Deadline management in intelligent agents
From Semantic Portal Wiki
{{#vardefine:category|Publication}}{{#vardefine:templatename|i.publication}}{{#vardefine:package|smwbp_instance_templates}}
| Edit |
Reference: {{#vardefine:pagename|deadline management in intelligent agents }}
- [[]]
bibtex
{{#vardefine:pagename|Deadline management in intelligent agents }}{{#vardefine:key| }}
abstract: This paper describes an approach to reasoning about goals with deadlines in an intelligent agent. The presented approach builds on the model of dynamic control where control plans describing the desirable behavior of an agent are dynamically constructed and used by meta-control decisions to choose the best behavior of the agent at each point in time. We extend this model by introducing scheduling behaviors that construct a control schedule of tasks required for time constrained goals, and by changing the meta-control decision criterion to consider both the current control plan and the current control schedule when deciding which behaviors to execute. We also present in this paper demonstrations and experimental results in the domain of office robots showing that our approach keeps the control cost low, and brings more run-time flexibility than current approaches.
download:
- paper:
- slides:
| Abstract | This paper describes an approach to reason … This paper describes an approach to reasoning about goals with deadlines in an intelligent agent. The presented approach builds on the model of dynamic control where control plans describing the desirable behavior of an agent are dynamically constructed and used by meta-control decisions to choose the best behavior of the agent at each point in time. We extend this model by introducing scheduling behaviors that construct a control schedule of tasks required for time constrained goals, and by changing the meta-control decision criterion to consider both the current control plan and the current control schedule when deciding which behaviors to execute. We also present in this paper demonstrations and experimental results in the domain of office robots showing that our approach keeps the control cost low, and brings more run-time flexibility than current approaches. -time flexibility than current approaches. |
| Author | Philippe Lal +, A +, and Barbara Hayes-Roth + |
| Bibtype | techreport + |
| Institution | Knowledge Systems, AI Laboratory + |
| Key | KSL-94-27 + |
| Month | May + |
| Number | KSL-94-27 + |
| Tag | Computer science + |
| Title | Deadline Management in Intelligent Agents + |
| Tr id | KSL-94-27 + |
| Year | 1994 + |
