Estimating properties of units using system state graph models

A method of estimating properties of a plurality of operational units, the method comprising:

• generating a central system state graph model representing the properties of the plurality of operational units as probabilities of transitions between states of the plurality of operational units, wherein the states represent operational data;
• generating a respective updated system state graph model for each of the plurality of operational units, based on the central system state graph model and based on new operational data for each respective operational unit, wherein each updated system state graph model represents properties of the respective operational unit as probabilities of transitions between states for the respective operational unit;
• determining a distance measure between each respective updated system state graph model of the updated models and a version of the central system state graph model;
• based on the distance measure between the respective updated system state graph models fulfilling a divergence criterion, generating a plurality of new central system state graph models, each new central state graph model representing properties of a respective subset of the plurality of operational units as probabilities of transitions between states for the respective subset of the plurality of operational units;
• and
• initiating control of at least one of the radio network nodes in one of the subsets based on the respective new central system state graph model.

A device for estimating properties of a plurality of operational units, the device comprising a processor, wherein the processor is configured to:

• generate a central system state graph model representing the properties of the plurality of operational units as probabilities of transitions between states of the plurality of operational units, wherein the states represent operational data;
• generate a respective updated system state graph model for each of the plurality of operational units, based on the central system state graph model and based on new operational data for each respective operational unit, wherein each updated system state graph model represents properties of the respective operational unit as probabilities of transitions between states for the respective operational unit;
• determine a distance measure between each respective updated system state graph model of the updated models and a version of the central system state graph model;
• based on the distance measure between the respective updated system state graph models fulfilling a divergence criterion, generate a plurality of new central system state graph models, each new central state graph model representing properties of a respective subset of the plurality of operational units as probabilities of transitions between states for the respective subset of the plurality of operational units;
• and
• initiate control of at least one of the radio network nodes in one of the subsets based on the respective new central system state graph model.