Robert Boute, Professor of Operations & Supply Chain Management at Vlerick Business School, and Joren Gijsbrechts, PhD student at the Research Centre for Operations Management at KU Leuven, discusses how AI can enable supply chains to achieve sustainability.
The way in which companies organise logistics today is not sustainable. Supply chains are not always the most environmentally friendly and it is impossible for organisations to sustain a balancing act of the best possible speed, flexibility, cost and carbon footprint when it comes to the shipping and delivering of their goods. In an ideal world, organisations would be able to have a sustainable, cost-effective and efficient supply chain for their products – however, this is isn’t always feasible using the current available methods for shipping.
However, new technologies such as Big Data analytics and AI can help companies make a positive change, ensuring their supply chains run as efficiently and sustainably as possible. Utilising AI can have a dramatic effect on supply chains, helping organisations to benefit from the fastest, cheapest and most sustainable routes for shipping, and combining these seamlessly.
So, how can firms implement this? Well, it all focuses around organisations shifting to a sharing economy when it comes to their supply chains. Using AI, data and innovative algorithms can drastically improve the sustainability and efficiency of supply chains by enabling organisations to work together. In fact, there are three very specific areas in which these technologies can applied to create a smart, efficient logistics chain.
Collaborative shipping, otherwise known as sharing shipping, refers to the shared use of shipping and transport methods between organisations. This is something that we have explored at Vlerick Business School and KU Leuven, by developing an algorithm to help organisations better identify opportunities to share their shipping data and collaborate with other transporting firms.
Using GPS data, this algorithm logs the collection and drop-off points of shipping organisations. The system remains aware of the state of the environment at all times, with regards to shipping, stocks, transport methods and the costs. By incorporating the sharing economy aspect into this, organisations can share details of their supply chain with other firms.
For example, if a truck is delivering goods and services to a specific location, by inputting this data into the algorithm, the system is aware of the amount of stock in the truck, where it is travelling to and the costs of this travel. If the truck is partially empty for instance, by using the algorithm organisations which are delivering to the same location or a location en route can share their delivery method, not only cutting costs but reducing pollution too – making their supply much more sustainable. This could also be the case for making effective use of trucks return trips with empty loads too. Using the AI algorithm and the data inputted into this, organisations can identify these empty returning trucks and use them for their own delivery purposes.
Not all packages are equally as urgent to distribute. In fact, many have intentionally long delivery times and some packages can actually change in urgency after they have originally been shipped. Using the physical internet, organisations can adapt a synchromodality system, which involves combining a variety of transport methods in a sustainable way and taking into account the urgency of these deliveries, without comprising on the flexibility of the shipping.
Using a real-time data system, the transportation method of a delivery can be adapted whilst a shipment is en route, meaning that throughout its transportation the algorithm can select the most cost-effective and environmentally friendly supply chain in real-time, continuously shifting to the most efficient and sustainable delivery method possible.
Deep reinforcement learning
Deep reinforcement learning is a specific element of machine learning and involves training an algorithm to make the best possible decisions. This is done through a trial and error process, where the robot is guided to the correct decision through positive feedback on its actions. By positively rewarding the robot, it will learn to narrow down its random actions, and only repeat those that have a good outcome for the organisation.
Organisations using this deep reinforcement learning are able to train AI to make complex and positive supply chain decisions which involve a number of variables. In doing so, AI could determine the exact number of products to ship, when to ship it these and which mode of transport is best to use. This could also be used to train smart algorithms to support companies to collaboratively ship, use synchromodality and replenish an organisation’s inventory smartly, linking all of the aspects of AI together to create the most sustainable and efficient supply chain possible for the organisation.
Not only does integrating AI and new technologies to supply chains benefit the organisation, but from an environmental standpoint utilising these technologies can reduce pollution and the organisation’s carbon footprint, creating a much more sustainable supply chain and enabling a company to make a positive impact on many of the world’s pressing environmental issues.