Scotsburn Dairy Group Uses a Hierarchical Production Scheduling and Inventory Management System to Control Its Ice Cream Production

In this paper, we discuss a hierarchical production planning approach to schedule ice cream production, a continuous batch production process with sequence-dependent setup times and highly seasonal demand. We use mixed-integer linear programming models to optimize production plans for long-, medium-, and short-term planning. The long-term model, the monthly model, is used to plan aggregate production and inventory levels for the year to meet demand each month at the lowest cost possible. The medium-term model, the weekly model, is used to disaggregate the long-term plan to minimize weekly setup and holding costs over a 13-week period. The short-term model, the daily model, is a detailed scheduling model that determines an optimal daily production sequence for the products and run lengths determined by the second model, while meeting the labor schedule defined by the first model for the upcoming production week. When used together, the three decision models produce feasible results at each stage, and short-term operations reflect the goals of the long-term plan. Synchronizing the model breakdown with Scotsburn’s management hierarchy provides support at each decision-making level. The hierarchical plan reduces costs, improves production efficiency, and creates better linkages between the decisions of each management level.