New group decision making models with heterogeneous information based on different frameworkschangeable contexts, non-homogeneous experts and web 2. 0

Resumen

Decision Making, that is, selecting the best alternative (or alternatives) from a feasible set, is a very common task present in almost every human activity. Thus, it provokes a great interest in the study of decision making situations and mechanisms that allow to solve decision making problems, not only in Decision Theory, but also in other disciplines as Artificial Intelligence, Economy, Sociology, Engineering and so on. However, basic decision models have little in common with real decision making. Many real decision making processes are developed in environments where objectives, restrictions and feasible options are not exactly known and defined. Thus, it is necessary to study and refine those decision models in order to be able to represent this uncertainty. A practical and powerful way to handle uncertainty in human knowledge was proposed by professor Zadeh in 1965: Fuzzy Sets Theory [Zad65]. The application of Fuzzy Sets Theory to solve information uncertainty in decision making was proposed by Bellman and Zadeh in 1970 [BZ70] and since that moment it has been widely used because of its utility. Fuzzy Sets Theory has provided a much more flexible framework where it is possible to easily represent and tackle imprecision of human judgements. Usually, decision problems require to make some analysis of the different alternatives and the problem that we face. However, not every decision problem is solved by means of a completely rational process. In fact, many external and subjective factors affect the decision processes, and thus, the final solution for a decision problem can change if the conditions in which the problem is presented vary. It is obvious that the comparison of different actions according to their desirability in decision problems, in many cases, cannot be done by using a single criterion or a unique person. Thus, we interpret the decision process in the framework of group decision making (GDM) [KF90, Rou97]. This approach has led to numerous evaluation schemes and has become a major concern of research in decision making. A GDM problem appears when there is a question to be solved, a set of alternatives from where to choose, and a set of persons (experts) ,which express their opinions or preferences about the available options. Experts should have the intention of reaching a collective decision about the problem. Sometimes there exists a particular person, called moderator, which is in charge of the direction of the whole resolution process until the experts reach an agreement about the solution to choose. To correctly solve a GDM problem two main different processes have to be carried out [HHVV96b]: The consensus process and the selection process. Both have been widely studied by different authors and in different GDM contexts [FR94, KF88, KF90]. The first one refers to how to obtain the highest consensus or agreement among experts about the set of alternatives. The second one (which is also called the algebraic consensus process) refers to how to obtain the final solution set of alternatives from the opinions expressed by experts. Both processes work together sequentially. First of all, the consensus process is developed to reach the maximum consensus degree among experts' preferences. In every step of the process the current consensus degree is measured, and if it does not reach an acceptable level, experts are encouraged to discuss their points of view and change their opinions to increase the proximity of their preferences. Once a certain level of consensus have been reached the selection process is applied and the final solution is obtained. Thus, a GDM process can be defined as a dynamic and iterative process in which experts change their opinions until their preferences about the solution are close enough, therefore allowing the obtention of a solution of consensus by means of the application of the selection process. We will pay attention to different kinds of GDM situations along this memory, analyzing the current GDM models and trying to improve them. In order to carry out this study, this memory is divided in two parts. First one is devoted to the problem statement and the discussion of the results. Second one corresponds to the publications associated to this study. In Part I we begin by developing the problem statement introduced in this section and the techniques proposed to solve it with the following subsections: subsection 1.1 introduces the consensus approaches in fuzzy Group Decision Making situations, subsection 1.2 describes GDM problems with heterogeneous information in changeable contexts, subsection 1.3 presents GDM problems with and different experts' importance and subsection 1.4 illustrate GDM situations in web 2.0 communities. Next we indicate the open problems which justify the realization of this memory in section 2. The objectives pursued in this memory are described in Section 3. Section 4 provides a summarized information about the proposals and most interesting results obtained in each part. Section 5 summarizes the results obtained in this memory and present several conclusions about them, moreover, in section 6 we point out several open future works which remain open from the results of the present memory. Finally, in order to develop the goals set, this memory is constituted by eight publications distributed in four sections which will be developed in Part II. They are the following: 1. Analyzing Consensus Approaches in Fuzzy Group Decision Making: Advantages and Drawbacks. 2. Mobile Decision Support Systems Based on Heterogeneous Information and Changeable Contexts. 3. A New Consensus Model for Group Decision Making Problems with Non Homogeneous Experts. 4. A Linguistic Consensus Model for Web 2.0 Communities.