This paper investigates the reliability of TRANSYT-7F optimal solutions fro
m two perspectives: (1) the effect of the selected optimization criterion o
n the values of various system performance measures; and (2) the extent to
which TRANSYT-7F signal plans remain optimal when evaluated in a microscopi
c and stochastic traffic environment. The well-established CORSIM model was
used for the latter test. A customized code has been developed that can au
tomatically generate statistical distributions of selected performance meas
ures in CORSIM based on various TRANSYT-7F optimization runs. The method ha
s been applied to a calibrated test network in Chicago consisting of 9 sign
alized intersections with 50 nodes and 74 links. It was found that while go
od correlation existed between the corresponding performance measures in TR
ANSYT-7F and CORSIM under light traffic, their correlation diminished under
congested flow. In particular, the best signal strategy identified in TRAN
SYT-7F was not the best in the CORSIM environment. The discrepancies are un
doubtedly due to the basic difference between the two models: TRANSYT-7F is
deterministic, CORSIM is stochastic.