Strategy optimization and evolution analysis for congestion alleviation in deterministic networks.

The surge in variety and quantity of B5G/6G network services will lead to serious traffic aggregation and network congestion, which is difficult to meet the deterministic requirements of time-sensitive services. Dealing with the congestion alleviation problem in deterministic networks, the real-time influence of network node congestion states on the packet forwarding behaviours is fully considered in this paper from a macro perspective. A time-varying continuous congestion alleviation system model is proposed, and an optimization problem to jointly minimize the network delay, strategy cost and congestion degree is formulated. The optimal strategies and congestion states of network nodes are solved based on the Hamilton variational method, based on which a deterministic congestion alleviation (DCA) algorithm is proposed to obtain the analytical solution to the optimal strategies. Simulation results indicate that, by means of the optimal DCA strategies, the congestion degree of each network node is gradually reduced. Moreover, compared with the random strategies, the total loss of both the whole network and each node of the proposed model is lower, so as to minimize the delay under the lowest strategy cost, and effectively alleviate the network congestion degree.