MI4 presents its new project based in the use of AI

Submitted by Silvia Jiménez… on 14 December 2020

MI4 has the pleasure to present one of the newest projects of the Cluster, titled "Production planning using automatic learning and optimization algorithms".

The main objective of this project is the development of an advanced data software prototype, based on automatic learning algorithms, which allows to automate and optimize the process of planning of non-serial production in industrial companies.

The process is carried out from the collection of data from the systems of the industrial company, to the creation of a query API that allows predictions of time and related costs, through training and selection of a predictive algorithm with the highest possible efficiency.

This is a collaborative project, funded by the Program to support Innovative Business Groups (AEI’s in Spanish) of the Ministry of Industry, Trade and Tourism in which three industrial companies participate Mecanizaciones y Fabricaciones, S.A. (Mefasa), Ingeniería y Diseño Europeo, S.A (Idesa), which will be joined by thyssenkrupp Norte, with which the key data will be identified and which, together with the participating technology developers Izertis and Talento Transformación Digital, will work on the project.

The project defines among the main milestones to be achieved:

  1. Study and interaction with industrial companies to check the presence of data for a series of processes to be modelled. Based on this information, a series of viable processes can be defined on which to work, since sufficient quality information will be available to forecast the viability of generating a predictive algorithm.
  2. Collection of all available information from the defined processes to classify and process it automatically (elimination of erroneous data, treatment of empty data, etc.).
  3. Performing a training of different algorithms and different configuration parameters automatically from the data collected for each process.
  4. Automatic selection of the most appropriate trained algorithm for a particular process, through a validation of the algorithms against a set of test data chosen from the original data obtained from the companies and not used in the training process.
  5. Definition of an automated system to simplify the algorithms obtained.
  6. Implementation of a generic API that can include such an algorithm to make the predictions and in turn receive calls with data in a defined structure that is common to all processes.

The project “Production planning using optimization and automatic learning algorithms” is a project funded by the Ministry of Industry, Trade and Tourism, through the Support Program for Innovative Business Groups 2020. This program is aimed exclusively at those national Clusters and AEI’s, and the companies that comprise them, which have been recognized by the Ministry with the Bond Spain Cluster seal included in the Registry of Innovative Business Groupings of the Ministry itself.

Share this Article