Skip to Main content Skip to Navigation
Theses

Modélisation et pilotage de la phase de délibération dans une décision collective : vers le management d'activités à risques

Abstract : The management of risk activities involves many decisions that concern a group of actors with different areas of expertise or action to design or operate a complex system. First, we focus on the role of an interactive group decision support system (GDSS) in the management of risk activities. The functionalities of the system are specified so as to minimize the impact of human and organizational errors that may affect collective decision-making. The GDSS is seen as the mediator between human beings and the system they aim to control: it supports them detecting critical situations, understanding situations, interpreting and diagnosing before remedy, but it also promotes collective resolution as a medium for communication and coordination of stakeholders. The decision is seen as a dynamic process whose response time depends on the efficiency with which the deliberation phase is managed. Several variables and models to control the debate are thus proposed. A first type of decisional situation depicts a manager who requires the support of a panel of experts to make a decision based on a process of fusion of expert advice. The uncertainty attached to the evaluation of alternatives is due in part to the imprecision of expert opinion, divergences of opinion between experts. A control model is proposed to manage this uncertainty and identify the criteria on which to focus the panel discussion. We then turn to the concept of influence in a social network to offer two models of control of the deliberation phase. One is based upon argued deliberation and the other one on stochastic simulations with a formalism of state equations to represent the evolution of beliefs during the debate. Next, decisions that concern improvement of a complex system are presented. Strategic and operational points of view are confronted in such a process. When achievable goals are negotiated, a model based on a problem of constraint programming techniques is used to compute the most efficient actions to be carried out to perform the improvement. In this case, either the collective is seen as a set of collaborative agents and deliberation is driven by an efficient decision; or the agents are simply cooperative, with collective goals but individual issues and thus the negotiation model is based on argumentation theory. Our models all deal with the control of a cognitive process, i.e., collective decision: cognitive automation aims to reduce human and organizational errors that could affect the decision in particular assessment and coordination errors. Conclusion and perspectives complete this manuscript which is illustrated with several examples related to the management of risk activities.
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/pastel-00658558
Contributor : Bibliothèque Mines Paristech <>
Submitted on : Tuesday, January 10, 2012 - 4:05:04 PM
Last modification on : Sunday, October 25, 2020 - 7:05:32 AM
Long-term archiving on: : Wednesday, April 11, 2012 - 2:31:42 AM

Identifiers

  • HAL Id : pastel-00658558, version 1

Citation

Abdelhak Imoussaten. Modélisation et pilotage de la phase de délibération dans une décision collective : vers le management d'activités à risques. Gestion et management. École Nationale Supérieure des Mines de Paris, 2011. Français. ⟨NNT : 2011ENMP0048⟩. ⟨pastel-00658558⟩

Share

Metrics

Record views

1004

Files downloads

1364