國立虎尾科技大學 |

A Hybrid Pattern Based Design to Improve Real-Time System Adaptation.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	A Hybrid Pattern Based Design to Improve Real-Time System Adaptation./
作者:	Phillips, Brian J.
面頁冊數:	1 online resource (201 pages)
附註:	Source: Dissertation Abstracts International, Volume: 78-12(E), Section: B.
Contained By:	Dissertation Abstracts International78-12B(E).
標題:	Systems science. -
電子資源:	click for full text (PQDT)
ISBN:	9780355096453

A Hybrid Pattern Based Design to Improve Real-Time System Adaptation.
Phillips, Brian J.

A Hybrid Pattern Based Design to Improve Real-Time System Adaptation. - 1 online resource (201 pages)

Source: Dissertation Abstracts International, Volume: 78-12(E), Section: B.

Thesis (Ph.D.)--Stevens Institute of Technology, 2017.

Includes bibliographical references

This dissertation investigates hybrid design strategies to improve performance of adaptive systems that must alter their own structure in real-time based upon changing operational conditions. The strategies focus on reducing the time to choose an adaptation based on information collected about the operational environment and its own internal states along with other inhabitants of that environment. This novel approach combines behavior selection algorithms (e.g. classifiers) with different temporal domains and embeds them into a two-system ensemble whose goal is to improve adaptation speed. This hybrid approach provides a method where adaptive systems collect information from their environment and their own internal states and use this information to choose behaviors that satisfy one or more end goals.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9780355096453Subjects--Topical Terms:

1148479
Systems science.
Index Terms--Genre/Form:

554714
Electronic books.

A Hybrid Pattern Based Design to Improve Real-Time System Adaptation.
LDR:03425ntm a2200373Ki 4500 001 916775
005 20180928111500.5
006 m o u
007 cr mn||||a|a||
008 190606s2017 xx obm 000 0 eng d
020 $a 9780355096453
035 $a (MiAaPQ)AAI10275777
035 $a (MiAaPQ)stevens:10397
035 $a AAI10275777
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Phillips, Brian J. $3 1190604
245 1 2 $a A Hybrid Pattern Based Design to Improve Real-Time System Adaptation.
264 0 $c 2017
300 $a 1 online resource (201 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 78-12(E), Section: B.
500 $a Adviser: Mark Blackburn.
502 $a Thesis (Ph.D.)--Stevens Institute of Technology, 2017.
504 $a Includes bibliographical references
520 $a This dissertation investigates hybrid design strategies to improve performance of adaptive systems that must alter their own structure in real-time based upon changing operational conditions. The strategies focus on reducing the time to choose an adaptation based on information collected about the operational environment and its own internal states along with other inhabitants of that environment. This novel approach combines behavior selection algorithms (e.g. classifiers) with different temporal domains and embeds them into a two-system ensemble whose goal is to improve adaptation speed. This hybrid approach provides a method where adaptive systems collect information from their environment and their own internal states and use this information to choose behaviors that satisfy one or more end goals.
520 $a This research uses design techniques to improve the performance of real time systems with time constraints. It uses a metric of system survivability as a key performance parameter. A pattern-based design technique offers a novel method of addressing this problem. This research embraces a wide variety of behavior selection algorithms, embedding them into the two-system ensemble to improve adaptation speed.
520 $a This research focuses on design strategies that seek to overcome problems associated with time limitations. It incorporates classification techniques into the research approach. It assumes that no classification technique is sufficient to speed up adaptation decisions while ensuring that those decisions remain significant, and contribute to overall system performance (i.e. survivability).
520 $a To verify design strategies, this research effort uses a simulation environment composed of agents that combat each other within a virtual two-dimensional battlefield. Combat scenarios yield insights into the adaptation speed problem because it focuses on time-limited battles between simulated tanks with and without adaptive behaviors. It presents system survivability problems based upon these simulated tanks, and allows them to adapt and avoid failures while competing directly in combat, or for collecting limited resources.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Systems science. $3 1148479
650 4 $a Artificial intelligence. $3 559380
655 7 $a Electronic books. $2 local $3 554714
690 $a 0790
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Stevens Institute of Technology. $b Systems Engineering. $3 1190413
773 0 $t Dissertation Abstracts International $g 78-12B(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10275777 $z click for full text (PQDT)