VMware, Inc.
TWO-LEVEL GRID-BASED ANOMALY AREA IDENTIFICATION AND SOLUTION NOMINATION FOR RADIO ACCESS NETWORKS

Abstract:

A system can include a network analysis platform for a two-level grid-based anomaly area identification and solution nomination in a radio access network. The network analysis platform can map key performance indicators for user sessions in the network to a grid that overlays a geographic area. The grid can be based on a military grid reference system. A machine-learning model can take a vector of key performance indicator samples as input and identify a problem for the grid. The network analysis platform can nominate cells to attempt to remediate based on ranking poor performing bins in the grid and determining the cells that contribute most to the problem in each bin. For a nominated cell, the network analysis platform can perform remediation actions to solve a coverage problem, throughput problem, or both. The problem can be solved for the grid while preventing conflicts between individual cells.