Martino Bardi, Hicham Kouhkouh
Published 2022-08-01Version 1
We study singular perturbations of a class of two-scale stochastic control systems with unbounded data. The assumptions are designed to cover some relaxation problems for deep neural networks. We construct effective Hamiltonian and initial data and prove the convergence of the value function to the solution of a limit (effective) Cauchy problem for a parabolic equation of HJB type. We use methods of probability, viscosity solutions and homogenization.
Stochastic optimal control in Hilbert spaces: $C^{1,1}$ regularity of the value function and optimal synthesis via viscosity solutions
Stochastic Optimal Control for Systems with Drifts of Bounded Variation: A Maximum Principle Approach
Modelling information flow in stochastic optimal control: How Meyer-$σ$-fields settle the clash between exogenous and endogenous jumps
![QR code for arXiv:2208.00655 [math.OC]](http://oss.arxitics.com/images/favicon.svg)