Kennen Sie das Bierspiel? Falls Sie jetzt an Bierpong denken, sind Sie falsch abgebogen – oder haben das so einfache wie effiziente Rollenspiel noch nicht kennengelernt. Beim Bierspiel nehmen die Teilnehmenden verschiedene Positionen in der Supply Chain ein. Das erklärte Ziel: Die Kosten der gesamten Wertschöpfungskette so gering wie möglich zu halten.Durch die Tatsache, dass wie im echten Tagesgeschäft die einzelnen Beteiligten einer Lieferkette ihre Informationen nicht vollständig austauschen, sondern lediglich die Bestellmengen übermitteln, liegt der Fokus meist nur auf der Bewältigung der eigenen Situation, der jeweiligen Rolle.
Dies führt dazu, dass sich die Bestellmengen sehr schnell aufschaukeln, wie man es vom Bullwhip-Effekt kennt. Trotz einfachster Regeln haben selbst erfahrene Logistiker Schwierigkeiten, das Aufschaukeln der Bestellmengen zu verhindern. „Die Erkenntnis dieses Rollenspiels ist, dass die Systemstruktur das Systemverhalten bestimmt. Probleme in der Supply Chain werden also nicht vorwiegend durch externe Faktoren verursacht, sondern durch den fehlenden Informationsaustausch ausgelöst“, erklärt Eric Weisz, Gründer und CEO des Start-ups Circly, der derzeit am Institut Transportwirtschaft und Logistik der WU Wien promoviert. Er untersucht, wie man künstliche Intelligenz nutzt, um den Bullwhip-Effekt in Lieferketten zu glätten. Weisz legt dabei den Fokus auf die Management-Perspektive, da die Technologie selbst bereits gut erforscht wurde, wie er sagt. Das Start-up Circly hat Weisz gemeinsam mit Armin Kirchknopf gegründet, um KI-Anwendungen für Supply Chain Mitglieder von der Produktion bis hin zum Einzelhandel kosteneffizient und kinderleicht zugänglich zu machen.
Concern about the data
Almost all events on the topic of artificial intelligence talk about data as a problem. And not the lack of data, but above all the transfer of this data. "The problem with the supply chain is therefore the information situation," says Weisz. The fact that most companies lack trust and are unwilling to share information leads, among other things, to various forecasting methods within the supply chain, which can cause a bullwhip effect either through the use of incorrect data or imperfect forecasting techniques. Other consequences include an increase in storage or manufacturing costs, replenishment times or transport costs.
The bullwhip effect could be stopped if all parties involved in the supply chain adopt a cooperative mindset and exchange data voluntarily and efficiently. Eric Weisz also has a prime example to share with Dispo: "Procter & Gamble and Walmart, both major players in their industries, have found a way to utilise information technology by sharing data about their joint supply chains. As a result, activities along the supply chain are better coordinated and there is less need for stock, but a higher yield by focussing on the needs of the end customer."
Can artificial intelligence avoid the bullwhip effect?
Artificial intelligence allows supply chain players to look into the future by recognising trends in a data-driven and unemotional way, thereby bypassing the poor or sometimes non-existent communication between the individual participants in the supply chain.
One application example is timeseries forecasting. Here, historical data is combined in the form of a timeline and the technology recognises the dependencies between the data. Because the information on various factors, such as seasonality, weather, events and the like, is known in advance, the AI-supported software recognises the combination of these factors and determines the actual demand. The quality of the AI results is determined by the input data. Around three years of historical data are required to get started.
As a result, the producer is no longer dependent on the retailer's forecast, but can plan precisely by using AI. This is not based on simple statistics and human experience, but on many different factors.
But time series predictions are not the only application example; in addition to the bullwhip effect, other supply chain problems can also be solved or contained using artificial intelligence.
For example, personnel deployment planning can be optimised using an efficiency key by calculating the effort in advance, or route planning can be made more efficient to prevent empty runs. When it comes to customs clearance, artificial intelligence is able to automatically cross-check, complete and process customs declarations. "Another field of application is stocktaking, which may require robots. One exciting example is Ubica from Germany, which uses a robot and suitable AI to take stock," explains Eric Weisz.
Logistics industry as a pioneer in AI
The logistics industry has long recognised the importance of artificial intelligence, if a recent Bitkom study is to be believed. According to the study, one in five logistics companies in Germany uses AI - either in demand forecasting, sales planning or transport optimisation. A further 26 per cent are planning to do so or are discussing it.
The results come from a representative survey of more than 400 logistics companies with 20 or more employees in Germany. This refers exclusively to logistics service providers, i.e. transport logistics (land transport, shipping, aviation), warehousing and postal, courier and express services.
In comparison, only nine per cent of German companies use artificial intelligence as a strategic component. "The logistics industry has recognised the potential that AI offers in both warehousing and transport and is using the technology in a targeted manner to increase efficiency, quality and productivity," comments Bitkom CEO Bernhard Rohleder on the results.
More than half of logistics companies (58 per cent) believe that artificial intelligence will be widely used in the future and will take over many tasks in logistics, such as planning routes or avoiding empty runs.
In the logistics sector, the majority of companies (56 percent) consider themselves to be pioneers in the field of digitalisation - in comparison, this figure is only 32 percent in the economy as a whole.
This article was published on 22.11.2022 at https://dispo.cc/lieferkette/wie-bier-zu-einer-resilienten-supplychain-beitragen-kann/
Credit: Michaela Holy-Zwickelstorfer
EN 
DE