Reduced headways of connected and automated vehicles (CAV) provide opportunities to address traffic congestion and environmental adversities. This benefit can be utilized by deploying CAV-dedicated lanes (CAVDL). This paper presents a bi-level optimization model that captures CAV market size uncertainty. The upper level determines the links (and number of lanes) for CAVDL deployment to minimize emissions. It considers lane reallocation policies that account for the prospect of smaller width of CAV-dedicated lane due to the smaller lateral wander of CAV tire tracks. This can increase the total number of lanes on wide highway sections. At the lower level, equilibrium and demand diffusion models capture travelers’ route and vehicle-type choices. The bi-level model is formulated as a min–max mathematical program with equilibrium conditions and solved using the cutting-plane scheme and active-set algorithm. The computational experiments indicate that the robust plans have superior performance compared to the deterministic plan in pessimistic cases.

The rising demand for electric vehicles (EVs), motivated by their environmental benefits, is generating an increased need for EV charging infrastructure. Also, it has been recognized that the adequacy of such infrastructure helps promote EV use. Therefore, to facilitate EV adoption, governments seek guidance on continued investments in EV charging infrastructure development. Such investment decisions, which include EV charging station locations and capacities, and the timing of such investments require robust estimates of future travel demand and EV battery range constraints. This paper develops and implements a framework to establish an optimal schedule and locations for new charging stations and decommissioning gasoline refueling stations over a long-term planning horizon, considering the uncertainty in future travel demand forecasts and the driving range heterogeneity of EVs. A robust mathematical model is proposed to solve the problem by minimizing not only the worst-case total system travel cost but also the total penalty for unused capacities of charging stations. This study uses an adaptation of the cutting-plane method to solve the proposed model. Based on two key decision criteria (travelers' cost and charging supply sufficiency), the results indicate that the robust scheme outperforms its deterministic counterpart.