Abstract
Every year a great amount of money is expended for the rehabilitation and reconstruction of roads and pavements in most countries. One of the most usual methods in evaluating pavements distresses is to determine the PCI of the Pavement Condition Index. As a result of large number of variables and complicated decision - making algorithm using the information obtained in this method, may have some difficulties. Presenting an analytic theoretical method mixed with the PCI method may be the bases for the development of a theoretical empirical method in evaluation of concrete pavements distresses & can remove the difficulties. This paper presents a new approach to the rough-set theory in a Pavement Management System PMS database that enables pavement engineers to discover the shortest subsets of condition attributes having quality equal to the general quality of defined characteristics in the information system, to assess and describe pavement conditions, and to derive decision rules for rehabilitation and reconstruction of the pavements. To evaluate the results, the best algorithm of defined attributes in the information system is determined by making use of Artificial neural network (ANN) method and the result is compared with rough-set ones. The results of the research indicate that the rough-set theory has a better and stronger operational capability in identifying the effective parameters for the severity evaluation of typical distresses in pavements and in decision-making for selecting the type of repair.