Computing Performance Budget Management for Anytime Algorithms

Solution Overview

Problem

In real-time systems, managing computing performance to ensure timely execution of applications, especially anytime algorithms, is challenging due to unpredictable computing time availability and the need to balance quality of service with execution time, while also optimizing resource utilization.

Innovation Solution

A method that assigns a computing-performance budget to applications, allowing for re-execution or redistribution of remaining budget based on available resources and result quality, using scheduling algorithms to manage and redistribute computing performance effectively.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If computing-performance budget is assigned to anytime algorithms, then quality of service is improved, but execution time becomes unpredictable

Engineering Contradiction:

Improvequality of serviceVSAvoidexecution time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent implements dynamic budget management where the computing-performance budget is not fixed but can be adjusted based on system state. The controller monitors available computing resources and dynamically assigns or releases budget portions to anytime algorithms, allowing the system to adapt between quality improvement and time constraints based on current workload and resource availability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the parameter of computing budget allocation from static to variable. By introducing a controller that monitors system state and adjusts budget assignments in real-time, the system can modify the computing-performance parameters allocated to anytime algorithms, enabling flexible trade-offs between quality of service and execution time based on actual system conditions.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If computing device is activated for re-execution of anytime algorithm, then quality of result is improved, but computing resource utilization increases

Engineering Contradiction:

Improvequality of resultVSAvoidcomputing resource utilization

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent implements a feedback mechanism where the controller monitors the quality of results produced by anytime algorithms and uses this information to make informed decisions about re-execution. The system evaluates whether the additional computing resources required for re-execution will actually improve the result quality sufficiently to justify the increased resource utilization, creating a closed-loop control system.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent applies partial action by allowing re-execution of anytime algorithms only when necessary and only to the extent needed to achieve satisfactory result quality. Rather than continuously re-executing algorithms or always allocating maximum resources, the system performs re-execution partially - only when the current result quality threshold is not met and resources are available.

Inventive Principle:

Principle #16Partial or excessive action

3Productivity

If remaining computing-performance budget is released to further applications, then resource utilization is optimized, but quality of service for first application may deteriorate

Engineering Contradiction:

Improveresource utilizationVSAvoidquality of service

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent implements dynamic budget management where the computing-performance budget is not fixed but can be adjusted based on system state. The controller monitors available computing resources and dynamically assigns or releases budget portions to anytime algorithms, allowing the system to adapt between quality improvement and time constraints based on current workload and resource availability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent implements a feedback mechanism where the controller monitors the quality of results produced by anytime algorithms and uses this information to make informed decisions about re-execution. The system evaluates whether the additional computing resources required for re-execution will actually improve the result quality sufficiently to justify the increased resource utilization, creating a closed-loop control system.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11567798B2Method and apparatus for managing computing performance in a data processing system

Publication Date: 2023.01.31 ROBERT BOSCH GMBH

AI summary

A method for managing computing performance in a real-time data processing system, having at least one apparatus for controlling at least one computing device that is embodied for execution of a first and of at least one further application, at least the first application encompassing at least one anytime algorithm. The method includes: assigning a first computing-performance budget, as a function of a first point in time, to the first application; activating the computing device for execution of the first application, at the first point in time by the computing device; checking whether, after execution of the first application, a portion of the first computing-performance budget is still available and, in the event a portion of the budget is still available, performing (i) activating the computing device for re-execution of the first application, or (ii) assigning the portion of the first computing-performance budget to at least one further application.