Abstract
This paper addresses the problem of assigning drivers to cover tasks when the tasks are the movements of containers. The tasks must be started within certain time windows and their servicing times are uncertain while the decisions are made over time. These decisions can be changed if new information is received. The problem is formulated in a stochastic optimization framework with the objective of minimizing the costs of current driver-task assignment and the expected future costs. A time-window sliding solution procedure is developed to estimate the expected future costs by solving the minimum cost flow problems iteratively. Furthermore, the results of the numerical experiments that assess the efficiency of the algorithm and the benefits of considering uncertainty in the model are reported.