US 12,282,828 B2
Methods and apparatuses for determining optimal configuration in cognitive autonomous networks
Anubhab Banerjee, Munich (DE); and Stephen Mwanje, Dorfen (DE)
Assigned to NOKIA TECHNOLOGIES OY, Espoo (FI)
Filed by NOKIA TECHNOLOGIES OY, Espoo (FI)
Filed on Jul. 16, 2021, as Appl. No. 17/377,662.
Claims priority of provisional application 63/056,084, filed on Jul. 24, 2020.
Prior Publication US 2022/0027789 A1, Jan. 27, 2022
Int. Cl. G06N 20/00 (2019.01); G06F 9/445 (2018.01)
CPC G06N 20/00 (2019.01) [G06F 9/44505 (2013.01)] 1 Claim
OG exemplary drawing
 
1. A method, comprising:
receiving, by a controller of a 5G network apparatus, a request from at least two cognitive functions in a cognitive autonomous network (CAN) to calculate or recalculate a configuration weight;
transmitting, by the controller to the at least two cognitive functions, a request to provide information;
receiving, by the controller from the at least two cognitive functions, the information comprising a configuration range set for the at least two cognitive functions, a first utility function and a second utility function generated by the at least two cognitive functions, and the configuration weight representing a relative importance that an input configuration has on an output for the at least two cognitive functions, wherein a value for the configuration weight of each cognitive function of the at least two cognitive functions is set by:
determining a maximum output value for all pairs of any combination of input values of a plurality of input values of each cognitive function,
performing a Shapley value calculation using the maximum output value for all pairs of any combination of input values of a plurality of input values of each cognitive function; and
set the value of the configuration weight of each cognitive function based on a result of the Shapley value calculation;
based on the received information, calculating, by the controller, a first value for configuration of the cognitive autonomous network (CAN), wherein the first value for the configuration is based on meeting combined interests of the at least two cognitive functions, and a value for the configuration that provides a best possible compromise when taking into account interests of the at least two cognitive functions in the network, wherein the calculating comprises determining the first value for the configuration using a Fisher market game calculation of ƒ1(s)Two1,p1, ƒ2(s)Two2,p1, wherein:
an input value of the plurality of input values of the at least two cognitive functions is denoted by p1,
a first configuration weight of a first cognitive function of the at least two cognitive functions is denoted by wo1,p1,
a second configuration weight of a second cognitive function of the at least two cognitive functions is denoted by wo2,p1,
a value of periodicity is denoted by T,
the first utility function of the first cognitive function of the at least two cognitive functions is denoted by ƒ1(s),
the second utility function of the second cognitive function of the at least two cognitive functions is denoted by ƒ2(s),
s is a sample value within the configuration range set for the at least two cognitive functions, and
the first value is a value of s at which the Fisher market game calculation is maximum; and
using the first value for the configuration to reconfigure the at least two cognitive functions and to reconfigure the CAN by updating the configuration range set using the first value,
wherein the controller is configured to use the reconfigured CAN to perform functions associated with operation of the apparatus, which includes precoding of antenna gain/phase parameters, encoding of individual bits forming a communication message, decoding of individual bits forming a communication message, formatting of information, and processes related to management of communication resources.