Deep Q-Learning for Self-Organizing Networks Fault Management and Radio Performance Improvement

  • 2019-01-10 12:41:35
  • Faris B. Mismar, Brian L. Evans
  • 0


We propose an algorithm to automate fault management in an outdoor cellularnetwork using deep reinforcement learning (RL) against wireless impairments.This algorithm enables the cellular network cluster to self-heal by allowing RLto learn how to improve the downlink signal to interference plus noise ratiothrough exploration and exploitation of various alarm corrective actions. Themain contributions of this paper are to 1) introduce a deep RL-based faulthandling algorithm which self-organizing networks can implement in a polynomialruntime and 2) show that this fault management method can improve the radiolink performance in a realistic network setup. Simulation results show that ourproposed algorithm learns an action sequence to clear alarms and improve theperformance in the cellular cluster better than existing algorithms, evenagainst the randomness of the network fault occurrences and user movements.


Introduction (beta)



Conclusion (beta)