This week, our partner CQM published a great use case on "valorization." The article is in Dutch, so I asked ChatGPT to translate it.
https://cqm.nl/nl/cases/vips-maakt-planning-vion-datagedreven
In agriculture, food production, and food processing industries, "valorization" is crucial. It increases the value, utility, or profitability of a product. In the meat processing industry, this means utilizing every part of the animal as efficiently as possible and making the most out of the meat. It’s a highly complex process with many variables. Together with meat producer Vion, CQM developed the VIPS application, which has made Vion's planning data-driven.
Vion Food Group's largest production site is located in Boxtel. Vion is an international producer of meat, meat products, and plant-based alternatives. From this location, fresh meat and processed meat products are supplied to national and international retailers, food service industries, and the meat processing industry. Due to their scale and operation within demand-driven supply chains, Vion can supply large quantities of raw materials in various specifications, precisely tailored to market demand. In Boxtel, 2,700 employees work in two shifts every day, ensuring the production of safe, high-quality meat products and maximizing valorization.
A Complex Puzzle
Every day, Vion faces the challenge of planning how the supply is processed so that each product reaches its best final destination with the highest yield. This is a complex puzzle, explains Luuk van Boxtel, head of planning at Vion Boxtel: “We don’t produce a bottle of cola that’s always the same. We deal with varying supplies from different suppliers to meet an ever-changing market demand. Many variables determine how, when, and on which production line the meat must be processed. Manual planning is no longer feasible, so we wanted to build a smart, data-driven application.”
Matching and Scheduling
To understand the valorization choices, the process must first be explained. Broadly, there are two steps. After slaughter, each pig is automatically scanned to classify each part of the animal by value and assign it to customer orders based on the daily order package. This step is called "matching." The second step is "scheduling": deciding which production line will process what, when, and by whom. In Boxtel, there are dozens of production lines, each with around 20 employees. They must all receive specific tasks for particular cuts of meat while maintaining line continuity, minimizing changes or downtime. The planning for matching and scheduling is done daily the evening before production, presenting a massive challenge.
Under Pressure
Previously, 14 Vion planners in Boxtel manually handled matching and scheduling. “Valorization often came under pressure during this process,” says Max Houben, product owner of Vion's Innovation Hub, where specialists collaborate on data-driven improvements. “Decisions were sometimes made to complete customer orders for the day or run large batches, but this wasn’t always the best choice for valorization. While Vion uses standardized processing methods for its products, these standards were often ignored because planners were accustomed to processing orders differently for certain customers. In hindsight, there was often a significant difference between what we thought we were selling and what we actually shipped. Due to the complexity of valorization and the desire for consistency, our request to CQM was clear: make our planning data-driven.”
VIPS
“Together with Vion’s planners and specialists, we mapped out the variables,” explains CQM senior consultant Arno van den Eijnden. “Vion provided the database with the yields from different classifications and standard processing methods. With this data, we built the VIPS application: Vion Planning System. The application—developed with AIMMS software, a long-standing technology partner of CQM—calculates the optimal matching between supply and sales orders and creates a production plan for scheduling the following day. By following the plan, maximum valorization is achieved.” VIPS also considers feasibility during production, says Arno: “Previously, a customer order would preferably be scheduled on one production line. Due to valorization, VIPS can now schedule such an order across five different lines in five different batches. However, the algorithm ensures that there aren’t too many changes on a production line. It also considers the various capacities of each line, special transport conditions, and other factors. So, in addition to VIPS's primary task of valorization, the application also considers the practical feasibility of the plan.”
Yield Improvement
VIPS has been fully operational for a few months at Vion Boxtel. The question is, does the system deliver higher valorization? Max Houben says, “A few years ago, we ran a VIPS pilot at our smaller production site in Apeldoorn. The results there were significantly above the business case. In Boxtel, we still need to verify the results, but we expect it to deliver significant yield improvements here as well. We’re also seeing that we’re working much more uniformly in Boxtel. Since VIPS is based on Vion’s standard processing methods, we’re much closer to the benchmark yields than before. Every part of the animal is being utilized to its fullest.”
New Way of Working
So, do the planners in Boxtel have much more free time now? Luuk van Boxtel explains, “Our work is changing. When you make something data-driven, you need to maintain and validate the data, and that now takes up a lot more time. Plus, we’re still in the startup phase with VIPS, so we sometimes need to make adjustments, which requires a lot of effort from the planners. We also need to adapt to the new way of working.” “VIPS isn’t a magic button,” adds Arno van den Eijnden. “We still need planners to make good decisions. For instance, if the supply delivered differs from what was expected or the proportions are off, sales orders or transportation might need to be adjusted, or a line may need to stay open longer. It’s always a coordination process.”
Opportunities
Max Houben is very pleased with the collaboration with CQM: “CQM has been a great partner from start to finish, always constructively challenging our business processes. VIPS is a planning application with a one-day horizon. With the knowledge CQM has passed on, we now want to develop a tool with a longer horizon to further improve our planning.” Arno van den Eijnden sees even more opportunities for data-driven optimizations at Vion: “Internal logistics, transport planning, demand and supply forecasting models, bottleneck analyses, and simulations. In agriculture, food production, and food processing, there are numerous opportunities to achieve higher valorization with mathematical applications. We’re eager to tackle these challenges.”