Network representations of performance analysis parameters and relationships are described. In these models, nodes represent variables or other information and links represent relationships. The intersection performance analysis procedures implicit in the Highway Capacity Manual are formalized as an example of a network representation. These representations make use of semantic net and frame representations originally developed in the field of artificial intelligence. The purposes of network representations of analysis procedures are (1) to aid flexible, computerbased implementations;(2) to facilitate interpretation of results; and (3) to provide a check on the completeness and uniqueness of particular analysis procedures. For computer-aided analysis and design, these representations provide a means to provide localized or variable specific representation of" intelligent" solution schemes.

Performance analysis for many transportation facilities is complicated by the numerous design characteristics, facility attributes, and measures of effectiveness involved. For example, the evaluation of the performance of an intersection may involve over a hundred separate parameters representing input data, intermediate analysis results, and desired final results. In this paper, we discuss network representations of analytical relationships and procedures. These network representations can provide procedural information for analysis, aid in the interpretation of performance, and serve as a check on analysis aids. They can also serve as representational and analytical building blocks for knowledge-based expert systems. As an example of network representation, we formalize the intersection performance analysis procedures implicit in the High-