-quality solutions. By introducing additional suitable intensification mechanisms, the non-dominated solutions can be attained with significantly fewer iterations and ants, which consequently indicates the effectiveness of the system, and further verifies the modelling capability of ant colony optimisation. The performance of the non-dominated solutions can be further investigated and validated using the flow-capturing analysis method.