Interprétation automatique de données hétérogènes pour la modélisation de situations collaboratives : application à la gestion de crise

Abstract : The present work is applied to the field of French crisis management, and specifically to the crisis response phase which follows a major event, like a flood or an industrial accident. In the aftermath of the event, crisis cells are activated to prevent and deal with the consequences of the crisis. They face, in a hurry, many difficulties. The stakeholders are numerous, autonomous and heterogeneous, the coexistence of contingency plans favours contradictions and the interconnections of networks promotes cascading effects. These observations arise as the volume of data available continues to grow. They come, for example, from sensors, social media or volunteers on the crisis theatre. It is an occasion to design an information system able to collect the available data to interpret them and obtain information suited to the crisis cells. To succeed, it will have to manage the 4Vs of Big Data: the Volume, the Variety and Veracity of data and information, while following the dynamic (velocity) of the current crisis. Our literature review on the different parts of this architecture enables us to define such an information system able to (i) receive different types of events emitted from data sources both known and unknown, (ii) to use interpretation rules directly deduced from official business rules and (iii) to structure the information that will be used by the stake-holders. Its architecture is event-driven and coexists with the service oriented architecture of the software developed by the CGI laboratory. The implemented system has been tested on the scenario of a 1/100 per year flood elaborated by two French forecasting centres. The model describing the current crisis situation, deduced by the proposed information system, can be used to (i) deduce a crisis response process, (ii) to detect unexpected situations, and (iii) to update a COP suited to the decision-makers.
Document type :
Theses
Complete list of metadatas

Cited literature [277 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02173328
Contributor : Abes Star <>
Submitted on : Thursday, July 4, 2019 - 1:56:26 PM
Last modification on : Monday, July 8, 2019 - 1:41:04 PM

File

FertierAudreyDiff.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02173328, version 1

Collections

Citation

Audrey Fertier. Interprétation automatique de données hétérogènes pour la modélisation de situations collaboratives : application à la gestion de crise. Autre [cs.OH]. Ecole des Mines d'Albi-Carmaux, 2018. Français. ⟨NNT : 2018EMAC0009⟩. ⟨tel-02173328⟩

Share

Metrics

Record views

73

Files downloads

53