Search

Optimization problems often require the use of optimization methods that permit the minimization or maximization of certain objective functions. Occasionally, the problems that must be optimized are not linear or polynomial; they cannot be precisely resolved, and they must be approximated. In these cases, it is necessary to apply heuristics, which are able to resolve these kinds of problems. Some algorithms linearize the restrictions and objective functions at a specific point of the space by applying derivatives and partial derivatives for some cases, while in other cases evolutionary algorithms are used to approximate the solution. This work proposes the use of artificial neural networks to approximate the objective function in optimization problems to make it possible to apply other techniques to resolve the problem. The objective function is approximated by a non-linear regression that can be used to resolve an optimization problem. The derivate of the new objective function should be polynomial so that the solution of the optimization problem can be calculated. [ABSTRACT FROM AUTHOR]

E-nose systems are becoming increasingly important instruments across all industries, especially the fields of food and beverages and biomedicine. Given the inaccurate, unsafe and unreliable dependency on the human nose to detect smells that are highly risky and hazardous to human health, e-nose systems offer a tremendous advantage. E-noses are convenient, highly efficient and can be used in real life to detect various types of odors. This paper presents a virtual organization of agents that integrates different classification techniques and neural networks to perform information fusion from parameters retrieved by the E-nose. The integral brain in e-noses is the data processing system, which classifies odors that have been detected by the detection part of its system. The system mimics how a human brain classifies odors. [ABSTRACT FROM AUTHOR]

Abstract: When working in conjunction with multiagent systems, virtual organizations attempt to simulate the functions and interactions of entities in different environments. Recent studies have addressed the problem of assigning roles to agents that form part of the organization, and incorporating new agents to carry out certain tasks. However, these studies are limited to defining the norms and rules that determine the behavior of the organization. The present study proposes a virtual organization model for egovernment environments to assign resources and minimize the required personnel by forecasting workloads. To this end, a neural network, queuing theory, and CBR are used to obtain an efficient distribution. Queuing theory can establish the number of agents with a specific role that are necessary to maximize profits, while the network distributes roles among agents according to their respective efficiency. The final part of the paper is focused on validating the plan developed inside a case study centered on e-government in order to obtain empirical results. [Copyright &y& Elsevier]

Abstract: Microarray technology provides a great amount of data requiring analysis, and the use of new intelligent algorithms that identify the relevant information for the classification process has become essential. This study presents a classification tool called MicroCBR that uses the case-based reasoning paradigm and incorporates a novel filtering technique based on statistical methods, a new clustering technique that uses ESOINN (Enhanced Self-Organizing Incremental Neuronal Network), and a knowledge extraction technique based on the J48 algorithm. MicroCBR has been applied to classify 91 CLL patients and the results obtained are shown in this paper. [Copyright &y& Elsevier]

Clustering is a branch of multivariate analysis that is used to create groups of data. Most of the existing clustering techniques require defining additional information, including the actual number of clusters, before they can be carried out. This article presents a novel neural network that is capable of creating groups by using a combination of hierarchical clustering and self-organizing maps, without requiring the number of existing clusters to be specified beforehand. The self-organized cluster automatic detection neural network is described in detail, focusing on the density, the average distance, the division algorithm, the update algorithm and the training phase. Three case studies have been carried out in this research in order to evaluate the performance of the neural network, and the results obtained are presented within this article. [ABSTRACT FROM PUBLISHER]

This paper presents an intelligent environment developed for monitoring patients’ health care in execution time in hospital environments. The CBPMP (case-based planner for monitoring patients) is an autonomous deliberative case-based planner designed to plan the nurses’ working time dynamically, to maintain the standard working reports about the nurses’ activities, and to guarantee that the patients assigned to the nurses are given the right care. The planner operates in wireless devices and is integrated with complementary software into an intelligent environment, named AmI-P (Ambient Intelligence for patients). CBPMP description, its relationship with the complementary technology, and preliminary results of system prototype in a real environment are presented. [Copyright &y& Elsevier]

Abstract: The following article proposes the implementation of a system based on neural networks of CounterPropagation type and intelligent agents for analysis and assessment of the risk of flood caused by rain, in addition to the implementation of agents in mobile devices for the presentation of alerts. Because as is known, natural phenomena have always existed, but in recent years as a result of global warming we have seen that floods are becoming more frequent, which has forced the creation of specialized agencies and intelligent mechanisms to prevent the loss of human lives due to these phenomena. [Copyright &y& Elsevier]

Abstract: This paper presents a novel Ambient Intelligence based solution for shopping assistance. The core of the proposal is a CBR system developed for guiding and advising users in shopping areas. The CBR incorporates a neural based planner that identifies the most adequate plan for a given user based on user profile and interests. The RTPW neural network is based on the Kohonen one, and incorporates an interesting modification that allows a solution or a plan to be reached much more rapidly. Furthermore, once an initial plan has been reached, it is possible to identify alternatives by taking restrictions into account. The CBR system has been embedded within a deliberative agent and interacts with interface and commercial agents, which facilitate the construction of intelligent environments. This hybrid application, which works on execution time, has been tested and the results of the investigation and its evaluation in a shopping mall are presented within this paper. [Copyright &y& Elsevier]