This work presents a stochastic control formulation for stability in multi-layer logistics systems. The system is modeled as a high-dimensional discrete-time dynamical process with stochastic perturbations and cross-layer coupling. We define a control objective based on minimizing expected deviation from a stability region in state space, replacing traditional decentralized optimization of individual subsystems. The proposed approach incorporates covariance-weighted loss functions and finite-horizon control. A theoretical stability result is provided under standard boundedness assumptions. Simulation results demonstrate reduced variance and suppression of oscillatory dynamics compared to reactive control baselines. This work contributes to the study of complex operational systems, including logistics networks, distributed supply chains, and cyber-physical infrastructure.
Dmitry Chistyakov (Fri,) studied this question.