Transportation network design methods under the assumption of elastic demand

Theoretically speaking, in any supply design model the transportation demand should be assumed elastic because supply changes generally influence at least mode choice. Sometimes, the errors connected with the assumption of rigid demand are negligible, while in other cases demand cannot be assumed as rigid, especially if the design objective is to change the modal split (Travel Demand Management) or when transportation systems share elements of the network (buses and cars that use the same lanes).Nevertheless, in the literature monomodal design models, which analyse only one transportation system and consider everything else invariable, are fully discussed both for road and transit systems. However, these models are useful when projects chiefly affect the analysed system without major repercussions on other systems.As highlighted above, a multimodal approach, which analyses several modes simultaneously, is required when developing plurimodal projects (combined design) or monomodal projects with major effects on other modes (for example the pricing). This approach has received little attention due to computational difficulties inherent in the problem.In this paper, we proposed two multimodal network design models that take into account the following aspects: the modal split is elastic and therefore depends on network costs; the road link costs depend on the number of (road and transit) vehicles on the link; the transit link costs are constant in exclusive bus lanes and depend on the number of (road and transit) vehicles on the link in the case of shared lanes.The first model allows to determine optimal transit fare; it is based on the optimisation of an objective function that estimates explicitly operational costs, traffic revenues, system user generalized costs and social, political and environmental external costs.The second model concerns the optimisation of parking prices in urban networks. In particular, almost all the parking pricing strategies usually adopted are (destination) area-based (D policy), that is parking charges vary with the (destination) zone: this means that the user pays only according to trip destination. Instead, this paper proposes a new parking fare strategy based on both trip origin and destination (OD policy). The main advantage of the proposed policy is that parking fares can be calibrated so as to "penalise" only the users that have an efficient transit system on some OD pairs; it is thus possible to "force" the modal split, limiting the decrease in accessibility; this "penalisation" seems correct also from a political point of view, since only part of the costs of transit systems are covered by users (tickets), while most are subsidised by society.In conclusion, initial results of multimodal approach applications for solving fare design problems show the validity and functionality of the proposed models and algorithms. Finally, research perspectives could be addressed to analyse other multimodal problems, combine both proposed problems with the road-pricing problem and propose more effective solution algorithms.