OPTIMAL PLANNING OF MULTIMODAL FREIGHT TRANSPORT NETWORK

Designing a multimodal freight transport network is essential especially for regions and countries where lack of attention is given on the planning of multimodal transport systems. This paper proposes a model that can be used as a tool for strategic level of multimodal transport planning, particularly in freight terminal development and freight transport network design. The model determines an optimal freight transport network expansion plan, which selects an optimal set of actions from a number of possible actions.The model is developed within the framework of bi-level programming problem, where a multimodal multi-user assignment technique is incorporated within the lower level problem that explicitly takes into account passenger and freight flows, and the various activities within the terminals. The upper level problem optimises the combination of actions such that the freight-related benefit-cost ratio is maximised. The feasible actions include improving the existing infrastructure as well as establishing new roads, sea links and freight terminals. This type of problem can be considered as a combinatorial optimisation problem, where evolutionary computation techniques based on genetic algorithms can be applied as solution procedures.The model is tested on the freight transport network in Java Island, Indonesia, where freight transport network development is desired to increase the utilisation of other modes rather than road-based vehicles. Results revealed that the procedure based on Genetic Local Search could provide better performance and could adequately find the best combination of actions among available alternatives.