KI-basierte Planung als Regelkreis für die optimale Produktion (Bild: Adobe Stock, Grispb)

AI-based Planning as a Control Loop for Optimal Production

Guest author: Kevin Klaus, Executive Manager, Product Management at MPDV

In times of increasing demands on fast turnaround, costs, and flexibility, traditional production planning is no longer sufficient. With the Advanced Planning and Scheduling System (APS) FEDRA from MPDV and AI-based planning, manufacturing companies are relying on an optimizing control loop. An optimizing control loop ensures that production is continuously adapted to changing conditions – automatically and with maximum cost efficiency.

Production planning generally focuses on the overall goal of manufacturing end products quickly, reliably, and in the best quality at minimum cost. The decisive factor here is that machines are running at maximum capacity, only good parts are produced, and material stocks remain as low as possible to minimize tied-up capital. Careful planning also reduces throughput, idle, and setup times, making adherence to deadlines and delivery times a competitive factor.

How APS FEDRA implements AI-based planning as a control loop

With the integration of AI-based planning methods, APS FEDRA enables continuous optimization of all processes. The core element is the dynamic incorporation of all framework conditions:

  • Machine utilization & processing times: AI is used to calculate perfect machine-tool-article combinations that keep throughput and setup times to a minimum.
  • Automatic sequence optimization: Similar orders are sorted so that as little retooling as possible is needed.
  • Minimized transition times: Operations in different process steps follow each other in quick succession, significantly reducing waiting and storage times.
  • Maximum adherence to deadlines: AI recognizes critical deadlines and prioritizes orders accordingly, ensuring that delivery deadlines are met or delays are reduced to a minimum.
  • Capacity, tool, and material management: The availability of personnel, tools, and materials is factored into current planning throughout the entire process.

Ever-changing conditions, whether due to sudden machine failures, staff shortages, or delays in raw materials, require planning that is continuously updated. This control loop can only be achieved with real-time data: By integrating machine data collection, personnel time management, and tool management in real time, APS FEDRA knows the actual status of your operations at all times. MPDV’s HYDRA Manufacturing Execution System (MES) can function as a source for providing real-time data.

Real-life example: the planning cycle stays up to date

Picture this: everything in your production facility is running according to plan. But suddenly, a vital machine breaks down or an employee falls ill at short notice. Traditional planning systems quickly reach their limits in this situation: Existing plans have to be painstakingly adjusted, often manually, which can lead to delays, unplanned downtime, and ultimately increased costs. The question of how such incidents affect subsequent work steps and delivery dates often remains unanswered in conventional systems.

The situation is quite different if you use AI-based planning as provided by APS FEDRA. As soon as an unexpected event occurs, its effects are immediately identified – whether that be machine failure, material delays, or short-term rescheduling of personnel. Artificial intelligence (AI) automatically calculates how current planning needs to be changed based on new data, so that the overarching goals such as adherence to deadlines and minimal costs can still be met.

For example, FEDRA not only shifts the affected operation, but also evaluates the domino effect on all subsequent work steps throughout the entire production network. The planning software therefore not only adjusts individual machine assignments but also reorganizes entire production chains if necessary. One option, for example, is to switch to alternative machines or bring forward specific orders to avoid bottlenecks. All relevant conditions, such as tool and material availability or staff qualifications, are directly factored in.

At the same time, the concept of fixing production ensures that planning remains stable for a set period of time, e.g., for the next three or five days. This gives employees in production, material issue, and the workshop the necessary planning security and helps them to organize their work reliably. Planning is also adjusted flexibly to new events.

Marcus Pottendorfer, Head of Supply Chain Management at KLINGER Fluid Control GmbH, uses AI-based planning and confirms: “Thanks to AI-powered functions in FEDRA, all our operations are planned in a coordinated fashion, while taking all constraints into account. This resulted in a significant increase in finished product output of +25%.”

Thanks to real-time data collection, AI optimization, and targeted planning, production remains responsive, efficient, and on schedule at all times – even in the event of unforeseen disruptions.

Benefits of AI-based planning as a control loop

State-of-the-art manufacturing requires flexibility and reliability, which is exactly what AI-based planning as a control loop provides with APS FEDRA. But what does “control loop” mean in this context?

The principle of the control loop is a recurring control process in which production is constantly monitored and continuously adjusted to the optimum setting. In the case of APS FEDRA, this means that operating data such as machine statuses, personnel availability, material stocks, and tools are continuously recorded during ongoing operations and fed into the system. Real-time information shows exactly what the current situation in production is and highlights any deviations, delays, or disruptions.

The actual AI-supported replanning takes place at night. AI processes all the data collected during the day: new customer orders, machine downtime, staff availability, material bottlenecks, and other changes are intelligently compared with each other. This automatically generates a new, optimized production plan for the days following the fixation. So, your production team starts each morning with an up-to-date plan that is as accurate as possible.

The control loop of dynamic AI planning clearly shows how automatic and autonomous processes interact. (Image: MPDV)
The control loop of dynamic AI planning clearly shows how automatic and autonomous processes interact. (Image: MPDV)

The closed control loop is evident in that this comparison is repeated regularly – every night whereby planning and reality are continuously aligned with each other. Unforeseen events that occur during the day are mitigated at short notice by direct intervention from production staff or by supportive suggestions. Artificial intelligence then performs long-term optimization across the entire planning horizon overnight. The synergy between these factors enables maximum efficiency, adherence to deadlines, and cost control.

Michael Wetzel-Staar, Production Manager at VACOM Vakuum Komponenten & Messtechnik GmbH, knows this all too well. He also uses AI-based planning with APS FEDRA: “When we confirm an order, we can tell the customer when they will receive their bespoke component, and we can keep to that deadline. The quality of our deadlines will improve significantly.”

Future outlook

With the ongoing development of digital manufacturing technologies and increasing computing power, companies are now on the threshold of fully continuous, automated optimization in real time. The goal is for AI systems to evaluate new data continuously in the near future, not just at night but throughout the entire operation, and to instantly realign planning whenever there is a relevant change. All of this makes for even more agile, resilient, and efficient manufacturing.

By using AI-based planning as a control loop, manufacturing companies are paving the way for future-proof, flexible, and data-driven production. This means they are ready to meet the challenges of the market at any time.

Leave a Reply