5. Conclusions and Future Work The modification of the continuous time scale to a discrete one was crucial for scheduling the tanks, because the different steps of their cycle have shorter times than the arrival of the batches from the pipeline. The Rolling Horizon strategy allowed the scheduling of a month of operations by dividing it into smaller periods. It is concluded that the number of periods for the Rolling Horizon depends on the month under study. When the model was applied to the real case study (data from the Portuguese CLC company) an improved result was obtained in terms of percentage of extremes ratio (full or empty state) for all products and tanks compared to the company’s proposal. As a future work it is desirable that the model converts automatically the input data into shifts, in order to accelerate the scheduling problem. It would also be desirable the automatic transfer of data between consecutive periods of the Rolling Horizon.