ESCORT: Engineering Systems Classification and Ordering Technique

Out of the study of the general problem of optimization in the presence of many objectives, there is growing a new approach which consists of ranking in order of attractiveness a discret~ set of alternatives on the basis of their performance on a discrete set of specified criteria. There have been two principal lines of investigation of ordering methodologies. The one consists of proceedin~ by successive dichotomies of the set of alternatives until the irreducible subset (core) of equally attractive alternatives is isolated. The other consists of seeking complete (linear) orderings of the alternatives. This repart tries to pursue the second line of investigation. A methodology is proposed and implemented in the algorithm ESCORT (Engineering Systems Classification and Ordering Technique), based on the theory of inductive learning. It is a report intended to provide a preliminary treatment of the topic and show encouraging results of cpplication. *School of Civil Engineering~ PUrdue University, W. [4fayette, IN 47907 **Department of Computer Science, Purdue University~ w. Lafayette~ IN 47907