Conaprole, the biggest dairy production company in Uruguay, produces more than 150 SKUs in their ice cream plant, using five production lines, and up to five different packaging configurations for each line.
The company plans ice cream manufacturing capacity on a 12-month rolling basis as part of the Sales & Operations Planning process, and the demand plan varies a lot due to seasonality. The factory management needs to prepare the production lines for the peak season during the low season, taking into account product shelf life and warehouse freezing capacity and costs. The factory was often unable to meet the high season demand which generated stock-outs, and the management found it very difficult to quickly reschedule their detailed plans due to the challenges they faced.
Other factors, including bottlenecks and constraints in the manufacturing processes and variations in human resource availability that can occur randomly, made capacity planning analyses even more difficult.
The management’s challenge was to be able to reformulate their plans in order to balance supply and demand and make sure they would avoid stock-outs in key products. They also sought ways to optimize the use of their manufacturing capacities. Ite Consult found simulation modeling to be the best tool to carry out manufacturing optimization and provide Conaprole with the solution to these problems.
The objectives of the manufacturing simulation model were:
- To analyze various production scenarios for the following twelve months of changing demand.
- To carry out manufacturing optimization to avoid stock-outs for all SKUs.
- To optimize a production line’s capacity using the optimized production.
Using AnyLogic’s discrete event modeling capabilities, the consultants designed and developed a manufacturing optimization model. It was integrated with the company’s S&OP planning platform and SAP Material Management and Production Planning. The created solution included three experiments with the model of the production system. Each of them addressed one of the objectives above and helped solve the business problems questioned.
In the first experiment, the model examined the initial manufacturing capacity plan, detecting stock-outs and backorders that could be expected if production followed this plan. It allowed the management to explore production needs based on demand and initial inventory. This experiment also gave users the ability to find out, by manually modifying parameters, how different situations could impact performance, for instance: the need to close lines during certain periods, the necessity to modify equipment efficiency, extend resource availability, or change human resources’ schedules. Users could manually change priorities of SKUs and analyze the expected impact of such actions on revenue (costs associated with stock-outs’ differed by SKU). Additionally, they could define minimal used manufacturing capacity and some other policies.
The parameter variation experiment ran the model of the system 100 times and searched for the solution to fulfill the demand and keep products’ shelf life as long as possible while minimizing warehouse costs.
The last experiment optimized the use of lines by freeing production capacity in peak periods. Manufacturing was scheduled as close to the beginning of planning periods as possible in order to leave free capacity in all manufacturing lines as a buffer.
Manufacturing Capacity Planning — Model Statistics
The input data included:
- Demand per SKU
- Inventory levels per SKU
- Batch size, priorities, life shelf, warehouse space occupation
- Overall efficiency and mean absolute percentage error (MAPE), if needed
The system considered the following:
- SKUs’ allocation by lines and sublines
- Lines’ and sublines’ capacities
- Line and packaging restrictions
- Warehouse limitations
- Production times
- Production schedules
All simulation results segmented by month and SKU were exported to Excel. Additionally, the model presented changes in demand and inventory levels by month in histograms. It also gave information about stock-outs, in case they happened.
By using the model, the Conaprole management was able to:
- Discover the processes in each production line by SKU.
- Optimize manufacturing capacity planning to better meet demand while maximizing product shelf life and minimizing warehouse costs.
- Improve production line utilization to secure additional manufacturing capacity in case of an increased demand.
The simulation model provided the management with the insight to choose the solution that would increase revenue and minimize the risk of stock-outs.