From data driven decision making (DDDM) to automated data driven model based decision making (MBDM)

Ioan Doré Landau (1) , Vlad Landau (2)

1 GIPSA-SLR - GIPSA - Systèmes linéaires et robustesse
2 Freelancing Consultant

Data driven decision making (DDDM) constitutes the modern operation of systems and organizations. However a prediction model of the system can be estimated from data. The availability of data driven prediction models allows on one hand to test in simulation various decision policies and on the other hand to move toward an automated decision making procedure. These automated decision making procedures have a feedback structure and the design of the decision algorithms relies on the knowledge of the prediction model of the system. This procedure can be called automated data driven model based decision making in short (MBDM). The paper explores basic aspects to be considered in the design of the decision algorithms related to optimal and safe operation of the full system.

Mots clés: Key-Words :Data mining; Prediction model; Parameters estimation; Feedback control; Stability; Predictive control

Domaines: Automatique; Informatique [cs]; Sciences cognitives

Ioan Doré Landau, Vlad Landau. From data driven decision making (DDDM) to automated data driven model based decision making (MBDM). 2017. ⟨hal-01527766⟩