With many factors to account for and an extensive number of events to sequence, planning and scheduling for pharmaceutical sites is incredibly complicated to execute. This complexity has made way for multiple planning and scheduling software systems to hit the market over the years in order to help plants manage their operations. These systems use a variety of different technical philosophies to generate production plans for your plant. Two of the most common approaches are discrete event simulation and math programming, which we’ll discuss more in depth in this article.
What Is Discrete Event Simulation?
Discrete event simulation is an older and very common technology that software providers like SchedulePro use to create a manufacturing schedule. In fact, the software you currently use for planning and scheduling may be run with this type of approach. In discrete event simulation, every action involved in the manufacturing process is assigned a particular time stamp to create a chronological sequence. The software can then run simulated events through the sequence to observe and record predictions under different circumstances to help the site identify its schedule.
There are some advantages to building a schedule with this method. When every action, component, and constraint is included in the model, it can produce accurate results. It also supports running simulations under many different conditions, such as with different amounts of raw material on hand, which can offer helpful insights. However, discrete event simulation also comes with some significant disadvantages that are difficult to overcome.
Drawbacks of Discrete Event Simulation
At every point in generating a production schedule, there are an incredible number of decisions that must be made. To generate a potential schedule, discrete event simulators will follow one set of decisions at a time, using rules of thumb to dictate each decision. Unfortunately, the rules of thumb don’t apply to every situation, so simulators require a lot of attention and maintenance from programmers to remain functional. Similarly, discrete event simulators also typically offer features that are standard across many industries, which means the software may not be tailored to the specific problems pharmaceutical manufacturers experience. As a result, the insights the tool produces may not address the real bottlenecks the plant is facing or may require substantial effort from the end user to keep them working.
Additionally, the sequencing nature of discrete event simulation means that the backend of the program can become extremely complex very quickly. For pharmaceutical plants in particular, the many processes and moving parts of their systems often result in a tool that is very convoluted and can only be managed by one or two expert users. This can be a problem because the tool becomes slower and more difficult to use. If the tool isn’t user-friendly, adding updates whenever the plant model changes becomes very time-consuming.
The complexity of discrete event simulation becomes especially challenging when manufacturers need to add in a new constraint to the system. Because of the time, resources, and processing power it takes to update the model, many manufacturers choose not to make updates regularly. However, if the tool doesn’t account for all current constraints, such as hold times or material expirations, the schedule it produces is less likely to be accurate and more likely to result in errors, delays, and less output.
How Math Programming Adds Flexibility & Accuracy
One alternative to scheduling with discrete event simulation is math programming technology. Math programming is a more contemporary solution than simulation technologies, as it offers the ability to move back and forth in time to find the best combination of decisions based on your plant’s constraints. Most companies avoided powering their scheduling software with mathematical programming because the equations needed to build the tool require advanced expertise. Additionally, they struggle some struggle to overcome the computational cost of formulating and solving the complex math equations that generate the optimized schedules. However, if a piece of software can overcome those barriers, it will provide much more flexibility within the model, identify more opportunities for optimization, and require less effort from the end user.
Our team at APCI has spent the last three decades developing VirtECS, our planning and scheduling tool with a math programming framework. VirtECS is built on highly sophisticated algorithms written by our PhD engineers and specifically designed to model real-world manufacturing processes. Our tool overcomes the barriers associated with developing the equations because it understands and models the process physics of real manufacturing plants, and it incorporates additional data gathered from ERP systems, MES, and process historians. It also uses augmented intelligence to continuously improve the program and advance its scheduling capabilities.
As a result, the VirtECS interface can be updated much more quickly when plants need to add new constraints, products, or equipment to their model. It can also handle more specific constraints than many of our competitors, such as upstream train selection, maintenance schedules, and product changeover activities. With a more detailed model, VirtECS can provide insights on how plants can account for uncertainty and variability to create the most optimized and accurate production schedule. If you’re interested in learning how VirtECS can transform modeling, planning and scheduling for your manufacturing site, sign up for a short introductory call with one of our experts so we can learn more about your unique needs.
