Numerical Recipes for Supply Chain

00:08 Introduction
00:23 Joannes, why do you think it is important to revisit the concept of “numerical recipes”?
02:35 What are algorithms missing out?
05:06 Are algorithms only useful when we are facing a clear and defined problem?
08:29 How reliant are you on the skills and expertise of the Supply Chain Scientist creating the numerical recipe?
12:34 How do you ensure that the numerical recipes created by the Supply Chain Scientists are all up to a good level?
17:48 What about the supply chain industry? How open are modern companies to using numerical recipes?
20:58 To conclude, why are numerical recipes so important? Why is it so important to change them?


Much like a great chef in a Michelin star kitchen, the best data scientists have to craft statistical solutions that adapt and evolve to every scenario. As such, we investigate what it takes to create these numerical recipes and what characterizes the solutions built for our supply chains.

The naive rationalism is the idea that you have a problem and a solution that goes with it. But the reality is not so simple, as usually the type of solution shapes the problem - and vice-versa. Creating solutions for supply chains is often an entire journey, with unexpected bumps along the road, due to the pure complexity of supply chains.

We use the term “numerical recipe” to specifically explain that we’re talking about more than just algorithms. For example, when using the archetypal “sorting” algorithm, where you have a collection of objects with an order relationship and you can “sort” them, with a well defined series of steps, the algorithm will have many aspects, such as the memory consumption, the number of steps, stochastic, deterministic, etc.

However, using the sorting algorithm usually makes sense for a somewhat clear-cut situation, with a problem statement that is completely non-ambiguous and where there is a mathematical clarity. On the contrary, when it comes to solving real world problems in actual supply chains, things get much more muddy. For example you’ll have MOQs, which aren’t set in stone like the laws of physics, but are usually negotiated in person with suppliers and are subject to change.

This is not to say we don’t use algorithms at Lokad (of course we do). Algorithms are of course very useful and extensive research allows us to know their pros and cons. But algorithms are ultimately just one small screw in a much larger machine.

To conclude, we talk about the importance of the expertise and insights of Supply Chain Scientists and go into more detail about how practically all companies actually operate with numerical recipes but don’t realise it.