Catálogo de publicaciones - libros

Cellular manufacturing is one of the most powerful management innovations in actualizing mass customization. In this paper, we develop a data mining algorithm for designing the conventional cellular manufacturing systems.

In this work, we propose a compromised large-scale neighborhood, which is embedded in simulated annealing to solve a cargo loading planning problem arising in logistics industry. It is “compromised” because it makes a tradeoff between the extensive backward checking work incurred in traditional subset-disjoint restriction and the possible infeasibility resulting from the relaxing the restriction. Extensive experiments have shown the competitive advantages of the heuristic approach. The proposed neighborhood search method is generally applicable.

This paper is the first in a series which address the development of a temporal planner using timed Petri net (TPN) models of concurrent temporal plans. Unlike previous work which uses Petri net unfoldings to develop partially ordered plans (in the untimed case), the approach presented here is purely algebraic, and is based on a (min,max,+) discrete event dynamical system description of a TPN. This paper focuses primarily on , with subsequent work detailing how to use this approach for efficient optimal temporal planning.

This paper deals with the validation of a Reinforcement Learning (RL) policy for dosage optimization of Erythropoietin (EPO). This policy was obtained using data from patients in a haemodialysis program during the year 2005. The goal of this policy was to maintain patients’ Haemoglobin (Hb) level between 11.5 g/dl and 12.5 g/dl. An individual management was needed, as each patient usually presents a different response to the treatment. RL provides an attractive and satisfactory solution, showing that a policy based on RL would be much more successful in achieving the goal of maintaining patients within the desired target of Hb than the policy followed by the hospital so far. In this work, this policy is validated using a cohort of patients treated during 2006. Results show the robustness of the policy that is also successful with this new data set.

This paper presents a new approach to image segmentation of colour images for automatic pipe inspection. Pixel-based segmentation of colour images is carried out by a support vector machine (SVM) labelling pixels on the basis of local features. Segmentation can be effected by this pixel labelling together with connected component labelling. The method has been tested using RGB, HSB, Gabor, local window and HS feature sets and is seen to work best with the HSB feature set.

This paper discusses a model to decide when the predicates of various types of sentences should use Spanish subjunctive mood in Japanese to Spanish machine translation. It consists of a hierarchy for the elements of the sentences and a series of rules applied to them at transfer phase. We compare our results against some commercial machine translation systems. Our experiments show that our model outputs more accurate results, a necessary step to get a fairly good and natural translation.

In this paper, we propose to introduce a nonstandard analysis into a logical modeling of continuous dynamics and present a new framework called hyper-finite hybrid automaton (HHA). HHA is a nonstandard interpretation of hybrid automata in the domain of ℝ. We also enlarge the linear temporal logic LTL to LTL to describe the system specification. By using this framework, we examine the validation of the system consistency of the hybrid system, especially the existence and reachability of Zeno point.

Dimensional reduction may be effective in order to compress data without loss of essential information. It is proposed to reduce dimension (number of the used web-areas or ) as a result of the unsupervised learning process maximizing a specially defined average log-likelihood divergence. Two different web-areas will be merged in the case if these areas appear together frequently during the same sessions. Essentially, roles of the web-areas are not symmetrical in the merging process. The web-area or with bigger weight will act as an attractor and will stimulate merging. In difference, the smaller cluster will try to keep independence. In both cases the powers of attraction or resistance will depend on the weights of the corresponding clusters. The above strategy will prevent creation of one super-big cluster, and will help to reduce the number of non-significant clusters. The proposed method is illustrated using two synthetic examples. The first example is based on an ideal matrix, which characterizes weights of the and relations between them. The matrix for the second example is generated using a specially designed web-traffic simulator.

Intelligent service robots should understand the needs of human beings, recognize their environment, and perform reliable tasks. To enable robots to achieve these activities, we need to understand the behavior patterns of human beings first. In recent research, there were some attempts to recognize human intentions through multimodal information systems. In this paper, human behaviors in indoor daily life were analyzed to obtain a pattern. We also suggested the use of the semantic segmentation method for the prediction of succeeding actions and for a better understanding of the robots’ intentions. Finally, the validity of the suggested method was evaluated through real behavior data.

This paper describes a genetic programming (GP) approach to binary classification with class imbalance problems. This approach is examined on two benchmark and two synthetic data sets. The results show that when using the overall classification accuracy as the fitness function, the GP system is strongly biased toward the majority class. Two new fitness functions are developed to deal with the class imbalance problem. The experimental results show that both of them substantially improve the performance for the minority class, and the performance for the majority and minority classes is much more balanced.