A human-computer interactive scheduling system is developed with an automat
ic scheduling function. This function is designed based on the schedulers'
experiences which have been collected in a previous study. By conducting an
experiment, the total setup time and scheduling time of the subjects with
the automatic scheduling function and without this function are compared. T
he result reveals that the total setup time and scheduling time of each cir
cumstance are significantly reduced using the automatic scheduling function
, especially in more complicated circumstances. In addition, the hypotheses
tests of variances in setup time and scheduling time reveals that the auto
matic scheduling function could balance the scheduling skills of human sche
dulers. Moreover, a brief discussion of total setup time between four commo
n dispatching rules and experimental data is presented. For further study,
instead of developing the entirely automatic scheduling system, we might de
velop a system which can properly balance the work of the computer and the
human by providing the assisting functions or information. However, we can
still combine the advantages of simulation-based methods and AT methods to
the human-computer interactive scheduling system to make the scheduling mor
e efficient.