This scenario could become very real in the near future: A particular car part (e.g. a shock absorber) continuously collects data about its current status using sensors, and autonomously alerts the car manufacturing company in case of imminent replacement – just before the part fails, of course. The car manufacturer receives a detailed specification of the car type and the address of a car repair shop, with which the part has already scheduled an appointment (in sync with the car owner’s calendar in his smart phone and a defined delivery time). The machines in the car factory autonomously configure themselves based on the data received, produce the respective car part and send it to the defined destination.
The fourth industrial revolution
The scenario described above originates from the vision of a new technology trend, which currently is on the agenda of some research institutions and companies. Industry 4.0 is the focus of the high-tech strategy of the German Government, which serves as the main driver of development in Europe. There is also a similar concerted research effort in the U.S. under the term Industrial Internet. Whatever title you give it, the vision is nothing less than the claim to a fourth industrial revolution.
The first industrial revolution started with the invention of the steam engine in the 18th century, when machines took over procedures formerly performed by human laborers to a greater extent. The second era of technical revolution resulted in an increase of productivity at the beginning of the 20th century, when assembly lines achieved a major breakthrough with the production of Ford’s model T. And then there is the third era we live in right now, sometimes referred to as the Digital Revolution, which has led to the entrance of electronics and information technology into the production process.
The drivers for the next revolution are the developments in the field of robotics and the pervasiveness of internet-ready devices and the corresponding infrastructure. The integration of almost any electronic device via the internet, the so called Internet of Things (IoT), will become a part of everyday life. In the case of consumer electronics such as TVs or Blu-Ray-players, wireless communication is almost standard, and every day, more devices are following suit. Part of this development is a result of take-overs, for example Google’s purchasing of Nest Labs, and more recently Smart Things being bought by Samsung. These moves have drawn great attention from the media, but this is only the beginning. According to the consulting firm Gartner, approximately 30 billion devices will be connected to the Internet via an IP address in 2020 – many of them will not be smart phones or similar devices, but rather autonomous machines.
Customer-specific mass production
These communicating machines will – in line with the vision of Industry 4.0 or Industrial Internet – be the heart of our factories. In the end, so called Smart Factories will be designed to produce almost automatically – without human interaction involved. This is projected to achieve an increase in productivity of over 30%. One of the main goals of setting up automatic factories is to create more flexibility in large-scale production. Customer-specific mass production (which is already the standard for new cars) of formerly “off the rack” products may become possible in the future. In the past, only very big batch sizes were profitable – the future goal will be a profitable batch size of 1, thus making contract manufacturing possible for practically everybody. A very ambitious goal, based on the assumption that robots will be able to configure themselves automatically using transmitted data and will not need to be changed over for every individual order. This is the vision of the Smart Factory.
Operations Research is the Ghost in the Machine
Advanced robotics, wireless communication e.g. via RFID and almost infinite data storage capacities – all prerequisites for the Smart Factory – seem to be readily available. A very important ingredient in this scenario is often neglected: intelligent decisions. Even today’s complex production environments cannot be managed with simple If-Then heuristics. The Smart Factory needs to think ahead, use its resources as efficiently as possible and cope with exceptional circumstances. These features are offered by forecast and optimization algorithms from Operations Research, which today, embedded in software, already optimize supply chain processes in many production companies. Thus, Operations Research can be tagged as the “Ghost in the machine”, using the famous phrase of the British philosopher Gilbert Ryle. In a way, the machines are the body of the Smart Factory, whereas Operations Research is the intelligent mind (or the “ghost”).
Today’s production in version 3.0 can already be very “smart,” all the more reason why The Smart Factory will have to rely on Operations Research. It is highly likely that the Operations Research algorithms will be able to achieve even better results in such an environment. Since each product and each process will be digitally documented in detail, the algorithms will be able to process a higher volume of data than today and thus achieve even more precise forecasts and even better optimization. In the end, the machines in the smart factory will thus make even more intelligent decisions than would be possible today.
If you think of the Smart Factory as a particular part of a supply chain, the full digitalization of processes, as described in the shock absorber example, allows for a respective full horizontal integration of the supply chain. This way, Operations Research provides a holistic optimization, of which suppliers, producers and retailers profit in equal measure.
Automatic inventory management
The road to the Smart Factory and Industry 4.0, however, is still a long one, because the developments are still in an early phase and really convincing examples (beyond prototypes) are hard to find right now. But it is still worthwhile to envision the new challenges supply chains will face in the future. Since I personally focus on inventory management topics, the procurement processes in production are of particular interest to me. Machines that will have to automatically keep supplies coming have to make intelligent decisions – here are three examples:
– Even if highly-developed robots allow for customer-specific mass production, it should not be forgotten that this type of production coupled with high customer expectations regarding delivery times will likewise create high demands on the management of purchased parts and raw materials. Although the machines are supposed to be able to automatically identify low stock levels and make new orders by themselves, without anticipatory planning, there will surely be bottlenecks which will delay production.
– In the end, a Smart Factory will have limited capacities, much like today’s factories. A purely order-based production will – depending on product types and hard customization – not be possible in the case of seasonal high demand. In these cases, at least pre-production of semi-finished parts will be necessary, combined with the corresponding requirements on procurement of raw materials and scheduling of production orders.
– Automatic orders will become standard in Industry 4.0. An increasing automation of orders has already led to lower costs per order. But this does not mean that the total order costs are also decreased; order sizes have merely become smaller, but more frequent. Besides the “triggering” transaction costs, the “downstream” transaction costs also need to be considered e.g. handling costs of incoming goods. The challenge here could be finding the right balance between storage and transaction costs.
The challenges of these cases can already be solved today by algorithms from Operations Research. Of course it is hard to forecast all consequences of the developments since we are still in an early stage of Industry 4.0. But one thing is for sure: in the end, the customer will profit immensely from these advances. Many products will be easily customized according to the customers’ desires, delivery times will be reduced, and full transparency of the whole supply chain due to the continuous digital documentation will be available. Going back to the shock absorber example I gave earlier, the only thing the customer needs to do is keep an eye on his calendar, so that his shock absorber or electric window lift does not schedule an inappropriate appointment at the car repair shop.
What supply chain challenges do you see on the horizon as we approach the Industry 4.0/Industrial Internet era?