To simplify the problem, further assumptions are made. (1) The volume of the delivery products and pickup
products are the same. This means that ‘‘pileup’’ issues in a vehicle are not discussed. (2) The time to start a
service should be shorter than the time it takes for a vehicle to travel from the DC to the customer location.
This is because if the travelling time of a new vehicle form the DC to the new order is too short, the new order
will be assigned to a new vehicle rather than to the dispatched vehicles. This guarantees that the environment
is dynamic. (3) The time window constraints include both outer and inner service time window constraints. An
order should be served within the inner service time window, and should be rejected if it cannot be fulfilled
within the outer time window. (4) This study assumes that vehicles travel at a steady speed, and travelling time
is the distance divided by the travel speed plus 10 percent for random deviation. There is no significant differ-
ence in traffic conditions. (5) The number of vehicles is 10 and the moving speed is 100 units of distance per
unit of time. (6) Each vehicle returns to the DC to reload products when it has insufficient products to serve
more customers and to unload products when it is fully loaded. There is no limitation on the frequency of
returning to the DC.