Catálogo de publicaciones - libros

The calculus of relations has been very important during the past 40 years from theoretical as well as from practical point of view. The development of fuzzy set theory, particularly in the framework of relational calculus has undoubtly increased the interest in this domain of science.

The Golden ratio is one of the most beloved numbers in human society. It is a symobol of combination of beauty and practical use. We consider the Golden ratio through four models —(i) optimization, (ii) inequality, (iii) identity and (iv) opertator —. We introduce the Golden matrices whose characteristic values are the Golden number and its conjugate. We show that the Golden matrices take an important role in the four models. Further the role is essentially equivalent.

Pattern discovery from string data is an important problem with many applications. In this paper, we give a brief overview of our work on the , which integrates numerical attribute information into the string pattern discovery process.

We take institutions seriously as both a rational response to dilemmas in which agents found themselves and a frame to which later rational agents adapted their behaviour in turn. Medieval corporate bodies knew that they needed choice procedures. Although the social choice advances of ancient Greece and Rome were not rediscovered until the high middle ages, the rational design of choice institutions predated their rediscovery and took some new paths. Both Ramon Llull (ca 1232-1316) and Nicolaus of Cusa (a.k.a Cusanus; 1401-64) made contributions which had been believed to be centuries more recent. Llull promotes the method of pairwise comparison, and proposes the Copeland rule to select a winner. Cusanus proposes the Borda rule, which should properly be renamed the Cusanus rule.

Voting might be needed in any institution ruled by more than one person, where decisions could not simply be handed down from above. Medieval theologians no doubt believed that God’s word was handed down from above; but they well knew that they often had to decide among rival human interpretations of it. The Church faced its own decision problem every time a new Pope needed to be elected. Bodies not directly in the hierarchy of the Church had to evolve their own decision procedures. The chief such bodies were commercial and urban corporations; religious orders; and universities.

The disagreement between Llull and Cusanus raises the issue: should voting be regarded as a method of aggregating judgments or as a method of aggregating interests? In the former interpretation (only), voting procedures are a solution to a problem of approximate reasoning. There is an unknown, true state of affairs (for medieval thinkers, divine will). A voting procedure aggregates unreliable individual perceptions of the will of God to a more reliable group judgment of it. In the rougher world of Cusanus, and probably of electors to the papacy and to Dogeships, only at most lip service is paid to the will of God, and voting is a process of aggregating interests.

In Group-Choice Decision Making (GCDM) where a number of stakeholders are involved in choosing a single solution from a set of available solution options, it is common for the stakeholders to form coalition during negotiation in order to increase their individual welfare. It is also common to use Multi-Agent Systems (MAS) to automate GCDM processes. In such MAS, agents have to form coalitions like their human counterparts, and within each coalition, the individual agents behave according to the strategies of their clients. This paper presents a coalition formation engine that has two coalition formation algorithms. One of the algorithms is based on the concept of static coalition formation, and the other is based on the concept of dynamic coalition formation. Moreover, the coalition formation engine is coupled with algorithms that govern the social behavior of the agents in their coalitions, to form an agent negotiation engine. Finally, this paper presents an example and simulation results that illustrate the operational effectiveness of the two coalition formation algorithms, as well as the algorithms that govern the social behavior of the agents.

We present a multicriteria fuzzy system using gradual rules and fuzzy arithmetic. We first present a multicriteria problem and its solution for the case of precise information. Then we extend the model to treat pieces of information that may involve imprecision/vagueness. We show that the use of residuated implication operators, employed by gradual rules, coupled with similarity relations offer a better treatment of the problem than a Mamdani-like approach.

In this paper, we discuss a behavioral model of decision making using weather information, making use of the so-called fuzzy target based decision model. Due to forecasting uncertainty in weather forecasts, many decision problems in practice influenced by weather information have been formulated as that of decision making under uncertainty. After introducing the fuzzy target based decision model which states that, after assessing a fuzzy target, the decision maker should select the decision which maximizes the probability of meeting his target, we will show that different behaviors of the decision maker about his target can lead to different decisions. This behavioral analysis not only provides an interpretation for influence of psychological personality features of the decision maker on his decisions, but also has a corresponding link to attitudes towards risk in terms of utility function.

In group decision making (GDM) processes, prior to the selection of the best alternative(s), it would be desirable that experts achieve a high degree of consensus or agreement between them. Due to the complexity of most decision making problems, individuals’ preferences may not satisfy formal properties. Consistency is one of such properties, and it is associated with the Obviously, when carrying out a rational decision making, consistent information, i.e. information which does not imply any kind of contradiction, is more appropriate than information containing some contradictions. Therefore, in a GDM process, consistency should also be sought after.

In this paper we present a consensus model for GDM problems that proceeds from consistency to consensus. This model includes a novel consistency reaching module based on consistency measures. In particular, the model generates advice on how experts should change their preferences in order to reach a solution with high consistency and consensus degrees.

In this paper we introduce a multi-stage decision making procedure where decision makers sort the alternatives by means of a fixed set of linguistic categories, each one has associated a numerical score. First we average the scores obtained by each alternative and we consider the associated collective preference. Then, we obtain a distance between each individual preference and the collective one through the Euclidean distance among the individual and collective scoring vectors. Taking into account these distances, we measure the agreement in each subset of decision makers, and a weight is assigned to each decision maker: his/her overall contribution to the agreement. Those decision makers whose overall contribution to the agreement are not positive are expelled and we re-initiate the decision procedure with only the opinions of the decision makers which positively contribute to the agreement. The sequential process is repeated until it determines a final subset of decision makers where all of them positively contribute to the agreement. Then, we apply a weighted procedure where the scores each decision maker indirectly assigns to the alternatives are multiplied by the weight of the corresponding decision maker, and we obtain the final ranking of the alternatives.

By selecting and asking the user to label only the most informative instances, active learners can significantly reduce the number of labeled training instances to learn a classification function. We focus here on how to select the most informative instances for labeling. In this paper we make three contributions. First, in contrast to the leading sampling strategy of halving the volume of version space, we present the sampling strategy of reducing the volume of version space by more than half with the assumption of target function being chosen from nonuniform distribution over version space. Second, via Halving model, we propose the idea of sampling the instances that would be most possibly misclassified. Third, we present a sampling method named CBMPMS (Committee Based Most Possible Misclassification Sampling) which samples the instances that have the largest probability to be misclassified by the current classifier. Comparing the proposed CBMPMS method with the existing active learning methods, when the classifiers achieve the same accuracy, the former method will sample fewer times than the latter ones. The experiments show that the proposed method outperforms the traditional sampling methods on most selected datasets.