Planning Cyclic Queuing and Forwarding for Time-Sensitive Networking via Heuristic and Neurodynamic Optimization

Time-sensitive networking (TSN) is a potential means to provide high-quality communication for cyber-physical systems in industrial automation and manufacturing. IEEE 802.1Qch cyclic queuing and forwarding (CQF) is a mechanism for transmitting time-sensitive packets in TSN systems. This paper addresses the planning of periodic time-sensitive flows for CQF in TSN systems. An integer programming problem is formulated to plan the periodic time-sensitive flows. In view of the NP-hardness of the integer programming problem, a polynomial-time heuristic algorithm is developed for CQF planning based on flow efficiency sorting and packet swapping to compute time offsets of multi-period flows. The heuristic algorithm is proven to converge to a CQF plan with a constant worst-case performance ratio with respect to the global optimal plan under given assumptions and derived sufficient conditions. Based on the heuristic solution, neurodynamic optimization is used to search for global optimal plans. Experimental and simulation results on open-loop network flow planning and closed-loop networked model predictive control are elaborated to substantiate the efficacy of the planning algorithms.