An Efficient Slicing Method in Integrated Network of Synaptic Computing

In the era of rapid technological progress, people have made great achievements in the construction of 5G networks. In the next decade, the integration of the digital and physical world in all dimensions is one of the themes of 6G network development. In this context, holography-type communication (HTC) presents exciting educational application prospects. The application of HTC in the field of training and education can offer students a holographic classroom experience without space restrictions, redefining learning environments, enabling automated management, and providing immersive learning experiences. However, HTC has high requirements for bandwidth, latency, and computing power, which traditional networks such as 5G cannot fully satisfy. This paper proposes a dynamic slicing algorithm based on multi-agent reinforcement learning to realize an integrated network of sensing communication and computing. The network slicing technology is used to slice according to channel bandwidth, computing resources, and cache resources, and deep reinforcement learning is used to dynamically adjust resources to meet the changes of different application requirements, thereby improving network performance and utilization. The results show that the proposed scheme can effectively meet the diverse Quality of Service (QoS) requirements and provide a new solution for the implementation of HTC.