Skip to main content

MEDIDAS DE SIMILITUD DIFUSAS Y METRICAS GENERALIZADAS CON APLICACIONES A PERCEPCION (Y ROBOTICA)

PID2022-139248NB-I00

•
Funding agency name Agencia Estatal de Investigación
Funding agency acronym AEI
Program Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprogram Subprograma Estatal de Generación de Conocimiento
Call Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Call year 2022
State Plan 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Beneficiary centre UNIVERSIDAD DE LAS ISLAS BALEARES

Resultados relacionados

Resultados totales (Incluyendo duplicados): 8
Encontrada(s) 1 página(s)

Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

Aggregation of Fuzzy Relations: The Bounded Partial Metric Case

  • Mir-Fuentes, A.
  • Valero, O.
  • Pajoohesh, Homeira
[eng] In 2002, Pradera and Trillas faced the problem of merging a special kind of fuzzy relations, the so-called bounded metrics. They gave a description of functions that allow to fuse bounded metrics defined on a set into a bounded metric defined on the aforementioned set and preserving the bound. However, in many real-world scenarios, bounded metrics may fall short in capturing appropriately relationships among entities. In response to this limitation, the concept of bounded partial metric, in the sense of Matthews, emerges offering a broader perspective on distance functions. As introduced in 1994 by Matthews, the concept of partial metric extends the notion of metric allowing, among other things, that self-distances can be different from zero. Inspired, on the one hand, by the importance that bounded partial metrics have acquired in computer science and artificial intelligence and, on the other hand, by the observation that this type of fuzzy relations can be derived through aggregation, in this paper, we study the aggregation problem of bounded partial metrics. In this way, we provide the necessary conditions for a function to be able to serve as an aggregator for bounded partial metrics.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

Una aplicación de conjuntos difusos a la asignación óptima de tareas

  • Jaume-Martín, G.
  • Antich, J.
  • Valero, O.
[spa] Uno de los principales retos al trabajar con sistemas multi-robot consiste en asignar, en cada instante de tiempo, la tarea a realizar mas adecuada para cada robot. En muchas situaciones reales, estas tareas tienen ligados unos plazos de tiempo límite, que se tienen que cumplir. Dada su probada efectividad a la hora de lidiar con el problema que nos concierne, en este artículo, nos inspiramos en los métodos de umbral de respuesta (“Response-Threshold Methods”), los cuales son un caso particular de los métodos de enjambre (“Swarm-like Methods”). Siguiendo esta metodología, los robots deciden cual es la tarea que deben realizar basándose en los estímulos procedentes tanto de la tarea que están llevando a cabo en ese momento, como de la tarea a la cual consideran desplazarse. De esta forma, planteamos una alternativa innovadora a las estrategias de asignación de tareas existentes actualmente en la literatura, que tiene la capacidad de manejar tareas con plazos de tiempo límite asociados. Mas concretamente, los estímulos percibidos por los robots son modelados mediante conjuntos difusos, de manera que el robot es capaz de determinar cual es la tarea más apropiada en cada momento a través de la célebre técnica de optimización difusa de Bellman y Zadeh. Con el objetivo de poder evaluar el nuevo enfoque matemático, hemos llevado a cabo un extenso numero de simulaciones. Así, hemos podido comprobar que la alternativa propuesta es capaz de modelar la evolución de un sistema de enjambre cuando existen tareas con plazos límite de ejecución.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

An application of fuzzy sets to optimal task-allocation problem

  • Jaume-Martín, G.
  • Antich, J.
  • Guerrero, J.
  • Valero, O.
[eng] When dealing with multi-robot systems, one of the main challenges is assigning to each robot the best-possible task to perform at any given time, which is known as the optimal task-allocation problem. Furthermore, in numerous real-world scenarios, it is also crucial to consider time deadlines associated with the tasks. In this paper, we draw inspiration from response-threshold methods (a type of swarm-like methods), which have proven to be effective in addressing the aforesaid optimal task-allocation problem. In such methods, each robot processes a collection of stimuli to evaluate the suitability of a given task. We model the aforementioned stimuli using appropriate fuzzy sets in such a way that each robot determines the most suitable task to perform at every moment of time via the so-called Bellman-Zadeh fuzzy optimization technique which involves aggregation functions. In order to assess the efficiency of the new mathematical approach, we have conducted an extensive set of simulation experiments testing different aggregation functions. The obtained results demonstrate that the proposed approach effectively models the evolution of a swarm-like system when considering tasks with deadlines and that the most effective task allocation is achieved when OWA operators are taken into account.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

On the Use of Modular Indistinguishability Operators in RBFNN-like Models

  • Ortiz, A.
  • Valero, O.
  • Miñana, J.J.
[eng] Radial Basis Function Neural Networks (RBFNN) have become popular machine learning models with a simple structure but at the same time strong non-linear function approximation and effective modeling capabilities. In this work, we explore the use of Modular Indistinguishability Operators (MIO) in RBFNN-like structures to replace the RBFs that populate the hidden layer, to give rise to MIO-based Neural Networks (MIO-NN). In this respect, we introduce a new distance function and prove that it is a modular metric, to next use it to derive two MIOs to be evaluated as the key component of MIO-NNs. As an additional contribution, we describe Self-Defining MIO-NN (SD-MIO-NN) as an approach capable of configuring MIO-NNs in a parameterless way. SD-MIO-NN comprises a first step that defines the size of the hidden layer, a second step that determines the parameters of the hidden neurons and a last step that calculates the weights of the hidden-to-output layer connections. The experimental results show the effectiveness of the proposed MIOs for multi-class classification, and by extension of SDMIO-NN, which in turn compares well with other similar solutions.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

Espacios métricos parciales acotados y el problema de agregación

  • Mir-Fuentes, A.
  • Valero, O.
[spa] En 1981, Borsík y Dobos investigaron el problema de la agregación para espacios métricos. Así, caracterizaron aquellas funciones que permiten combinar una colección de métricas, cada una de ellas definidas en un conjunto, para obtener una sola métrica definida en el espacio producto como resultado. Posteriormente, en 1994, Matthews introdujo la noción de espacio métrico parcial con el objetivo de proporcionar una herramienta matemática adecuada para modelar ciertos procesos que surgen de modo natural en ciencias de la computación y en inteligencia artificial. Inspirados por la aplicabilidad mencionada de los espacios métricos parciales y por el hecho de que existen métricas parciales útiles en dicho campo que pueden ser inducidas mediante agregación, en 2015Alghamdi, Shahzad y Valero exploraron el problema planteado por Borsík y Dobos en el marco de los espacios métricos parciales. Sin embargo, muchas de las métricas parciales empleadas en las aplicaciones son acotadas. Motivados por el hecho de que dicho caso no ha sido tratado desde el punto de vista de la agregación en la literatura, en este trabajo ofrecemos una descripción general de cómo agregar, mediante una función, una colección finita de espacios métricos parciales acotados para obtener una métrica parcial acotada definida en el espacio producto como resultado. Así, una caracterización de dichas funciones es proporcionada.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

Addressing Multi-class Classification Tasks by means of RBFNN-like Models using Modular Indistinguishability Operators

  • Ortiz, A.
  • Valero, O.
[eng] As a machine learning model, a Radial Basis Function Neural Network (RBFNN) is an artificial neural network that comprises a single hidden layer whose neurons implement a set of radial-basis functions that are linearly combined at the output layer. The most popular form of RBFNN makes use of the Gaussian RBF combined with the Euclidean distance. In this work, we explore the use of Modular Indistinguishability Operators (MIO) in the hidden layer of RBFNN-like structures in replacement of Gaussian RBFs. Moreover, we introduce a new modular metric m(t, p, q) and next use it to derive a specific MIO IL(t, a, b) from the Luckasiewicz t-norm. The resulting MIO is next used as the activation function of the hidden neurons, whose parameters are determined on the basis of the samples of the underlying dataset. The performance of the new machine learning model is evaluated through a set of publicly available datasets in the context of multi-class classification tasks.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

A Fuzzy Mid-Point Based Decision-Support Method for the Assessment of Investments and Companies

  • Mulet-Forteza, C.
  • Serra-Moll, M. A.
  • Valero, O.
[eng] In decision-making, we face a situation where we have to decide whether it is worth performing an investment or not, based on a multitude of variables. The Net Present Value (NPV) indicates the profit that our investment will produce over a series of time periods within a fixed time horizon. Its computation begins with certain data from which we can obtain a cash-flow for each time period under consideration. This study will focus on the variables involved in the aforementioned computation that can take values within an interval of possibilities whose end-points are their pessimistic forecast (PF) and their optimistic one (OF). Now, we are in front of infinite different possible forecasts and the expert in charge needs to determine which data is more appropriate among them in order to make the final decision. The objective of this study is to introduce an analytical technique based on the use of L-fuzzy sets, aggregation functions and mid-points to select one forecast among the aforesaid infinite range. This methodology incorporates a penalization, fixed following an analytical procedure and thus reducing subjectivity, for the elapsed time and for the discrepancy between PF and OF in order to generate the NPV. Finally, the methodology is tested by applying it a five-year hotel assessment where real data is considered. Here, it is illustrated that the proposed methodology could be appropriate for the assessment of investments and companies.
Repositori Institucional UIB
Publication Capítulo o parte de un libro (BookPart)

A new decision-making method for the assessment of investments and companies based on fuzzy mid-points

  • Mulet-Forteza, C.
  • Serra-Moll, M. A.
  • Valero, O.
[eng] In decision-making, we face a situation where we have to decide whether it is worth performing an investment or not, based on a multitude of variables. The Net Present Value (NPV) indicates the profit that our investment will produce over a series of time periods within a fixed time horizon. Its computation begins with certain data or variables from which we can obtain a cash-flow for each time period under consideration. Some of these variables involved in the aforementioned computation can take values within an interval of possibilities whose end-points are their pessimistic forecast (PF) and their optimistic one (OF). From these values, the expert subjectively generates two distinct cash-flow projections: the Optimistic Cash Flow (OCF) and the Pessimistic Cash Flow (PCF). The Net Present Value (NPV) is then computed using a cash-flow obtained as the arithmetic mean of both projections. However, there are infinite possible intermediate cash-flows that could be considere to obtain the NPV. Given the lack of an analytical reason to select the arithmetic mean cash-flow as the most appropriate one, and considering that such a choice could be derived from an appropriate aggregation function, in this paper we introduce a new decision-making method for the assessment of investments and companies. This method is based on the use of fuzzy mid-points and aggregation functions. Moreover, the new methodology incorporates a penalization, fixed following an analytical procedure and thus reducing subjectivity, for the elapsed time and for the discrepancy between OCF and PCF in order to generate the NPV. Finally, all methodologies are tested by applying them to a paradigmatic example where real data is considered and the NPV of a five-year hotel assessment is computed. Here, it is illustrated that the proposed methodologies could be appropriate for the assessment of investments and companies, particularly when operating under uncertainty.