# US 2025/0328809 A1 · Replacement component management using machine learning

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

- Number: US 2025/0328809 A1
- Document ID: US20250328809A1
- Status: Published application
- Kind: A1
- Application number: 18/637,890
- Filed: 2024-04-17
- Published: 2025-10-23
- Earliest claimed: 2024-04-17
- Assignees: Dell Products L.P.
- Applicants: Dell Products L.P.
- Inventors: Parminder Singh Sethi; Lakshmi Saroja Nalam; Vasanth Sathyanarayanan; Praveen Kumar
- Web page: https://patentgrep.com/patent/US20250328809A1

## CPC

- G06N 20/00: Machine learning
- G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks
- G06N 3/0895: Weakly supervised learning, e.g. semi-supervised or self-supervised learning
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method comprises predicting one or more device types for which one or more components thereof will be replaced, wherein the predicting is performed using at least a first machine learning algorithm, identifying locations of respective devices of a plurality of devices corresponding to the one or more device types, and determining one or more component distribution sources that are in proximity to the locations of the respective devices, wherein the determining is performed using at least a second machine learning algorithm. At least one device of the respective devices qualifying for at least one replacement component is identified. The method further comprises causing dispatching of the at least one replacement component to a location of the at least one device from a component distribution source of the one or more component distribution sources in proximity to the location of the at least one device.

## Claims

### Claim 1

A method comprising:
predicting one or more device types for which one or more components thereof will be replaced, wherein the predicting is performed using at least a first machine learning algorithm;
identifying locations of respective devices of a plurality of devices corresponding to the one or more device types;
determining one or more component distribution sources that are in proximity to the locations of the respective devices, wherein the determining is performed using at least a second machine learning algorithm;
identifying at least one device of the respective devices qualifying for at least one replacement component; and
causing dispatching of the at least one replacement component to a location of the at least one device from a component distribution source of the one or more component distribution sources in proximity to the location of the at least one device;
wherein the steps of the method are executed by a processing device operatively coupled to a memory.

### Claim 2 (depends on 1)

The method of claim 1 further comprising causing dispatching of one or more replacement components to the one or more component distribution sources.

### Claim 3 (depends on 1)

The method of claim 1 wherein the first machine learning algorithm comprises a multiple linear regression algorithm.

### Claim 4 (depends on 1)

The method of claim 1 further comprising training the first machine learning algorithm with data comprising replacement component fulfillment data for a plurality of device types.

### Claim 5 (depends on 1)

The method of claim 1 wherein predicting the one or more device types for which one or more components thereof will be replaced comprises using the first machine learning algorithm to analyze an input dataset comprising one or more independent variables, wherein the one or more independent variables comprise, for respective device types of the one or more device types, at least one of a device model, one or more component models, one or more component firmware versions, central processing unit utilization, memory utilization, input-output utilization and operating system version.

### Claim 6 (depends on 1)

The method of claim 1 wherein identifying the locations of the respective devices comprises collecting location data from respective software agents in each of the respective devices.

### Claim 7 (depends on 6)

The method of claim 6 wherein identifying the locations of the respective devices further comprises computing a length of time each of the respective devices have been in a given location.

### Claim 8 (depends on 1)

The method of claim 1 wherein the second machine learning algorithm comprises a graph convolutional network.

### Claim 9 (depends on 8)

The method of claim 8 wherein the second machine learning algorithm further comprises one or more decision trees.

### Claim 10 (depends on 8)

The method of claim 8 wherein the second machine learning algorithm further comprises a k-nearest neighbor algorithm.

### Claim 11 (depends on 1)

The method of claim 1 wherein identifying the at least one device of the respective devices qualifying for the at least one replacement component comprises:
monitoring a degradation rate of at least one component in the at least one device using at least one software agent;
comparing the monitored degradation rate to a threshold degradation rate for the at least one component to determine whether the monitored degradation rate exceeds the threshold degradation rate; and
identifying the at least one component as qualifying for replacement in response to the monitored degradation rate exceeding the threshold degradation rate.

### Claim 12 (depends on 1)

The method of claim 1 further comprising prioritizing dispatching of respective ones of a plurality of replacement components to respective ones of a plurality of devices from the one or more component distribution sources based on at least one of component degradation rates of the respective ones of the plurality of devices, durations for the respective ones of the plurality of replacement components to reach the respective ones of the plurality of devices from the one or more component distribution sources, durations to replace degraded components in the respective ones of the plurality of devices with the respective ones of the plurality of replacement components and durations to test the respective ones of the plurality of replacement components in the respective ones of the plurality of devices.

### Claim 13 (depends on 1)

The method of claim 1 further comprising training the second machine learning algorithm with a training dataset comprising replacement component fulfillment data for a plurality of devices, location data for the plurality of devices and distribution source location data.

### Claim 14

An apparatus comprising:
a processing device operatively coupled to a memory and configured:
to predict one or more device types for which one or more components thereof will be replaced, wherein the predicting is performed using at least a first machine learning algorithm;
to identify locations of respective devices of a plurality of devices corresponding to the one or more device types;
to determine one or more component distribution sources that are in proximity to the locations of the respective devices, wherein the determining is performed using at least a second machine learning algorithm;
to identify at least one device of the respective devices qualifying for at least one replacement component; and
to cause dispatching of the at least one replacement component to a location of the at least one device from a component distribution source of the one or more component distribution sources in proximity to the location of the at least one device.

### Claim 15 (depends on 14)

The apparatus of claim 14 wherein the processing device is further configured to train the first machine learning algorithm with data comprising replacement component fulfillment data for a plurality of device types.

### Claim 16 (depends on 14)

The apparatus of claim 14 wherein, in predicting the one or more device types for which one or more components thereof will be replaced, the processing device is configured to use the first machine learning algorithm to analyze an input dataset comprising one or more independent variables, wherein the one or more independent variables comprise, for respective device types of the one or more device types, at least one of a device model, one or more component models, one or more component firmware versions, central processing unit utilization, memory utilization, input-output utilization and operating system version.

### Claim 17 (depends on 14)

The apparatus of claim 14 wherein, in identifying the at least one device of the respective devices qualifying for the at least one replacement component, the processing device is configured:
to monitor a degradation rate of at least one component in the at least one device using at least one software agent;
to compare the monitored degradation rate to a threshold degradation rate for the at least one component to determine whether the monitored degradation rate exceeds the threshold degradation rate; and
to identify the at least one component as qualifying for replacement in response to the monitored degradation rate exceeding the threshold degradation rate.

### Claim 18

An article of manufacture comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes said at least one processing device to perform the steps of:
predicting one or more device types for which one or more components thereof will be replaced, wherein the predicting is performed using at least a first machine learning algorithm;
identifying locations of respective devices of a plurality of devices corresponding to the one or more device types;
determining one or more component distribution sources that are in proximity to the locations of the respective devices, wherein the determining is performed using at least a second machine learning algorithm;
identifying at least one device of the respective devices qualifying for at least one replacement component; and
causing dispatching of the at least one replacement component to a location of the at least one device from a component distribution source of the one or more component distribution sources in proximity to the location of the at least one device.

### Claim 19 (depends on 18)

The article of manufacture of claim 18 wherein, in predicting the one or more device types for which one or more components thereof will be replaced, the program code causes said at least one processing device to use the first machine learning algorithm to analyze an input dataset comprising one or more independent variables, wherein the one or more independent variables comprise, for respective device types of the one or more device types, at least one of a device model, one or more component models, one or more component firmware versions, central processing unit utilization, memory utilization, input-output utilization and operating system version.

### Claim 20 (depends on 18)

The article of manufacture of claim 18 wherein, in identifying the at least one device of the respective devices qualifying for the at least one replacement component, the program code causes said at least one processing device:
to monitor a degradation rate of at least one component in the at least one device using at least one software agent;
to compare the monitored degradation rate to a threshold degradation rate for the at least one component to determine whether the monitored degradation rate exceeds the threshold degradation rate; and
to identify the at least one component as qualifying for replacement in response to the monitored degradation rate exceeding the threshold degradation rate.
