Section: New Results

Decision Support Systems Applied to Agronomy

Participants : Pierre Bisquert, Patrice Buche, Abdelraouf Hecham, Madalina Croitoru, Jérôme Fortin, Rallou Thomopoulos, Bruno Yun.

High-level decision-making needs to take into account the often-conflicting interests of different stakeholders with the goal of finding solutions to provide trade-offs and build consensus towards the adoption of so-called win-win solutions. In order to enrich the deliberation process we have proposed several complementary approaches that combine various methods for an unified approach towards decision making. This has been applied in practical domains as explained below.

First, in [17] we presented a systematic method to assess possible options, based on the complementarity of argumentation modeling and system dynamics (SD) simulation, in conjunction with field experimentation. Taking advantage of the argument analysis, SD simulations are used to: 1) compare different cultural strategies available to farmers in current operating, market and regulatory conditions; 2) propose plausible what-if scenarios anticipating technological progress, and exploring the impact of adopting potential incentives and dissuasive regulatory measures.

Second, voting theory has been applied at the service of decision making. We employed Computational Social Choice (CSC) and Argumentation Framework (AF) as a combination to propose socially fair decisions which take into account both (1) the involved agents' preferences and (2) the justifications behind these preferences. Furthermore we implemented a software tool for decision-making which is composed of two main systems, i.e., the social choice system and the deliberation system [16]. This work was evaluated in practice [18]. Note that the use of argumentation in practice, when not considering fully formalised domains is very challenging. This specifically concerns decision support systems as shown in [34] where we focused on the following research question: “How to define an attack relation for argumentative decision making in socio-economic systems?” To address this question we proposed three kinds of attacks that could be defined in the context of a specific application (packaging selection) and studied how the non-computer-science experts evaluated, against a given set of decision tasks, each of these attacks.