الفهرس 
Literature ReviewOnly 14 pages are availabe for public view
 |  | from | 297 |  |  |
| | | from | 297 | | |
Abstract
This thesis is concerned with building an expert system for masonry type determination.
To ensure long service life of a building, one has to use the proper type of material.
An improvement procedure for the choosing process is exposed in this thesis. The improvement comes from a new aspect of using computer in decision-making based on the artificial intelligence.
A shell working under Windows 95 environment is used for the development of the expert system of this work. Level-5 Object (ver.3) is the shell used here. It provides vast variety of benefits: ability to create and display graphics, mechanism for handling uncertainty, simple and comprehensive role language, on-line help, and tip wizard.
ESMW is an expert system for masonry type determination. It is developed as a part of this work. The source of Imowledge for the Imowledge base of ESMW is acquired through conducting meetings with human experts of the domain, and through processing data collected in field survey. Every meeting with the human expert as well as the data from the field are documented. Analyzing procedure is then applied to that data in onler to extract roles that govern the choosing process.
Factors affecting the decision of masonry type determination are divided into major groups. These groups are: The architectural featU..res of the building, the stroctura1 features of
•
the building, the wall properties (thic1mess an~ function), the sUlTOunding environment-
including outer and inner temperatures-, the repaUing and fmishing of waU, and the standard of the building and economics of ditTerent masonry types.
The role base of ESMW consists of many roles. Some of those roles are roles giving intennediate conclusions, while others are roles giving final conclusions -Le. the type of masonry to be used-. Inference network technique is used for the presentation of the system behavior. In an inference network, branches are roles relating variable’s values (at the beginning of the branch) to the intennediate conclusion (or the fmal conclusion) at the end of the branch. The operator that is shown at the branch node governs this relationship.
n VaIidation and ewIuation of the developed system are conducted through studying actual cases. Two different cases are studied here. The strategy is that the conditions -i.e. the variables’ wIues- are p-esented to human expert of the domailL The human expert is asked to decide which type of masonry is to be used. Then the system is applied to the variables’ wIues and the conclusion is compared to that of the human expert.
Using of ESMW (for deciding which type of masonry to be used) leads to better dtnbility and serviceability. This is because the choosing of the proper type ensures that a minor (or may be no) repairing will be needed along the life of building.