A decision system for tunnel support design allows questions to be posed and answered in relation to the information stored in a rafael design knowledge base, Typically, such a system possesses an inferential capability and, in particular, has the capability to infer from promises that are imprecise, vague, or open-ended. The success of such a system depends on the ability to extract information from geology, rook mechanics, and tunnel technology and translate it into a form or forms which help the user to make a more intelligent decision for tunnel design. Potential applications of fuzzy methodology to such geological problems are many and includes topics such as fuzzy decision making, fuzzy pattern recognition, fuzzy clustering, fuzzy reasoning, and many etaergo Fuzzy methodology takes into account professional expertise and judgment in expert systems. This paper presents a preliminary discussion of approaches to the development of such systems.


Proper analysis of information in the field of tunnel design is hindered by the degree of specialization and the variety of professional disciplines involved in the design. Designers require special knowledge in geology, rock mechanics, tunnel technology, statistics, etc. if they are to handle the sheer volume of geologic information encountered in a tunnel project. Though specialists, they must still communicate with each other as they deal with the essential nature of the geological environment; a nature that operates as a complex whole, a nature that behaves independently of any analytical divisions made to it for the sake of human and academic convenience. Rock mechanics has passed the stage where exports can hope to comprehend and integrate the breadth of knowledge in the fields that contribute to rock mechanics. This has led to a concentration of the expertise that does exist into areas of limited, regional application. Research scientists rarely have the chance to discuss the practical consequence of their work with qualified practitioners. Practitioners in turn are generally not familiar with the possibilities and limitations in application of the latest analytical and numerical techniques. Extensive and detailed knowledge of this field is required to make a 'good' interpretation of frequently ambiguous, ill- defined geological information. It is now possible, through 'intelligent' systems analysis, to design computer programs which can, like a human, improve its interpretation techniques by experience, and also handle subtle variations in the problems addressed. Specialized knowledge in rock mechanics and rock engineering may be divided broadly into two schools. In the first school, often referred to as the 'classification' school, knowledge is organized in a way that reflects how it is applied to problems of interpretation. A detailed description is attempted of the rock, mineral, ground water, or geologic structure under consideration. The features of the object described are compared with the characteristics of typical examples or prototypes that have been organized with a taxonomy, and allocated class names. These class names are associated with the features or relationships, in common with other members of the class. The classes are then examined to find a 'best-fit,' according to what it is hoped are well- established criteria of what constitutes 'best.' Devotees of the second or 'behavioral' school take a markedly different approach to the problem of interpretation.

You can access this article if you purchase or spend a download.