Autonomous vehicle (AV) ride-hailing fleets are no longer hypothetical: Waymo, Uber, and competing operators are deploying mixed AV and human-driven fleets across U.S. cities. State and local agencies must now decide which operating rules to impose, such as permit caps, congestion pricing zones, and limits on empty repositioning trips, without analytical tools that capture how multiple competing fleet operators actually respond to those rules and to one another. Existing models treat AV fleets as if they were human-driven fleets with different cost structures, missing the central control, repositioning capability, and regulatory susceptibility that distinguish AV operations. This project develops a two-level modeling framework for operational constraints and regulatory design in mixed AV and human-driven ride-hailing networks. The lower level represents the strategic interaction among competing fleet operators, travelers, and the road network, extending the team's prior network equilibrium model to include empty-vehicle repositioning that loads network links and human drivers who work for multiple platforms simultaneously. The upper level represents a regulator selecting permit caps, congestion-pricing tolls, and repositioning limits to optimize a system-level objective such as total travel time or system congestion. The team will develop solution methods, release an open-source solver, and apply the framework to representative California networks. The work will give state and local agencies a transparent, defensible tool to evaluate regulatory levers before deployment.