AN OPERATING SYSTEM FOR DATA MANAGEMENT, DECISION SUPPORT AND CONTROL IN A DISCRETE OPERATIONS ENVIRONMENT
Abstract
The objective of this research is to develop a formal model for designing and implementing a computerized decision support capability for automatic control in a discrete operations environment. The model provides information representation and intelligence capabilities through a synthesis of data base management and artificial intelligence techniques. The system includes a knowledge system for representation of declarative data and procedural knowledge, a problem processing system to combine relevant data and procedures for problem solving, and a language interface to management for system specification and process control. Illustrations from a manufacturing environment are used to demonstrate the modeling concepts. The knowledge system combines semantic networks and predicate calculus to model the system entities, attributes, and relationships. Knowledge is partitioned into three major sub-components: reference data, operational data, and decision control. Reference data pertain to static manufacturing system entities. Operational data denote states of the manufacturing system and define the logic of the manufacturing processes. Decision control determines how the reference and operational data can be used for problem solving in the modeled system. The intelligence capability is provided by a problem processing system. This system is based on a state-space approach to deduction for predicate calculus axioms. A procedure of control and decision support executes decision control axioms to determine which operational logic should be applied. For a given state of the modeled system, this procedure identifies operational logic pertinent to the state, determines which of these logic procedures are feasible, recommends, and possibly implements actions based on decision control evaluation. A decision support mode is invoked if automatic control evaluation is inconclusive. A language interface permits users to interact with the information representation and intelligence capabilities of the model. A user can specify the reference data, operational logic, and decision control in a high-level natural language. Augmented transition networks translate the natural language specification into the internal knowledge representation. A synthesis of semantic networks, predicate calculus, and network data base concepts is used for the organization and management of the knowledge system. A data base schema is constructed for a job shop scheduling problem, loaded with data instantiations, and used to illustrate control and decision support in a manufacturing environment.
Degree
Ph.D.
Subject Area
Business community
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.