ML-Based Resource Allocation for Task Prioritization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current systems for managing computing resources struggle to optimize task processing and resource allocation efficiently, leading to suboptimal utilization and increased processing times, as they lack the ability to adapt dynamically based on past experiences and priorities.

Innovation Solution

The implementation of machine learning algorithms that analyze past task execution data and feedback to determine patterns and models for optimizing computing resource allocation and scheduling, allowing for probabilistic measures to be used in prioritizing and allocating resources effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional resource allocation methods are used, then system simplicity is maintained, but resource utilization efficiency deteriorates

Engineering Contradiction:

Improveresource utilization efficiencyVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent implements feedback mechanisms where the machine learning algorithm continuously receives information about task completion status, resource availability, and processing outcomes. This feedback loop enables the system to dynamically adjust resource allocation decisions, improving resource utilization efficiency while managing complexity through iterative learning rather than complex predetermined rules

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The machine learning algorithm autonomously makes resource allocation decisions based on learned patterns from historical data, without requiring complex external control systems. The system serves itself by automatically optimizing resource distribution, which improves efficiency while keeping the overall system architecture relatively simple

Inventive Principle:

Principle #25Self-service

2Adaptability or versatility

If static resource allocation is used, then system complexity is reduced, but adaptability to varying task priorities deteriorates

Engineering Contradiction:

Improveadaptability to task prioritiesVSAvoidallocation system complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements dynamic resource allocation where the machine learning algorithm continuously adapts its decisions based on current task priorities and system state. Resource allocations are not fixed but dynamically adjusted in response to changing conditions, enabling high adaptability while the learning algorithm manages the complexity of handling dynamic variations

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes allocation parameters based on learned patterns from historical task data. The machine learning model adjusts resource distribution parameters dynamically according to task characteristics and priorities, providing adaptability without requiring manual configuration of complex allocation rules for each scenario

Inventive Principle:

Principle #35Parameter changes

3Productivity

If machine learning optimization is implemented, then resource allocation efficiency is improved, but computational overhead increases

Engineering Contradiction:

Improvetask processing efficiencyVSAvoidcomputational overhead

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The machine learning algorithm performs preliminary learning during offline training phases using historical task data, building predictive models in advance. During actual task execution, the pre-trained model makes rapid allocation decisions without requiring intensive real-time computation, thus improving processing efficiency while minimizing online computational overhead

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system applies machine learning optimization selectively to the most critical resource allocation decisions rather than all decisions equally. By focusing computational resources on high-impact allocation choices and using simpler rules for less critical decisions, the system achieves significant efficiency improvements while keeping overall computational overhead manageable

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10572307B2System and method of training machine learning algorithm to satisfactorily allocate resources for task execution

Publication Date: 2020.02.25 BANK OF AMERICA CORP

AI summary

In one or more embodiments, one or more systems, processes, and/or methods may receive first task sets that include respective first tasks and one or more of respective first priorities, respective first minimum computing resource allocations, and respective first maximum processing times; receive first satisfaction information associated with processing the first task sets; receive first execution metric information associated with processing the first task sets; determine a first pattern based at least on the first satisfaction information and based at least on the first execution metric information; receive second task sets that include respective second tasks and one or more of respective second priorities, respective second minimum computing resource allocations, and respective second maximum processing times; determine, based at least on the first pattern, computing resources allocations for the second task sets; and determine, based at least on the first pattern, a processing order for the second task sets.