Simulation of SARSA-Based Reinforcement- Learning Dynamic SDN Migration Process

This article presents a SARSA-RL-based dynamic SDN migration process, designed to tackle the dynamicity of Software-Defined Networking (SDN) migration process. Existing migration techniques often struggle with the dynamic nature of network conditions, resulting in suboptimal performance and disruptions. This work proposes a reinforcement learning (RL) approach using the SARSA algorithm, to create an intelligent agent capable of adapting the migration process based on changing network conditions, traffic demands, and budget constraints. The primary objective is to optimize the migration process under varying network conditions, improving network efficiency. Experimental results demonstrate the scheme's effectiveness in reducing network congestion and enhancing overall performance.