Multi-resource Task Scheduling with Deadline Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current multi-resource task scheduling algorithms result in prolonged average completion times for tasks, leading to poor user experience due to inefficient resource allocation, especially when dealing with tasks that have different resource requirements.

Innovation Solution

A method that classifies packets into those with deadlines and those without, using the EDF algorithm for packets with deadlines to minimize packet loss and the SJF algorithm for packets without deadlines to shorten completion times, while estimating virtual start and completion times to determine successful scheduling and potentially advancing the scheduling of packets without deadlines to reduce overall average completion time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If dominant resource scheduling is used to ensure fairness, then resource distribution fairness is improved, but average completion time increases

Engineering Contradiction:

Improveresource distribution fairnessVSAvoidaverage completion time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent segments packets into two categories: packets with deadlines and packets without deadlines. This segmentation allows different scheduling strategies to be applied to different packet types, resolving the contradiction by ensuring fairness for deadline packets while optimizing completion time for non-deadline packets through advanced scheduling opportunities.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic scheduling by allowing packets without deadlines to be advanced in the scheduling queue when resources become available. This dynamic adjustment enables the system to maintain fairness for deadline-critical packets while reducing average completion time by flexibly scheduling non-critical packets ahead when possible.

Inventive Principle:

Principle #15Dynamics

2Reliability

If EDF algorithm is used for packets with deadline, then packet loss rate is reduced, but average completion time increases

Engineering Contradiction:

Improvepacket loss rateVSAvoidaverage completion time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the packet queue into deadline-critical packets and non-critical packets. EDF algorithm is applied specifically to deadline packets to minimize packet loss, while non-critical packets can be scheduled in advance when resources are available, thus reducing their waiting time without affecting deadline packet delivery.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent allows packets without deadlines to be scheduled in advance when resources become available. This preliminary action reduces the average completion time for non-critical packets while leaving deadline packets to be handled by EDF algorithm, which ensures they are not lost by maintaining their deadline-based priority.

Inventive Principle:

Principle #10Preliminary action

3Loss of time

If SJF algorithm is used for packets without deadline, then average completion time is shortened, but packet loss rate increases

Engineering Contradiction:

Improveaverage completion timeVSAvoidpacket loss rate

Core Design Contradiction:

Loss of timeVSReliability

Solution Approach 1:

The patent segments packets into deadline-critical and non-critical categories. SJF algorithm is applied to non-critical packets to minimize average completion time, while deadline-critical packets are handled by EDF algorithm to prevent packet loss, thus resolving the contradiction between speed and reliability.

Inventive Principle:

Principle #1Segmentation

4Productivity

If multi-resource scheduling is implemented, then resource allocation efficiency is improved, but system complexity increases

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidscheduling system complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the scheduling problem into two independent sub-problems: scheduling deadline-critical packets using EDF and scheduling non-critical packets using SJF with advance scheduling opportunities. This segmentation simplifies the overall scheduling system while still achieving efficient multi-resource allocation by handling different packet types with appropriate algorithms.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS9389912B2Multi-resource task scheduling method

Publication Date: 2016.07.12 TSINGHUA UNIVERSITY

AI summary

A multi-resource task scheduling method includes: classifying concurrency packets to distinguish packets with deadline and packets without deadline; ranking packets with deadline using EDF algorithm and ranking packets without deadline using SJF algorithm; estimating a virtual start time and a virtual completion time according to ranking results; determining whether packets with deadline can be scheduled successfully; if yes, determining whether there is a packet without deadline, which can be arranged to be scheduled before the packets with deadline and can shorten average completion time, existing in the packets without deadline; and if yes, scheduling the packet without deadline, which can be arranged to be scheduled before the packets with deadline, in advance. The method can shorten the average completion time of all tasks greatly under multi-resource circumstance.