Controlled Deep Reinforcement Learning for Optimized Slice Placement

TL;DR

This paper introduces HA-DRL, a hybrid approach combining deep reinforcement learning and heuristics to improve network slice placement, achieving faster learning and higher acceptance ratios than existing methods.

Contribution

The paper proposes a novel hybrid ML-heuristic method, HA-DRL, that enhances DRL for slice placement by integrating heuristic guidance to improve exploration and efficiency.

Findings

HA-DRL accelerates learning of slice placement policies.

HA-DRL improves slice acceptance ratio.

HA-DRL outperforms state-of-the-art RL-based approaches.

Abstract

We present a hybrid ML-heuristic approach that we name "Heuristically Assisted Deep Reinforcement Learning (HA-DRL)" to solve the problem of Network Slice Placement Optimization. The proposed approach leverages recent works on Deep Reinforcement Learning (DRL) for slice placement and Virtual Network Embedding (VNE) and uses a heuristic function to optimize the exploration of the action space by giving priority to reliable actions indicated by an efficient heuristic algorithm. The evaluation results show that the proposed HA-DRL algorithm can accelerate the learning of an efficient slice placement policy improving slice acceptance ratio when compared with state-of-the-art approaches that are based only on reinforcement learning.