Distributed networked IRL for discrete-time systems with combined communication impairments

Develop a distributed networked integral reinforcement-learning method for unknown-dynamics discrete-time nonlinear systems that jointly addresses learning, control, communication scheduling, security, channel-capacity constraints, packet losses, denial-of-service attacks, quantization, and asynchronous updates while providing a true wall-clock predefined-time guarantee.

Background

The paper identifies a gap between existing results that separately address communication dropouts, denial-of-service attacks, resilient model-free adaptive dynamic programming, quantized predefined-time coordination, and bit-rate-constrained secure control. The unresolved problem is to integrate these concerns in a single distributed networked IRL framework for discrete-time nonlinear systems with unknown dynamics.

The desired guarantee must be measured in wall-clock time rather than only over successful-transmission or attack-free intervals. Consequently, the framework should explicitly characterize the tradeoffs among the prescribed deadline, communication and bit-rate requirements, packet-loss and denial-of-service severity, quantization, and asynchronous updates.

References

Although fixed-time RL has recently been developed for uncertain discrete-time nonlinear systems, the distributed networked IRL counterpart with unknown dynamics and combined communication impairments remains largely unresolved, despite the natural relevance of discrete-time models to digital CPS implementation.

— Predefined-Time Resilient Integral Reinforcement Learning for Input-Constrained Unknown Nonlinear Systems Under FDI Attacks and Disturbances: A Fully Data-Driven Approach  (2609.11815 - Vu et al., 10 Sep 2026) in Section Discussion, subsection Scalability

What remains is a discrete-time theory that carries the temporal class without relying on small-$h$ emulation, in particular for recursive architectures, where the discretization of one virtual-control level enters the next and observer and adaptation states add further clocks; the target is a class-preserving certificate for the complete augmented loop.

— From Asymptotic to Designer-Assigned-Time Control: A Review of Stability Notions, Design Mechanisms, and Controller Architectures  (2610.11964 - Bingöl, 8 Oct 2026) in Section 5.3, paragraph “Class-preserving discretization”

A scalable theory should decide whether a common physical deadline can be retained under heterogeneous resources and, if not, how agent-specific horizons are assigned, through laws relating the gains to the algebraic connectivity $\lambda_2$ of the graph Laplacian, compositional communication-layer certificates, and conditions valid under switching or intermittent graphs.

— From Asymptotic to Designer-Assigned-Time Control: A Review of Stability Notions, Design Mechanisms, and Controller Architectures  (2610.11964 - Bingöl, 8 Oct 2026) in Section 5.3, paragraph “Scaling over networks and heterogeneous resources”