Institut des Systèmes Complexes - Paris Île-de-France - 57-59 rue Lhomond, 75005 Paris
Why a few liars are essential to win a controversial public debate?
11 - 12:30 pm
Abstract: Controversial public debates driven by incomplete data where nobody can claim absolute certainty, due to the current state of knowledge, are studied within the frame of sociophysics. To adopt a cautious balanced attitude based on clear but inconclusive data appears to be a lose-out strategy. In contrast overstating arguments with incorrect claims which cannot be refuted appears to be a key ingredient to eventually win a public debate. The underlying key mechanisms of these puzzling and unfortunate conclusions are identified using the Galam Unifying Frame (GUF) of opinion dynamics. It reveals that the existence of inflexible agents and their respective proportions are the instrumental parameters to determine the outcome of incomplete data driven public debates. Acting on one's own inflexible proportion modifies the topology of the flow diagram making irrelevant the value of initial support. Thus focusing on open-minded agents may turn useless. Accordingly, the inflexibles drive the opinion of the population. The results shed a new but disturbing light on designing adequate strategies to win a public debate.
Speaker: Prof Serge Galam,
Centre National de la Recherche Scientifique (CNRS),
Upload the abstract
Multicriteria Decision Making Seminar
16 - 17:30 am
Abstract: Decision makers might face the problem of evaluating alternatives under several criteria. Multicriteria decision making was developed to fulfill this task. However, aggregating criteria can be a complicated task. The reason being that attributes are not always independent, they might interact. Hence, a simple weighted mean is not sufficient in general. There is a wide range of effects that can take place. Compensatory effects, when the lack of an attribute can be compensated by another attribute; redundancy, when we can exchange two attributes; synergy, when two attributes that are not very important separately acquire a great importance when found together. Fuzzy integrals are aggregation operators that consider the interactions above, by defining a measure over the power set of the attributes, hence giving a weight to every possible combination of attributes.
Fuzzy integrals can also express different behaviours for a decision maker given the same weights and utility functions, behaviours ranging from tolerance to intolerance. Fuzzy integrals have been applied to different problems, where significant improvement has been achieved compared to linear models.
Speaker: Michel Grabisch, Ph.D,
Centre de recherche de mathématiques
et économie mathématique (CERMSEM)
Centre d'Économie de la Sorbonne
Université Paris 1 (Panthéon-Sorbonne)
Upload the abstract and the biography