Query Optimization and Planning in Object-Oriented Knowledge Bases
Abstract
The purpose of the project entitled 'Query Planning and Optimization in Object- oriented Knowledge Bases' sponsored by AFOSR-90-0004 is to extend a deductive object base with knowledge-based problem solving and planning, which is intended to realize the concept of very-high level programming in a database system. The input to such a system is a specification of the problem to be solved (as a set of goals) and the output is a solution of the problem, where the knowledge-based problem solving system deals with problems that do not change the state of a database and the planning system processes goals that require some state changes in the database. In our approach, the knowledge-based problem solving system stores a set of problem models (such as graph problems) so that an input problem can be matched through an object-oriented specialization/generalization process. If no problem models can be matched by a given problem, the user should be provided with a high-level programming system that allows a top-down problem solving process be carried out until some matches can be found at detailed implementation stages. For the planning system, we have realized that most of the conventional approaches based on the formulation of operations-preconditions-postconditions have been proved to be inefficient. We have classified general planning problems into several classes so that each class can be solved individually and efficiently. With this approach, each class of planning problems can be constructed as a problem model and included in the general problem solving system.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1992
- Accession Number
- ADA256006
Entities
People
- Phillip Sheu
Organizations
- Rutgers School of Engineering