Temporal Knowledge Representation and Reasoning for Project Planning

Abstract

The author has written a working computer program which allows a user (or a different module of an AI planning program) to specify a set of temporal constraints including disjunctive constraints. In the context of a point-based temporal model, my program has required several innovative design choices. The program will find a feasible solution to the constraints if one exists, otherwise it will identify a contradiction. It is especially appropriate when adding new constraints to currently satisfiable existing constraints. The philosophical approach of dependency-directed backtracking is employed. This is implemented in a way that takes account of my problem's special structure. The program is particularly successful in certain examples for which chronological backtracking would be hopelessly inefficient. Other notorious examples require time-consuming search. These examples reveal important tradeoffs between explicitly storing temporal knowledge and deriving appropriate temporal knowledge on demand. Particular instances have been identified where explicit storage is critical for avoiding expensive search. Keywords: Knowledge based systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 24, 1988
Accession Number
ADA196075

Entities

People

  • Colin E. Bell

Organizations

  • University of Iowa

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Computational Complexity
  • Computer Programming
  • Computer Programs
  • Computers
  • Expert Systems
  • Intervals
  • Job Shop Scheduling
  • Knowledge Based Systems
  • Management Engineering
  • Management Planning And Control
  • Operations Research
  • Reasoning
  • Scheduling (Production)
  • Trees (Data Structures)
  • Universities

Readers

  • Applied Combinatorial Optimization and Logic Circuit Design.
  • Artificial Intelligence