No video

Inventory Optimization Models and Simulations (with Nicolas Vandeput) - Ep 110

Historically, programming has often been seen as an intimidating task reserved for computer whizzes and IT practitioners. Nowadays, however, with the rise of tools such as Python, programming has become that much more accessible. For this episode of LokadTV, we're joined, once more, by Nicolas Vandeput to discuss just how simple programming can actually be.
Since the last time Nicolas was on the show, he has published a new book "Inventory Optimization Models and Simulations". This publication puts forward what we believe to be a cornerstone of modern supply chain optimization, which is the need for programmatic expressiveness.
This programmatic expressiveness is the only way to be able to successfully deal with real world supply chains and the real world situations that accompany them. Instead of focusing on idealized supply chains, with mathematical frameworks that fall apart when confronted with the real word, Nicolas shows how simple and straightforward numerical recipes with Python can be used to address such real world situations.
Although programming has been around for quite a while, for a long time, companies were struggling to simply achieve an accurate digital counterpart of their supply chain, i.e. an ERP or WMS set-up to manage stock levels. This took some time, with the first ERPs being deployed at the end of the 70's, so we have four decades under our belts nowadays and are able to start thinking differently.
Most modern supply chains are run by Excel, and Python could be seen as a natural continuation of this, permitting more programmatic expressiveness than what Excel allows. There are many other alternative programming languages, like JavaScript or Java, but Python has multiple advantages, such as an extensive library of resources, which means that it's very difficult to get stuck. But maybe its biggest benefit is its simplicity.
However, although Python may be the most adapted programming language for supply chain, it definitely has its limitations.
To conclude, we talk more about common confusions within supply chains, such as the safety stock formula and why it's so often incorrectly used, and what supply chain can do to attract more brilliant minds.
You can find Nicolas' new book via: www.amazon.com/Inventory-Opti...
******
0:00:00 - Introduction
0:00:28 - Nicolas, what have you been up to since we last saw you on the show?
0:00:57 - Joannes, today we are going to discuss Nicolas’ book "Inventory Optimization: Models and Simulations"? What is so different about this book?
0:03:46 - Who is this book intended for?
0:05:13 - Programming has historically been reserved for IT departments. What is changing now?
0:08:25 - Why is Python the language of choice for this book? What does it provide that other programming languages do not?
0:11:42 - Can we do better than Python?
0:16:18 - Nicolas, your book mentions a few sources of confusions in inventory optimization. What are the common confusions that we should be aware of in the industry?
0:18:33 - Would you say this confusion is why people are so happy to pay so much for enterprise software?
0:22:31 - Nicolas, your book clearly explains many inventory optimization models and their limitations. Would you say that some of these limitations can be overcome? How?
******
Check out our website: www.lokad.com/
Follow us on LinkedIn: / lokad
Follow us on Twitter: / lokad
Read our blog: blog.lokad.com/
Sign-up for the Lokad newsletter: www.lokad.com/home#news
******

Пікірлер: 7

  • @fawadmahdi
    @fawadmahdi3 жыл бұрын

    Thanks guys. Very useful videos for someone who wants to get a practical perspective on optimising supply chains

  • @mohamedmehnaoui5588
    @mohamedmehnaoui55882 жыл бұрын

    Insightful content! Keep going guys

  • @davidcmpeterson
    @davidcmpeterson2 жыл бұрын

    @12:05 got to mention R & Julia! Heck, even MATLAB & Fortran justify more of a mention than the likes of Java when discussing supply chain modeling.

  • @Lokad

    @Lokad

    2 жыл бұрын

    Indeed R/Matlab/Fortran/Julia (and Python) are the popular data science languages. However, , from an enterprise software perspective, the math part of the job is a very small fraction of the challenge. If all the relevant business systems are already in Java, doing the supply chain modeling in Java is not nearly as odd as it looks. I have seen many more data science initiative fails due to "plumbing" issues compared to "modeling" issues. Then, even considering Julia (instead of Java), the point discussed in the episode stands: Julia is exposing to the engineer tons of concerns that have no net positive contribution to the supply chain challenge. Ex: multiple dispatch in Julia, very nice, but it takes a software engineer to master the sort of class of language capabilities. Cheers, Joannes

  • @samuelamihere2891
    @samuelamihere28913 жыл бұрын

    Pls I need a copy of the book

  • @Lokad

    @Lokad

    3 жыл бұрын

    You can buy the book at www.amazon.com/Inventory-Optimization-Simulations-Nicolas-Vandeput/dp/3110673916/ Hope it helps. Best regards, Joannes

  • @olivierjonard1872
    @olivierjonard18723 жыл бұрын

    You actually need clever people in Supply Chain, in order to : * implement APS that don't work, or only for a small part of your supply chains * pretend to the CFO that everything is fine, otherwise you'll never get funding for any other IS any more in a near future * rebuild everything is Excel, Access (now Power Query) or in BI so that you can actually make decent decisions. I actually reduced raw Materials inventory coverage from 8 months to 2,5 months only with Excel and re-order point on SAP. Starting with the "limited" Safety Stock Formula mentioned by Nicolas, that I had learned 20 years before * don't screw up too much * (motivate the teams, because you depend on people to compensate for systems screws up) On top you need to educate top management on the importance of Supply Chain and of the stupidity of some metrics * 99% line utilization is NOT a good way to reduce total costs * 100% service levels is NOT achievable, unless you want to go bankrupt * 85% forecasts accuracy does not mean anything... etc... => curious to discover Nicolas' book.