一种基于混合随机H2/H∞方法的能源互联网边缘计算系统控制策略

A Novel Stochastic Mixed H/H Control Strategy for Energy Internet Edge Computing System

作为一种新兴的计算范式,边缘计算可以将数据计算和处理任务下放到靠近用户的网络边缘侧执行,从而避免远距离数据传输所引发的高时延。随着边缘计算任务的增加,边缘计算中心的供能问题成为一个巨大的挑战。利用能源互联网对边缘数据中心进行供能是一个可行的方向,但可再生能源产能和边缘数据中心的耗能具有很强的不确定性和随机性。面对这一挑战,提出一种基于纳什均衡的随机最优鲁棒控制策略。该控制器通过最优控制最大化地使用可再生能源为边缘数据中心进行供能,同时也保证了边缘数据中心尽量以最大的工作功率进行数据运算。此外,该控制器还可以在有随机扰动的环境下保证母线频率偏差稳定。考虑到最优控制和鲁棒控制两个目标可能会相互冲突,该文通过纳什均衡使得二者达到一种平衡状态。最后,通过数值仿真验证了提出方法的可行性。

As an emerging computing paradigm, edge computing delegates data computing and processing tasks to the edge side close to users for execution, thereby avoiding the high latency caused by long-distance data transmission. With the increase of edge computing tasks, the problem of energy supply for edge computing centers has become a huge challenge. Using energy internet to power edge data centers is a feasible direction, but the generation of renewable energy and the energy consumption of edge data centers have strong uncertainty and randomness. Facing this challenge, this paper proposed a stochastic optimal robust control strategy based on Nash equilibrium. The controller maximizes the use of renewable energy to power edge data centers through optimal control, and also ensures that edge data centers perform data operations with the maximum possible working power. In addition, the controller also guarantees the stability of the bus frequency deviation in the presence of random disturbances from the environment. Considering that the two goals of optimal control and robust control might conflict with each other, this paper uses Nash equilibrium to achieve a balanced state between the two. Finally, this paper verifies the feasibility of the proposed method through numerical simulation.