A discrete-time predictive dynamic user-optimal departure time/route choice
problem has been formulated and solved with the nested diagonalization met
hod by Chen and Hsueh (1998a). That model did not impose constraints on eit
her origin departure times or destination arrival times, which are typicall
y required for regular work trips. For the purpose of better representing p
ractical situations, the time-window constraints associated with both depar
ture times and arrival times are incorporated into the previous dynamic use
r-optimal departure time/route choice model, resulting in a dynamic user-op
timal departure time/route choice model with time-window, where the off-pea
k and peak phenomena can be properly differentiated within the analysis per
iod, and the influence of earlier departures on later arrivals with respect
to perceived route disutilities can be appropriately modeled.