Network models for battery electric vehicles

In this thesis a nonadditive shortest path problem to model the route choice of battery electric vehicle (BEV) drivers has been proposed. Based on this nonadditive shortest path framework several multiuser (with heterogeneous risk attitude) network models which take congestion into account have also been proposed. The proposed route choice model relaxes several assumptions of earlier literature and allows for a continuum of range limits and heterogeneous drivers who have varying risk preferences. The model also accounts for nonlinearity in travel choices -- drivers value a small amount of charge more when they are close to running out of range than when the battery is close to full charge. A nonlinear nonconvex optimization problem is formulated and an approximation of the objective function leads to a convex problem which is solved using an outer approximation algorithm. A tour-based analysis, which is more appropriate for BEVs is considered; but a network transformation makes the formulation simpler. Numerical experiments on a small network demonstrate how the routes taken by BEV drivers are influenced by their risk attitudes and the uncertainty in the predicted range of the vehicle. The models developed in this thesis are applicable to networks with flows of BEVs. This work will hopefully inspire researchers to explore nonlinear travel models for BEVs and develop more general network models. These network models using survey data (extensive surveys will need to be carried out for this) will be able to predict system-wide effects of the choices made by BEV drivers and help planners and policy makers in their decision making.