OpenStudio and UCA identify territorial industrial synergies through AI

In a few weeks last year, the brutality of the Covid-19 pandemic highlighted the importance of being able to quickly find other sources of supply or of being able to adapt production tools to manufacture a product that can no longer be imported. Allowing manufacturers to quickly identify suppliers to relocate their production or explore avenues of diversification, this is precisely the vocation of the Atlas of productive synergies. This decision support tool, based on artificial intelligence and open data, was designed in collaboration between public and private actors, “in the service of productive resilience”, explain its designers. “We are looking to build relationships between companies that could work with each other,” explained Arnault Pachot, CEO of OpenStudio. For two years, this digital services company based in Clermont-Ferrand, Toulouse, Lyon and Paris, has been working on this research project with the University of Clermont Auvergne (UCA).

The Atlas of Productive Synergies maps the establishments present in a territory and makes it possible to visualize their productive potential. It is based on the detailed analysis of open data (such as the Insee, Sirene and Rome databases, those of Pôle Emploi or customs), as well as on the modeling of the productive fabric from the work of Harvard University on economic complexity. OpenStudio has already delivered various scientific publications on its research. His Atlas can already be explored online. However, it is more likely to be integrated into custom-made applications developed by OpenStudio for manufacturers who will go there to find potential products and suppliers related to their activities or capable, if necessary, of adapting their production tools to manufacture new products.

A multi-objective recommender system

The technologies implemented in the Atlas also make it possible to design a multi-objective recommendation system, for example for companies seeking to diversify their activities. “For an industrialist who could potentially manufacture 4 or 5 products, what is interesting is to have a decision support system that guides his choice according to certain dimensions associated with each product: complexity, competitive advantage, the level of resilience of the distribution circuit…”, explains Arnault Pachot. “The business manager may want to focus on economic performance, securing supplies, protecting the environment… We are building theoretical models in relation to this, it is one of our lines of work, which is constantly being improved”. On December 18, Arnault Pachot will notably present his work as part of the multidisciplinary conference EIRD (economics, engineering, sustainable development, social sciences) which addresses, among other things, the theme of the circular economy, in Tbilisi, Georgia.

Among the AI ​​technologies deployed in the Atlas of Productive Synergies, there are many semantic techniques, vectorization, with in particular Word2vec, for example to search for proximities in nomenclature designations, between products, etc. , quoted us as an example Arnault Pachot. Added to this are recommendation systems with multi-objective optimization, the use of statistics, cartography, geolocation and significant performance work to preload all data.

In a fairly similar field, OpenStudio has already developed, with the support of the Auvergne-Rhône-Alpes Region, the Phar’Eco marketplace for companies wishing to develop qualified local collaborations.