The CommonKADS methodology is a collection of structured methods for building knowledge-based systems. A key component of CommonKADS is the library of generic inference models which can be applied to tasks of specified types. These generic models can either be used as frameworks for knowledge acquisition, or to verify the completeness of models developed by analysis of the domain. However, the generic models for some task types, such as knowledge-based planning, are not well-developed. Since knowledge-based planning is an important commercial application of Artificial Intelligence, there is a clear need for the development of generic models for planning tasks. Many of the generic models which currently exist have been derived from modelling of existing AI systems. These models have the strength of proven applicability. There are a number of well-known and well-tried AI planning systems in existence; one of the best known is the Open Planning Architecture (O-Plan). This paper describes the development of a CommonKADS generic inference model for knowledge-based planning tasks, based on the capabilities of the O-Plan system. The paper also describes the verification of this model in the context of a real-life planning task: the assignment and management of Royal Air Force Search and Rescue operations.
[1]
Bob J. Wielinga,et al.
CML: The CommonKADS Conceptual Modelling Language
,
1994,
EKAW.
[2]
William J. Clancey,et al.
Heuristic Classification
,
1986,
Artif. Intell..
[3]
Nigel Shadbolt,et al.
Knowledge Level Planning in the Search and Rescue Domain
,
1995
.
[4]
Ashok K. Goel,et al.
Multistrategy adaptive path planning
,
1994,
IEEE Expert.
[5]
Austin Tate,et al.
O-Plan: a Knowledge-Based Planner and its Application to Logistics
,
1996
.
[6]
V. Richard Benjamins,et al.
Modeling Planning Tasks
,
1996,
AIPS.
[7]
Van de velde Breuker.
Common KADS Library for Expertise Modelling
,
1994
.