Computed data from the batch prediction job would be available on the cdemandQty column. This is then passed as a value to the key “ResultData” for a specific product which is under the Group ID of your planning object.
The complete source code can be found in this repository for your reference.
I find customers could benefit from Google Vertex AI when you can bring your own algorithms in most popular AI frameworks on a Compute Engine instance inside the Vertex AI workbench. When you have such a deep learning virtual machine image, with Google Vertex AI you don’t worry about the infrastructure components. Combined with carefully prepared data from SAP IBP, one can seamlessly reap the benefits of both the platforms.
SAP IBP offers a variety of algorithms that are used very successfully by a large number of productive customers. Therefore, this activity was not about investigating which forecast algorithms are better or worse, but to illustrate additional options available to SAP IBP customers by leveraging Google Vertex AI platform. From my experience working on the SAP IBP and the Google Vertex AI platform, I see both the platforms bring tremendous flexibility for customers to fine tune Supply Chain Planing in a combined tool set . One side you have the possibility to represent business data based on any kind of process in SAP IBP, on the other side, you now have the possibility to define your own constrains or capabilities in an algorithm, bring that to a machine learning platform, apply your existing business data and visualize after the computation. By complimenting each other, both the platforms are improving operational visibility and driving innovation in Supply Chain planning.
Special thanks to friends at Google and SAP who supported me during this work. Thanks for your reading, I would be happy to hear from you.
Domnic Savio Benedict
Note: This blog was originally published in Medium