The accident of a single freighter in the wrong place was enough to bring transport routes to a standstill. Of course, the pandemic has also contributed its part to the situation. Information about disruptions available at an early stage helps to avoid shutdowns – and artificial intelligence is becoming increasingly important for this purpose.
(Image: Blue Yonder, Inc.)
Disruptions in the supply chain typically result in certain parts or products not arriving at the desired location at the desired time. This disrupts production schedules and delivery commitments, which in turn affects customer satisfaction. In the worst case, penalties for late delivery can threaten. Even simple countermeasures can ensure greater safety. In the first step, it helps to have a complete overview of the supply chain, ideally in real time. Expert systems are available that combine data from the company with external information such as from ports, weather data, customs duties, etc. aggregate. The data needed for this can come from ERP systems, from transport management systems, from warehouse management or directly from the production machines. The disadvantage is that the real-time view is only ever a snapshot in time. On the other hand, it creates transparency about deviations and their consequences. For example, if a shipment of parts has to be diverted from Factory A to Factory B, there may be a shortage at the end in Factory A if it does not have safety stock to fall back on. Since disruptions often upset long-term plans and certain products have a long lead time, it should be known as far in advance as possible that a delivery will be late.
ETA and PTA
A common metric in logistics is estimated time of arrival (ETA). This indicates, for example, when a truck will reach its destination at the ramp. In production, however, predicting the timeliness of upstream steps in the supply chain is important. This Predicted Time of Arrival (PTA) metric provides important information, especially for products with long lead times. Companies that know four weeks before a shipment is scheduled to arrive whether it will be delayed, and by how much, have enough time to take countermeasures, reroute shipments or find alternative sources.
However, long-term planning reliability combined with real-time insights into the end-to-end supply chain can only work if the tools used provide clear insights. This is where dashboards and maps come in. Table views are often of limited help with complex supply chains. At the same time, more and more young employees are joining the companies, bringing with them a certain degree of digitization. As a result, there is also a certain demand on the tools used in terms of overview.
Intervention through AI
The shorter a disturbance occurs, the more important is the real-time insight. It’s even better if disruptions can be eliminated automatically. Manual interventions can be very time-consuming because many imponderables, consequences and possibilities have to be reconciled. Human employees need days to do this. However, if artificial intelligence is used, this process is reduced to a few minutes. An example of this is the sudden delay of a container ship. The AI knows what the ship has loaded, where the supplies are needed and can autonomously reroute flows of goods, access safety stock and prioritize materials. For AI-driven automated interventions in operations to be trusted by employees, they should know how the decision-making process comes about. AI must therefore be transparent and its decisions understandable, explainable and reproducible.
In order for an AI solution to accomplish this, it needs data. It is the same data that a supply chain monitoring tool requires. In order to be able to use this data, it is important that it is of good quality: it should be cleanly prepared and accessible. Cloud-based platform solutions to which all players along the supply chain have access can do this, for example. They are used to plan, control and monitor the supply chain. The control tower acts as an interface.
Production and logistics move closer together
As digitization progresses, production and logistics are moving ever closer together. Artificial intelligence helps to merge data from both areas, orchestrate and automate operations between both disciplines. Control towers will also become more important to manage supply chains and see where quick intervention is needed.
Date:2. February 2022
Authors:Gabriel Werner is Vice President Manufacturing DACH at Blue Yonder.