Back to Search Start Over

Situation-aware composition and execution in dynamic environments by automated planning.

Authors :
Lu, Qiang
Wilson, Justin
Chen, Yixin
Gill, Christopher
Thomas, Louis
Roman, Gruia-Catalin
Chen, Guoliang
Source :
Engineering Applications of Artificial Intelligence. Oct2014, Vol. 35, p215-236. 22p.
Publication Year :
2014

Abstract

Existing workflow management systems execute complex sequences of tasks by invoking corresponding services identified at execution time and ordered according to a graph that captures the workflow execution rules and data dependencies. The technology is mature and is used across the Internet to support a wide range of business transactions by taking advantage of a relatively stable communication and service infrastructure. This paper extends the applicability of workflow technology by considering settings in which service availability is highly variable due to the mobility of people and devices and due to frequent and extended disconnections from the wired infrastructure. This paper extends workflow technology to support goal-directed coordination among physically mobile people and devices that form a transient community over a mobile ad hoc network (MANET). Our approach is based on the notion of open workflows, i.e., workflows that are constructed dynamically using available knowledge and services. Because the workflow construction and execution are situation aware, different workflows are built under different circumstances even when the need being addressed is unchanged. We also introduce a framework that co-designs workflow allocation and host scheduling in a unified way, which can handle communication range limitations in a mobile environment while previous methods cannot. Specifically, we transform the collaboration problem into a temporal planning model and then solve it using automated planners. We integrate this framework into the CiAN workflow management system, which was developed in our previous work. Results of experiments we have conducted show that our approach significantly broadens the scope of previous research by solving the challenges of availability, mobility, and communication constraints in MANET domains. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
09521976
Volume :
35
Database :
Academic Search Index
Journal :
Engineering Applications of Artificial Intelligence
Publication Type :
Academic Journal
Accession number :
97483034
Full Text :
https://doi.org/10.1016/j.engappai.2014.06.024