Network node and a method performed in a wireless communication network for handling configuration of radio network nodes using reinforcement learning
A method is herein provided, performed by a network node for handling configuration of radio network nodes in a wireless communication network. The network node such as a O&M node or similar calculates a configuration for one or more radio network nodes by using a machine learning model with a search space of parameters, wherein the search space is reduced based on an importance factor for parameters of the radio network nodes and/or the wireless communication network.
1 . A method performed by a network node for handling configuration of radio network nodes in a wireless communication network, the method comprising:
calculating a configuration for radio network nodes using a machine learning (ML) model with a search space of parameters, wherein the search space is reduced based on an importance factor for parameters of the radio network nodes and/or the wireless communication network, wherein the importance factor is defined as a ratio defining an interest and emphasis of a parameter.
2 . The method according to claim 1 , wherein the interest and emphasis is for a plurality of radio network nodes and based on a business intent.
3 . The method according to claim 1 , wherein the search space is further reduced, before reducing the search space using the importance factor for parameters, by using a similarity matrix for clustering radio network nodes of similar parameters.
4 . The method according to claim 3 , wherein the ML model is using the output of similarity matrix and the output of the reduction based on the importance factor for parameters.
5 . The method according to claim 3 , wherein the similarity matrix is generated by hierarchical clustering of similar sites of radio network nodes, wherein the clustering is based on similarity of load and resource factors.
6 . The method according to claim 1 , wherein input to the ML model is a combination of network and demand parameters for radio network nodes selected in the wireless communication network, and wherein output of the ML model is a set of parameter values for each radio network node attaining a state of low energy consumption while maximizing demand requirements.
7 . The method according to claim 6 , wherein the selected radio network nodes are based on a similarity matrix for clustering radio network nodes of similar parameters.
8 . The method according to claim 1 , wherein an output of the ML model is weighted based on an average of all quality of experiences (QoE) which can be predicted from network parameters and a penalty factor for each unit of added energy consumption.
9 . The method according to claim 1 , further comprising detecting a mobility pattern in the wireless communication network; and wherein the ML model used in calculating the configuration is selected based on the detected mobility pattern.
10 . The method according to claim 1 , further comprising
sending the calculated configuration to the radio network nodes.
11 . A network node for handling configuration of radio network nodes in a wireless communication network, the network node comprising:
processing circuitry; and
memory coupled with the processing circuitry, wherein the memory includes instructions that when executed by the processing circuitry causes the network node to perform operations comprising:
calculate a configuration for radio network nodes using a machine learning (ML) model with a search space of parameters, wherein the search space is reduced based on an importance factor for parameters of the radio network nodes and/or the wireless communication network, wherein the importance factor is defined as a ratio defining an interest and emphasis of a parameter.
12 . The network node according to claim 11 , wherein the interest and emphasis is for a plurality of radio network nodes and based on a business intent.
13 . The network node according to claim 11 , wherein the search space is further reduced, before reducing the search space using the importance factor for parameters, by using a similarity matrix for clustering radio network nodes of similar parameters.
14 . The network node according to claim 13 , wherein the ML model is using the output of the similarity matrix and the output of the reduction based on the importance factor for parameters.
15 . The network node according to claim 13 , wherein the similarity matrix is generated by hierarchical clustering of similar sites of radio network nodes, wherein the clustering is based on similarity of load and resource factors.
16 . The network node according to claim 11 , wherein input to the ML model is a combination of network and demand parameters for radio network nodes selected in the wireless communication network, and wherein output of the ML model is a set of parameter values for each radio network node attaining a state of low energy consumption while maximizing demand requirements.
17 . The network node according to claim 11 , wherein the operations further comprise:
detect a mobility pattern in the wireless communication network; and
select the ML model used in calculating the configuration based on the detected mobility pattern.
18 . The network node according to claim 11 , wherein the operations further comprise:
send the calculated configuration to the radio network nodes.