# US 2026/0101289 A1 · Apparatus and a method for adapting power allocation constraints in a wireless communication system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0101289 A1
- Document ID: US20260101289A1
- Status: Published application
- Kind: A1
- Application number: 19/331,221
- Filed: 2025-09-17
- Published: 2026-04-09
- Earliest claimed: 2024-10-04
- Assignees: Nokia Solutions and Networks Oy
- Applicants: Nokia Solutions and Networks Oy
- Inventors: Shashwat MISHRA; Lou SALAUN; Muhammad Karam SHEHZAD
- Web page: https://patentgrep.com/patent/US20260101289A1

## CPC

- H04W 52/346: {distributing total power among users or channels}
- H04W 52/367: {Power values between minimum and maximum limits, e.g. dynamic range}
- G06N 3/02: Neural networks
- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/048: Activation functions
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- H04B 17/3913: {Predictive models, e.g. based on neural network models}
- H04W 52/223: {predicting future states of the transmission}
- H04W 52/42: in systems with time, space, frequency or polarisation diversity
- H04W 72/0473: {the resource being transmission power}
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04B: TRANSMISSION

## Abstract

Embodiments herein disclose a method performed by a first apparatus, the method comprising inputting, to a power allocation model employed by the first apparatus, an initial dataset comprising user data and constraints relating to power allocation in a system. The method comprises inputting to the power allocation model a classification of each constraint as a soft constraint or a hard constraint. The method comprises training the power allocation model to generate, using an optimization solver, a solution for the power allocation, based on the constraints in the initial dataset.

## Claims

### Claim 1

A method performed by a first apparatus, comprising:
inputting, to a power allocation model employed by the first apparatus, an initial dataset comprising user data and constraints relating to power allocation in a system;
inputting, to the power allocation model, a classification of each constraint as a soft constraint or a hard constraint; and
training the power allocation model to:
generate, using an optimization solver, a solution for the power allocation, based on the constraints in the initial dataset.

### Claim 2 (depends on 1)

The method as claimed in claim 1, wherein the generated solution, based on the constraints in the initial dataset, is infeasible, and the method comprises:
training the power allocation model to:
modify at least one constraint in the initial dataset; and
generate, using the optimization solver, a feasible solution for the power allocation, based on one or more of the at least one modified constraint.

### Claim 3 (depends on 2)

The method as claimed in claim 2, comprising:
inputting, to the power allocation model, a training dataset comprising:
the constraints in the initial dataset;
the at least one modified constraint; and
the feasible solution.

### Claim 4 (depends on 1)

The method as claimed in claim 1, wherein each constraint is classified based on at least one of:
an impact of the constraint on the solution for the power allocation; or
a flexibility of the system to accommodate the constraint.

### Claim 5 (depends on 2)

The method as claimed in claim 2, wherein the training the power allocation model to modify the at least one constraint comprises:
modifying at least one soft constraint.

### Claim 6 (depends on 2)

The method as claimed in claim 2, wherein the training the power allocation model to modify the at least one constraint and generate the feasible solution, comprises:
modifying at least one soft constraint;
generating an infeasible solution for the power allocation based on the at least one modified soft constraint;
modifying at least one hard constraint; and
generating a feasible solution for the power allocation based on the at least one modified hard constraint.

### Claim 7 (depends on 2)

The method as claimed in claim 2, wherein the power allocation model is trained to modify the at least one constraint by performing a bisection search.

### Claim 8 (depends on 1)

The method as claimed in claim 1, wherein the power allocation model is a graph neural network.

### Claim 9

A method performed by a first apparatus, comprising:
inputting, to a power allocation model employed by the first apparatus, an initial dataset comprising user data and constraints relating to power allocation in a system comprising a plurality of second apparatus;
inputting, to the power allocation model, a classification of each constraint in the initial dataset as a soft constraint or a hard constraint; and
generating, using the power allocation model, a solution for the power allocation, based on the constraints in the initial dataset.

### Claim 10 (depends on 9)

The method as claimed in claim 9, wherein the generated solution, based on the constraints in the initial dataset, is infeasible, and the method comprises:
modifying, using the power allocation model, at least one constraint in the initial dataset, wherein the at least one modified constraint is used for generating a feasible solution for the power allocation; and
transmitting, to at least one second apparatus, among the plurality of second apparatus, the at least one modified constraint.

### Claim 11 (depends on 10)

The method as claimed in claim 10, comprising:
receiving, from the at least one second apparatus, at least one of:
an acknowledgement (ACK) of the at least one modified constraint; or
a rejection (NACK) of the at least one modified constraint.

### Claim 12 (depends on 11)

The method as claimed in claim 11, wherein on receiving the ACK from the plurality of second apparatus, the method comprises:
generating, using the power allocation model, a feasible solution for the power allocation based on the at least one modified constraint; and
performing the power allocation based on the feasible solution.

### Claim 13 (depends on 11)

The method as claimed in claim 11, wherein on receiving the NACK, the method comprises performing at least one of:
adding the at least one second apparatus, from which the NACK was received, to a drop pool for a current transmission time interval (TTI); or
postponing transmission from the at least one second apparatus, from which the NACK was received, to a later TTI.

### Claim 14 (depends on 13)

The method as claimed in claim 13, comprising:
inputting, to the power allocation model, a second dataset comprising user data and constraints relating to the power allocation in the system comprising the plurality of second apparatus excluding the at least one second apparatus from which the NACK was received;
inputting, to the power allocation model, a classification of each constraint in the second dataset as a soft constraint or a hard constraint; and
generating, using the power allocation model, a feasible solution for the power allocation, based on the constraints in the second dataset.

### Claim 15 (depends on 10)

The method as claimed in claim 10, wherein the modifying the at least one constraint in the initial dataset comprises performing a bisection search.

### Claim 16 (depends on 10)

The method as claimed in claim 10, wherein the power allocation model is a graph neural network.

### Claim at position 17

17-18. (canceled)

### Claim 19

A method performed by a second apparatus, comprising:
receiving, from a first apparatus, at least one modified constraint, the at least one modified constraint relevant for generating a feasible solution for power allocation in a system; and
transmitting, to the first apparatus, at least one of:
an acceptance of the at least one modified constraint; or
a rejection of the at least one modified constraint.

### Claim 20 (depends on 19)

A second apparatus, comprising:
at least one memory storing a plurality of instructions; and
at least one processor configured to execute the plurality of instructions to cause the second apparatus to perform the method as claimed in claim 19.
