| The design of underground electrical supply for residential development is presently carried out manually, resulting often in an overdesigned, costly, and nonstandardized solution. Up to now no comprehensive computerized tool exists that encompasses design of the secondary, primary, and street lighting systems which are components of the overall residential electrical supply. As an ill-structured and open-ended problem, this design is hard to automate with conventional methods such as operational research or CAD. An additional complexity in automating electrical plat design is imposed by the need to process spatial data such as circuit maps and construction plans. After a comprehensive knowledge acquisition, the author has proposed an Artificial Intelligence (AI) methodology, namely Intelligent Decision Support Systems (IDSS) to automate electrical plat design (AEPD). To accommodate diverse software environments, data models, and problem solving paradigms, the author has proposed a conceptually new problem solving architecture. This architecture, called the Blackboard Based IDSS, consists of different knowledge sources, including the human designer, concentrated around a global database. The knowledge sources are implemented in a Geographic Information System (GIS), procedural programs, and an expert system shell. In a such hybrid environment, the GIS principal task is to structure and formalize a "real world" representation required by other AEPD knowledge sources. To evaluate the feasibility of the proposed concept, the author has developed an AEPD prototype that functionally covers the secondary part of the electrical plat design. The results obtained by testing the AEPD prototype on an actual plat, and comparison with the results of the manual design, prove the feasibility of the proposed AEPD concept. An additional outcome of this dissertation is the successful demonstration of the application of novel technologies: GIS, and AI. These are used for structuring the required data and for capturing the designer problem solving technique, respectively. The synergy of these two agents, one "being able to see" and another "being able to think" provides the distribution system designers with a viable tool to support their decision making. |
