Autonomous network healing in hybrid contacts center infrastructures: A reinforcement learning approach
Cisco Systems, Inc., Department of Contact Centre, Dallas-Fort Worth Metroplex, TX, USA.
Research Article
World Journal of Advanced Research and Reviews, 2024, 23(01), 3199-3208
Publication history:
Received on 15 May 2024; revised on 25 July 2024; accepted on 28 July 2024
Abstract:
This research introduces an autonomous network healing system for hybrid contact center infrastructures that combines on-premises Contact Center Unified Contact Center Enterprise (UCCE) with cloud-based Webex Contact Center deployments. Network failures in hybrid environments create complex cascading effects that traditional monitoring systems cannot predict or resolve autonomously. Our reinforcement learning-based approach uses Deep Q-Networks (DQN) to learn optimal healing strategies from historical incident data, environmental state representations, and real-time network telemetry. The system achieved 89% automatic resolution rate for network incidents, reducing mean time to recovery (MTTR) from 23 minutes to 4.2 minutes across 15 enterprise deployments. The agent learned to predict failure patterns 12 minutes before occurrence with 94% accuracy, enabling proactive healing actions. Implementation demonstrates significant improvements in service level agreement (SLA) compliance and operational costs while reducing dependency on human intervention for routine network issues.
Keywords:
Autonomous Systems; Network Healing; Reinforcement Learning; Hybrid Cloud; Contact Centers; Deep Q-Networks; Predictive Maintenance
Full text article in PDF:
Copyright information:
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0