Multi-Objective Provisioning of Network Slices using Deep Reinforcement Learning

TL;DR

This paper introduces a deep reinforcement learning approach to real-time network slice provisioning, effectively handling complex QoS requirements and reducing costs and SLA violations in next-generation networks.

Contribution

It models network slice provisioning as an online multi-objective optimization problem and applies PPO for real-time traffic demand prediction, outperforming traditional solvers.

Findings

Lower SLA violation rate compared to state-of-the-art methods

Reduced network operation costs

Effective real-time provisioning under complex QoS constraints

Abstract

Network Slicing (NS) is crucial for efficiently enabling divergent network applications in next generation networks. Nonetheless, the complex Quality of Service (QoS) requirements and diverse heterogeneity in network services entails high computational time for Network Slice Provisioning (NSP) optimization. The legacy optimization methods are challenging to meet the low latency and high reliability of network applications. To this end, we model the real-time NSP as an Online Network Slice Provisioning (ONSP) problem. Specifically, we formulate the ONSP problem as an online Multi-Objective Integer Programming Optimization (MOIPO) problem. Then, we approximate the solution of the MOIPO problem by applying the Proximal Policy Optimization (PPO) method to the traffic demand prediction. Our simulation results show the effectiveness of the proposed method compared to the state-of-the-art MOIPO solvers with a lower SLA violation rate and network operation cost.