This paper raises issues related to statistical and reliability-based model aspects of the 1986 Guide for Design of Pavement Structures. The use of reliability-based models is well accepted and their application in pavement design is proper. Concerns are raised related to the proper use and proper development of the reliability-based model used in the guide along with concerns about problems in terms of statistical decision theory. It appears that some unintended misapplication of some reliability-based model aspects may have resulted. A description of various underlying assumptions and complications is given in regard to reliability theory and statistical design theory to illustrate the problems of attempting to use a new technique without having full understanding of the technique or its underlying requirements and their ramifications. Two specific concerns are raised. The first concern is with the model used for staged construction. It appears to be unrepresentative of the actual situation. The second concern is with the use of variable levels of reliability to represent consequences of failure for the various functional classes of road. Another technique might be more appropriate in terms of design. The understanding of reliability level in terms of the guide is raised. The problems not addressed in the guide that have wide importance are illustrated, including the problem of estimating reliability level over design life, after rehabilitation, for overlay, and after design life. While the implementation of reliability-based models is the issue raised with respect to the guide, the use of reliability-based models is well accepted and is a proper application. What is needed is future revisions.
Abstract