A new generation of workflow-management-systems: beyond Taylorism with MOBILE
ACM SIGOIS Bulletin - Special issue: groupware for self-organizing units
KQML as an agent communication language
Software agents
Using metalevel techniques in a flexible toolkit for CSCW applications
ACM Transactions on Computer-Human Interaction (TOCHI)
From Centralized Workflow Specification to Distributed WorkflowExecution
Journal of Intelligent Information Systems - Special issue on workflow management systems
Distributed problem solving and planning
Multiagent systems
The Design of Intelligent Agents: A Layered Approach
The Design of Intelligent Agents: A Layered Approach
Knowledge Engineering: Unifying Knowlegde Base and Database Design
Knowledge Engineering: Unifying Knowlegde Base and Database Design
Approximate Reasoning about Combined Knowledge
ATAL '97 Proceedings of the 4th International Workshop on Intelligent Agents IV, Agent Theories, Architectures, and Languages
AI '01 Proceedings of the 14th Australian Joint Conference on Artificial Intelligence: Advances in Artificial Intelligence
Strategic Multi-personal-agent Interaction
PRIMA 2001 Proceedings of the 4th Pacific Rim International Workshop on Multi-Agents, Intelligent Agents: Specification, Modeling, and Applications
Hi-index | 0.00 |
A multi-agent system manages high-level business processes. The conceptual agent architecture is a three-layer BDI, hybrid architecture. During processing the responsibility for a sub-process may be delegated. The delegation problem is the problem of choosing an individual to delegate responsibility to so as to achieve some corporate goal. An approach to the delegation problem uses an estimate of the probability that each individual is the best choice. This estimate is based on the values of observed parameters. These values are based on historic information, and are accepted as long as they are statistically stable. If variations in these observed values lie outside specified limits then the system attempts to deduce why this is so. If a reason for an unexpected value is quantifiable then that reason is used to revise subsequent values while that reason remains significant. The architecture has been trialed on a process management application in a university administrative context.