Coexisting Automated and Human-Driven Vehicles: Well-Posedness of a Mixed Nonlocal-Local Traffic Model
Rinaldo M. Colombo, Mauro Garavello, Claudia Nocita
TL;DR
The paper addresses a macroscopic traffic model where nonlocal autonomous-vehicle dynamics coexist with local, conventional vehicles, yielding a coupled system of integro-differential PDEs for densities $\rho^i$ and $r$. By decomposing the analysis into a nonlocal subsystem and a local Kruzhkov problem, the authors construct a global semigroup on $\mathcal{BV}^1(\mathbb{R};\mathbb{R}^k) \times \mathcal{BV}(\mathbb{R})$, proving global well-posedness and 1st-order stability with respect to initial data. They establish regularity, positivity, mass conservation, and Lipschitz dependence on initial data, using a recursive approximation scheme that couples nonlocal velocity laws $v_{NL}^i$ and local velocity $v_L$, along with a convolution kernel $\eta$. The results provide a robust analytical foundation for mixed traffic models with coexisting vehicle classes and highlight the nuanced regularity requirements for nonlocal versus local components, enabling potential extensions to nonlocal-to-local limits and higher-population settings. The work advances the mathematical understanding of mixed nonlocal-local traffic dynamics and offers a rigorous framework for future control and optimization studies in heterogeneous traffic scenarios.
Abstract
We present a macroscopic traffic flow model where standard vehicles coexist with vehicles informed on the traffic distribution. The resulting mixed nonlocal-local integro-differential PDEs is proved to generate a locally Lipschitz continuous semigroup whose orbits are uniquely characterized as solutions to the system, according to a natural definition of solution. The norms and function spaces adopted are intrinsic to the different nature of the equations.
