Several methods used for the prerouting analysis of printed circuit boards are compared. The quality of a method is expressed as the correlation between predicted at~f actual values of objective measures of (in interconnection problem complexity. The experimental data confirm that some indicators of a board complexity which are often used in practice, such as component density, are highly unreliable. The PAP program is based on the modelling of the actual routing process. It has been experimentally proved that the PAP program provides reliable results which can be used for the detailed analysis of an interconnection problem.
The design of a printed circuit board can be divided into several consecutive steps: placement, pin assignment, prerouting analysis, routing and path optimization. Prerouting analysis is performed when an interconnection problem is fully determined; ie when components are located, pins are assigned to signals, the routing area is described and routing (technological) rules are defined. The main task of prerouting analysis is to predict an interconnection problem complexity.
The topic of prerouting analysis has been a subject of several papers and it has been vigorously debated at different conferences, but no meaningful data has been offered in support of particular viewpo!nts. The present work is a comparative study of the reliability of various board complexity indicators on a statistically significant experimental basis. In order to have criteria for comparison, the objective measures of a board complexity are first developed.