DETERMINATION OF MOST REPRESENTATIVE RESIDENTIAL SUBDIVISION

A procedure is established for determination of the most representativ e residential subdivision in a group of randomly selected subdivisions described by N physical characteristic measures. The procedure allows an electric utility company to use a single subdivision for their ann ual estimated cost differential between underground and overhead power distribution to new subdivisions. In the context of this procedure, t he term most representative refers to the tendency of the selected sub division to exhibit central values for the characteristic measures use d to describe the subdivisions, as well as to the degree to which the underground versus overhead cost differential for the selected subdivi sion corresponds to the mean cost differential for the group of subdiv isions considered. The methodology developed involves regression and c orrelation analyses of the relationships between the physical characte ristics and the cost differentials for all the subdivisions under stud y as well as definition of quadratic measures of atypicality for each subdivision. The reliability of the process is established using the c hi-square distribution of the atypicality measure. Data provided by a particular electric power company are analyzed, and experiences with t he procedure are reported.