Methods for optimizing handover stability in wireless broadband networks.

Abstract

The invention relates to a world modeling method for optimizing and stabilizing handover processes in 5G networks. The process consists of four steps: (1) Collecting performance metrics over time, including handover success rate, HOM and TTT parameters; (2) Learning the implicit representation of network state through a variable self-coding model incorporating a multi-head centralization mechanism; (3) Predicting network state evolution using an MDN-RNN model based on the learned implicit representation; (4) Optimizing HOM and TTT parameters for each station cluster using a multi-agent enhancement learning algorithm. The method helps station cells coordinate optimally over time, improving the handover success rate and enhancing radio network stability.

Images