In order to investigate the use of fuzzy logic in decision-support systems (DSS) for dairy cattle breeding, a first-generation prototype software system was developed. The objectives were to determine the advantages and limitations of fuzzy logic for this type of application, and to establish a basis for the development of more complete DSS in the future. The goal of the prototype DSS was to make culling decisions on the basis of monthly production data. An analysis of the development process of this prototype demonstrated the importance of creating a thorough methodology for the elicitation and implementation of knowledge. A framework for the development of fuzzy decision-support systems was established, consisting of four phases: the project groundwork phase, elicitation of knowledge from the expert, implementation of that knowledge, and system validation. In this framework, it is proposed that, in the case of multiple experts, knowledge can be amalgamated or aggregated. Once this framework was established, a second-generation prototype DSS was developed. Contrary to the first-generation prototype, where the encoded expertise was limited to three experts from the same domain, the second-generation prototype considered the knowledge of two individuals from each of three domains (Dairy researchers, Producers, and Dairy herd improvement specialists). An aggregation approach was used which involved the development and maintenance of separate modules, each containing the compiled expertise of one of the six experts.

No description available.

The goal of this thesis was to develop a framework that could integrate and analyse industry-wide information for the support of on-farm decision-making in dairy-cattle breeding. Specific objectives included (i) describing a dairy breeding information system (DBIS); (ii) examining how the Internet could be exploited to improve the DBIS and its functioning; (iii) describing a process for implementing a unified data model to facilitate integrated user access to information in the DBIS; and (iv) developing software to support decision-making by facilitating access to a unified data model when implemented as a database management software. / The first objective was achieved by following a systems approach---defining a goal, boundary, functions, structure and performance---to describe multi-organisational information systems and, specifically, a DBIS in the Canadian dairy industry. Using this framework, the subsequent analysis of the DBIS looked at its overall effectiveness. The DBIS was also compared with other known systems, where the number of participants (as well as their roles) differs from the Canadian situation. Improvements were suggested for the Canadian DBIS by focussing on the decision-maker's ability to retrieve, integrate and consider required information through information technologies. The second objective involved using the systems approach to investigate the kinds of information (if any) provided on Web sites of the DBIS participants, and to see if the Internet could be exploited to improve this process, either in terms of improved transfer speed or data transformation. It was established that the Internet is being used for rapid, flexible access to support information by DBIS participants, but that it is being under-utilised, particularly where herd output information is concerned. Herd output information could be filtered, integrated and transformed to support specific user needs at appropriate levels of intelligence density. It was further postulated that these data could be exploited more effectively through the use of such information technologies as common data exchange mechanisms and decision-support systems. The third objective was achieved through applying information engineering methods to develop a data model to represent the DBIS. This unified model was described in conceptual, logical and physical terms, and facilitated transparent access for on-farm users to information from more than one source organisation. It was demonstrated that such a model could maintain the autonomy of participating organisations while simultaneously creating an amalgamated databank for decisio

No description available.