Supporting deliberative real-time AI systems: A fixed priority scheduling approach

Yanching Chu; Alan Burns

Supporting deliberative real-time AI systems: A fixed priority scheduling approach

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Constructing deliberative real-time AI systems is challenging due to the high execution-time variance in AI algorithms and the requirement of worst-case bounds for hard real-time guarantees, often resulting in poor use of system resources. Using a motivating case study, the general problem of resource usage maximization is addressed. We show how the issues can be leveraged by employing a hybrid task model for anytime algorithms, which is supported by recent advances in fixed priority scheduling for imprecise computation.

In particular, with a novel scheduling scheme based on Dual Priority Scheduling, hard tasks are guaranteed by schedulability analysis and scheduled in favor of optional and anytime components which are executed whenever possible for enhancing system utility. Simulation studies show satisfactory performance on the case study with the application of the scheduling scheme. With its basis on fixed priority scheduling, it is expected that it can be easily incorporated into existing real-time operating systems, promoting wider use of imprecise computing and providing a framework where real-time AI applications can be suitably facilitated.

Original language	English
Title of host publication	19th Euromicro Conference on Real-Time Systems, Proceedings
Place of Publication	LOS ALAMITOS
Publisher	IEEE Computer Society
Pages	259-268
Number of pages	10
ISBN (Print)	978-0-7695-2914-1
Publication status	Published - 2007
Event	19th Euromicro Conference on Real-Time Systems (ECRTS 2007) - Pisa Duration: 4 Jul 2007 → 6 Jul 2007

Conference

Conference	19th Euromicro Conference on Real-Time Systems (ECRTS 2007)
City	Pisa
Period	4/07/07 → 6/07/07

Keywords

ALGORITHMS
TASKS

Cite this

@inproceedings{7369d33b640a4f68a18b8c4e7b08cec4,

title = "Supporting deliberative real-time AI systems: A fixed priority scheduling approach",

abstract = "Constructing deliberative real-time AI systems is challenging due to the high execution-time variance in AI algorithms and the requirement of worst-case bounds for hard real-time guarantees, often resulting in poor use of system resources. Using a motivating case study, the general problem of resource usage maximization is addressed. We show how the issues can be leveraged by employing a hybrid task model for anytime algorithms, which is supported by recent advances in fixed priority scheduling for imprecise computation.In particular, with a novel scheduling scheme based on Dual Priority Scheduling, hard tasks are guaranteed by schedulability analysis and scheduled in favor of optional and anytime components which are executed whenever possible for enhancing system utility. Simulation studies show satisfactory performance on the case study with the application of the scheduling scheme. With its basis on fixed priority scheduling, it is expected that it can be easily incorporated into existing real-time operating systems, promoting wider use of imprecise computing and providing a framework where real-time AI applications can be suitably facilitated.",

keywords = "ALGORITHMS, TASKS",

author = "Yanching Chu and Alan Burns",

year = "2007",

language = "English",

isbn = "978-0-7695-2914-1",

pages = "259--268",

booktitle = "19th Euromicro Conference on Real-Time Systems, Proceedings",

publisher = "IEEE Computer Society",

address = "United States",

note = "19th Euromicro Conference on Real-Time Systems (ECRTS 2007) ; Conference date: 04-07-2007 Through 06-07-2007",

}

TY - GEN

T1 - Supporting deliberative real-time AI systems: A fixed priority scheduling approach

AU - Chu, Yanching

AU - Burns, Alan

PY - 2007

Y1 - 2007

N2 - Constructing deliberative real-time AI systems is challenging due to the high execution-time variance in AI algorithms and the requirement of worst-case bounds for hard real-time guarantees, often resulting in poor use of system resources. Using a motivating case study, the general problem of resource usage maximization is addressed. We show how the issues can be leveraged by employing a hybrid task model for anytime algorithms, which is supported by recent advances in fixed priority scheduling for imprecise computation.In particular, with a novel scheduling scheme based on Dual Priority Scheduling, hard tasks are guaranteed by schedulability analysis and scheduled in favor of optional and anytime components which are executed whenever possible for enhancing system utility. Simulation studies show satisfactory performance on the case study with the application of the scheduling scheme. With its basis on fixed priority scheduling, it is expected that it can be easily incorporated into existing real-time operating systems, promoting wider use of imprecise computing and providing a framework where real-time AI applications can be suitably facilitated.

AB - Constructing deliberative real-time AI systems is challenging due to the high execution-time variance in AI algorithms and the requirement of worst-case bounds for hard real-time guarantees, often resulting in poor use of system resources. Using a motivating case study, the general problem of resource usage maximization is addressed. We show how the issues can be leveraged by employing a hybrid task model for anytime algorithms, which is supported by recent advances in fixed priority scheduling for imprecise computation.In particular, with a novel scheduling scheme based on Dual Priority Scheduling, hard tasks are guaranteed by schedulability analysis and scheduled in favor of optional and anytime components which are executed whenever possible for enhancing system utility. Simulation studies show satisfactory performance on the case study with the application of the scheduling scheme. With its basis on fixed priority scheduling, it is expected that it can be easily incorporated into existing real-time operating systems, promoting wider use of imprecise computing and providing a framework where real-time AI applications can be suitably facilitated.

KW - ALGORITHMS

KW - TASKS

M3 - Conference contribution

SN - 978-0-7695-2914-1

SP - 259

EP - 268

BT - 19th Euromicro Conference on Real-Time Systems, Proceedings

PB - IEEE Computer Society

CY - LOS ALAMITOS

T2 - 19th Euromicro Conference on Real-Time Systems (ECRTS 2007)

Y2 - 4 July 2007 through 6 July 2007

ER -