US Patent No. 11,071,122

METHOD AND UNIT FOR RADIO RESOURCE MANAGEMENT USING REINFORCEMENT LEARNING


Patent No. 11,071,122
Issue Date July 20, 2021
Title Method And Unit For Radio Resource Management Using Reinforcement Learning
Inventorship Francesco Davide Calabrese, Kista (SE)
Pablo Soldati, Kista (SE)
Gunnar Peters, Kista (SE)
Euhanna Ghadimi, Kista (SE)
Assignee HUAWEI TECHNOLOGIES CO., LTD., Shenzhen (CN)

Claim of US Patent No. 11,071,122


1. A system comprising an agent node and a trainer node, wherein the agent node is configured to configure a radio resource parameter of a radio cell of a communication system wherein the trainer node is configured to:receive a training data message, associated with the radio cell, from the agent node, wherein the training data message comprises at least one of:a set of features representing a state of at least a part of the communication system at a first time period,
a control action performed by the agent node in the radio cell at the first time period,
the set of features at the second time period, or
a performance measurement;

store the received training data message in a database associated with the radio cell;
determine a control policy for the radio cell based on a Reinforcement Learning algorithm, and
transmit the determined control policy to the agent node; and, wherein the agent node is further configured to:obtain the control policy by receiving it from the trainer node;
determine the set of feature representing the state of at least a part of the communication system, at the first time period;
determine the control action to be performed for configuring the radio resource parameter in the radio cell at the first time period, out of a set of available control actions associated with the radio cell, based on the obtained control policy and the set of feature representing the state of the part of the communication system, at the first time period; and
configure the radio resource parameter of the radio cell based on the determined control action.