This paper investigates the joint network equilibrium of parking and travel route choices under the mixed traffic of shared-autonomous vehicles (SAVs) and private autonomous vehicles (PAVs). Specifically, we consider that some travelers travel with SAVs, and some travel with PAVs. The SAV ride-sharing service is provided by an operator, and SAV travelers do not need to concern parking. For PAV travelers, the vehicle can drive itself to park after dropping off the traveler. A bi-level model is developed which optimizes the SAV service fare and fleet flow in the upper level, and specifies travelers’ mode choice, parking location choice and routing of SAVs and PAVs in the lower level. In particular, PAV travelers choose their route and parking location to minimize their own travel time or cost, while the routing of SAVs is subject to the operation model of the operator. The SAVs controlled by the operator may minimize each vehicle’s travel time (user equilibrium, ‘UE’) or minimize the total travel time of all SAVs operated by the operator (Cournot-Nash equilibrium, ‘CN’). The joint equilibrium of travel and parking under either UE or CN routing for SAVs can be modeled as a Variational Inequalities (VI) problem. The uniqueness/non-uniqueness properties of the joint network equilibrium are investigated. Moreover, we examine the SAV ride service operator’s optimal operation strategies subject to the lower-level network equilibrium. Solution approaches are introduced to solve the joint equilibrium and the proposed bi-level model. Numerical studies are conducted to illustrate the model and analytical results, and also to provide further understanding.
 

Shared-autonomous vehicles;Private-autonomous vehicle;Parking choice;Route choice;Mixed traffic equilibrium;Operation strategies