Urban public transport optimization by bus ways: a neural network-based methodology

Abstract

This paper describes an approach for planning the introduction of bus lanes into the urban road network, that has been applied to the urban area of Palermo. The proposed modelling tool adopts a multi-agent objective function expressing the trade-off between the interests of diverse stakeholders: the generalized transport cost for car drivers and the travel time for public transport users. The reaction of car traffic to a certain planning scenario has been simulated by the DUE assignment technique and the positive impact of the modal shift on the objective function has been tackled by attaching a suitable weight to the time saving for bus passengers. The rise in the bus travel speed, owing to the bus lane solution, has been predicted for a set of urban roads by a neural network, so as to take into account many quantitative and qualitative road attributes. The optimal location pattern of bus ways has been searched by a greedy heuristic that through a step-by-step strategy builds the problem solution by keeping, at each stage, the best alternative.