Sag vertical curves (sags) are roadway sections along which the gradient increases gradually in the direction of traffic. Empirical observations show that, on freeways, traffic congestion often occurs at sags; actually, in some countries (e.g., Japan), sags are one of the most common types of freeway bottleneck. This thesis is the outcome of a research project funded by Toyota whose objective was to develop new concepts for traffic management at freeway sags based on a thorough understanding of the causes of congestion. The main scientific contributions of the thesis belong to the fields of traffic flow theory, modeling and control. The thesis contributes to the understanding of the causes of congestion at sags by analyzing empirical trajectory data and showing that the main cause is related to the fact that most drivers do not accelerate enough when they drive through this type of freeway sections. Also, the thesis contributes to the theoretical understanding and mathematical description of driving behavior on hilly freeways by proposing a new model that describes car-following behavior at sags more accurately than existing models. Furthermore, the thesis contributes to the field of traffic management by identifying new strategies to mitigate congestion at freeway sags. First, the thesis determines how vehicles equipped with in-car systems should ideally move at sags in order to minimize congestion by using an optimal control-based method. Second, the thesis proves the potential of mainline traffic flow control measures based on the use of variable speed limits by means of traffic simulation experiments. The findings and tools presented in the thesis can support policy makers, road authorities, traffic engineering and automotive companies in developing and evaluating specific traffic management measures to mitigate congestion at freeway sags.