City Research Online

Abstract

This Thesis is concerned with the application of Distributed Artificial Intelligence techniques for the design of instruments.

In this thesis it is argued that, the early stages of the design process can be automated by the use of Distributed Artificial Intelligence systems that are contractual in their communication and control.

A Distributed Problem Solver is proposed, and implemented, for the purpose of conceptual design of instruments. The system consists of a community of knowledge-based agents, with expertise on design of instrument sub-systems. The agents, use a task-sharing form of cooperation for dynamic problem decomposition and sub-problem distribution phases of the design problem solving. New design concepts are generated by suitable combination of partial solutions.

To incorporate learning capabilities into our Distributed Problem Solver, we have proposed the use of Classifier System Modules as inductive knowledge-based agents. The application of Classifier Systems and Genetic Algorithms in the context of a number of concrete instrument design problems is investigated.

A normalized formulation is applied to the multi-modal design optimization of a Linear Variable Differential Transformer. A number of important proposals for the application of classifier systems to the design automation of instruments are detailed. In particular, an implemented classifier system is used for the purpose of design heuristic extraction for corrugated diaphragms, using a set of dimensionless curves. In this application, the classifier system has produced a set of useful design heuristics by direct interactions with the specified mathematical model.

Publication Type:	Thesis (Doctoral)
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	School of Science & Technology > Engineering > Electrical & Electronic Engineering School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads