A metamodel calculation continuous time for aid decision systems applied to financial planning ; Un métamodèle de calcul en temps continu pour les systèmes d'aide à la décision appliqués à la planification financière

Open Access

Abstract

In the scope of Business Intelligence, planning aims to support multiple actors in their process of converging different views and problematics from different domains to get a shared business planning model. A major difficulty in business planning is that each actor states her/his views and problematics with a different time scale. Integrating them into a unique model that represents a common state of reality becomes very costly and awkward to manage when basing the construction of these models on discrete modeling techniques used by current tools of business planning. This doctorate thesis proposes a novel solution, beyond the state-of-the-art, for addressing these issues: it conceives a novel metamodel based on a continuous time calculus. Through the developed approach, it allows multiple actors to integrate the different business logics of their planning domain in a shared model as well as to observe it from different time scales. The advantages of our solution based on continuous time against solutions based on discrete time are presented through a case study. The conceived metamodel was implemented within a real industrial set in MGDIS (a company specialized in business planning for local governments) following an innovative service oriented architecture: this architecture segregates the modeling from the evaluation to allow the parallelization of model evaluation for big volumes of data. The overall solution conceived and implemented in this thesis was designed to be a real scale prototype to be applied to real scale problems. Besides the case study, it was validated by MGDIS experts on business planning against real requirements. ; Dans le cadre de l'informatique décisionnelle, la programmation physico-financière doit permettre à des acteurs d'une collectivité provenant de divers domaines de faire converger leurs problématiques vers un objectif commun. L'une des principales difficultés de la modélisation d'une programmation physico-financière est que chaque acteur exprime ses problématiques dans des ...