An algorithm that optimises pharmacies purchases

The drug market is particularly complex. To help Valpharmex, a company based in Sierre, find the best prices for its customers, Icare developed the AOptima platform. After two years of research, the prototype was industrialised as part of an Ark project.
Medicines pass through a multitude of players who offer very different prices and commercial conditions depending on the quantities sold, the time of year, the business relationships established, etc. In this opaque market, it is therefore very difficult for pharmacies to have an overview and to obtain an optimal price from suppliers. This is why they turn to specialised companies such as Valpharmex, which find them the best commercial conditions for the products they are looking for.
Until now, the information processed by Valpharmex was the result of a manual process due to the wide variety of media used (fax, e-mail, paper, voice messages, etc.). The analysis and management of this information was very complex, which could lead to significant shortfalls due to human error, constantly changing prices and the volatility of many other criteria.
The main scientific challenge of this CTI project was to centralise the process in a single system, capable of managing the heterogeneity and large volume of information while integrating a prediction and purchase optimisation algorithm. This algorithm is able to identify the best commercial condition taking into account more than 150 criteria (time, regionality, cross-selling, etc.).
In order to facilitate the search for items, semantic technologies have been exploited, establishing links between products. Thus, for each pharmaceutical or parapharmaceutical product, the AOptima decision support platform proposes the most advantageous supplier and predicts the ideal purchase period, which leads to significant cost reductions. The next step is to industrialise the solution resulting from the research by refining the semantic links and stabilising the optimisation algorithms in order to offer a robust and easy-to-understand solution.