A threshold-triggered Deep Q-Network-based Framework for self-healing in autonomic software-defined IIoT-Edge networks

Stochastic disruptions such as flash events caused by benign traffic bursts and switch thermal instability are primary contributors to intermittent network service interruptions in software-defined industrial networks deployed in offshore wind power plants (WPPs). These disruptions violate IEC 61400-25 Quality of Service (QoS) and service-level agreement (SLA) requirements, which are essential to ensuring high availability and the reliable transmission of critical, time-sensitive, and best effort data traffic. Failure to meet these IEC 61400-25 QoS and SLA standards can result in delayed or lost control signals, decreased operational efficiency, and increased risk of wind turbine generator downtime. To address these challenges, this study proposes a threshold-triggered Deep Q-Network (DQN)-based self-healing agent (TTDQSHA) that detects, analyzes, repairs network disruptions, and adapts network behavior and resource allocation.