The dataset constructed contains images of part of Arabica coffee trees affected by biotic stress... more The dataset constructed contains images of part of Arabica coffee trees affected by biotic stresses. The images were obtained using a smartphone Galaxy S8. The images were collected in September 2019 and March 2020 in the mountain regions of Santa Maria, Marechal Floriano, state of Espirito Santo, Brazil. The photos were taken in a coffee plantation. A total of 300 images of part of Arabica coffee trees were collected, including healthy leaves and diseased leaves, affected by one or more types of biotic stresses: leaf miner, rust, brown leaf spot, and cercospora leaf spot.
The dataset was developed with the purpose to evaluate deep learning algorithms for segmentation ... more The dataset was developed with the purpose to evaluate deep learning algorithms for segmentation and classification. it contains images of arabica coffee leaves affected by the main biotic stresses that affect the coffee tree: leaf miner, leaf rust, brown leaf spot, and cercospora leaf spot. The images were obtained using different smartphones (ASUS Zenfone 2, Xiaomi Redmi 5A, Xiaomi S2, Galaxy S8, and iPhone 6S). The leaves were collected at different times of the year in Santa Maria of Marechal Floreano in the mountains regions of the state of Espirito Santo, Brazil. The photos were taken from the abaxial (lower) side of the leaves under partially controlled conditions and placed on a white background. The acquisition of the images was done without much criterion to make the dataset more heterogeneous. A total of 1747 images of arabica coffee leaves were collected, including healthy leaves and diseased leaves, affected by one or more types of biotic stresses. The process of biotic stresses recognition for dataset labeling was assisted by an expert and performed with the captured images. From the obtained photos were generated two datasets. A dataset with the original images of the entire leaves and a second one containing only symptoms images. Details of each dataset are described in the following. Leaf dataset: It consists of the original images of the entire leaves. The images were labeled in relation to the predominant biotic stress of each leaf and its severity. Stress severity was calculated using the symptom and leaf segmentation mask using automatic image processing methods presented in Manso et al. (2019). For certain severity ranges, labels were assigned as follows: healthy (< 0:1%), very low (0.1% - 5%), low (5% - 10%), high (10% - 15%) and very high (> 15%). Symptom dataset: This dataset was created by cropping the isolated symptoms from the original images in a way that only single stress was present in each image. A total of 2147 symptom images were cropped. Each dataset is divided in trainin [...]
Over the past few years, different computer-aided diagnosis (CAD) systems have been proposed to t... more Over the past few years, different computer-aided diagnosis (CAD) systems have been proposed to tackle skin lesion analysis. Most of these systems work only for dermoscopy images since there is a strong lack of public clinical images archive available to design them. To fill this gap, we release a skin lesion benchmark composed of clinical images collected from smartphone devices and a set of patient clinical data containing up to 22 features. The dataset consists of 1,373 patients, 1,641 skin lesions, and 2,298 images for six different diagnostics: three skin diseases and three skin cancers. In total, 58.4% of the skin lesions are biopsy-proven, including 100% of the skin cancers. By releasing this benchmark, we aim to aid future research and the development of new tools to assist clinicians to detect skin cancer.
Abstract This paper describes how to solve numerical equations of hydraulic problems that involve... more Abstract This paper describes how to solve numerical equations of hydraulic problems that involve the calculation of free and forced channels. The problem is modeled by using the Manning equation. This equation allows the calculation of outflows, inclination of the ...
Machine and deep learning algorithms have increasingly been applied to solve problems in various ... more Machine and deep learning algorithms have increasingly been applied to solve problems in various areas of knowledge. Among these areas, Chemometrics has been benefited from the application of these algorithms in spectral data analysis. Commonly, algorithms such as Support Vector Machines and Partial Least Squares are applied to spectral datasets to perform classification and regression tasks. In this paper, we present a 1D convolutional neural networks (1D-CNN) to evaluate the effectiveness on spectral data obtained from spectroscopy. In most cases, the spectrum signals are noisy and present overlap among classes. Firstly, we perform extensive experiments including 1D-CNN compared to machine learning algorithms and standard algorithms used in Chemometrics on spectral data classification for the most known datasets available in the literature. Next, spectral samples of the SARS-COV2 virus, which causes the COVID-19, have recently been collected via spectroscopy was used as a case study. Experimental results indicate superior performance of 1D-CNN over machine learning algorithms and standard algorithms, obtaining an average accuracy of 96.5%, specificity of 98%, and sensitivity of 94%. The promissing obtained results indicate the feasibility to use 1D-CNN in automated systems to diagnose COVID-19 and other viral diseases in the future.
Convolutional neural networks have been attracted great attention in the realm of complex tasks, ... more Convolutional neural networks have been attracted great attention in the realm of complex tasks, mainly in image recognition. They were specifically designed to process images as inputs, as they act in local receptive fields, performing a convolution process. However, understanding the work principle of this network may not be an easy task, especially for beginners in the area of computational intelligence. Therefore, the objective of this work is to present in a didactic and intuitive way the convolutional neural networks. A case study involving alphabet character recognition is presented in order to ilustrate the feasibility of the approach.
Muitos métodos de tomada de decisão multicritério (do inglês, multi-criteria decision making, abr... more Muitos métodos de tomada de decisão multicritério (do inglês, multi-criteria decision making, abreviada por MCDM) têm sido propostos para lidar com problemas de tomada de decisão incertos. A maioria deles tem como base números nebulosos e não são capazes de lidar com risco no processo de tomada de decisão. Nos últimos anos, alguns métodos MCDM baseados na teoria da propensão para lidar com problemas MCDM têm sido desenvolvidos. Neste artigo, nós estendemos o Fuzzy TODIM para tomada de decisão em grupo, para que seja possível abordar o problema que envolve um grupo de tomadores de decisão. Um estudo de caso envolvendo derramamento de óleo no mar ilustra a aplicação do novo método. Os resultados mostram a viabilidade do método. Palavras chave: Tomadores de decisão multicritério, tomada de decisão em grupo, lógica nebulosa
The Extreme Learning Machine (ELM) is a fast and efficient learning algorithm for single-hidden l... more The Extreme Learning Machine (ELM) is a fast and efficient learning algorithm for single-hidden layer feedforward neural networks (SLFN). Recently,with the increase in data volume in real-world applications, and the need to process data from streaming, two problems have become recurrent in data classification: it is not possible to gather all the necessary data before training the algorithms, and it is difficult to manually label the data for the classification tasks. To address these problems, many variations of ELM have been proposed to allow semi-supervised learning, online sequential learning, or both. In this paper, we propose a variation of ELM called Semi-Supervised Online Elastic Extreme Learning Machine (SSOE-ELM), an algorithm that uses both labeled and unlabeled data to learn in an online sequential way (chunk-by-chunk). We compare our approach to the SOS-ELM in several benchmarks. Our experimental results show that SSOE-ELM outperforms SOS-ELM in accuracy, generalization ability and in training speed.
The Extreme Learning Machine (ELM) is a single-hidden layer feedforward neural network (SLFN) lea... more The Extreme Learning Machine (ELM) is a single-hidden layer feedforward neural network (SLFN) learning algorithm that can learn effectively and quickly. The ELM training phase assigns the input weights and bias randomly and do not change them in the whole process. Although the network works well, the random weights in the input layer can make the algorithm less effective and impact on its performance. Therefore, we propose a new approach to determine the input weights and bias for the ELM using the restricted Boltzmann machine (RBM), which we call RBM-ELM. We compare our new approach with a well-known approach to improve the ELM and a state of the art algorithm to select the weights for the ELM. The results show that the RBM-ELM outperforms both methodologies and achieve a better performance than the ELM.
Skin cancer is one of the most common types of cancer around the world. For this reason, over the... more Skin cancer is one of the most common types of cancer around the world. For this reason, over the past years, different approaches have been proposed to assist detect it. Nonetheless, most of them are based only on dermoscopy images and do not take into account the patient clinical information. In this work, first, we present a new dataset that contains clinical images, acquired from smartphones, and patient clinical information of the skin lesions. Next, we introduce a straightforward approach to combine the clinical data and the images using different well-known deep learning models. These models are applied to the presented dataset using only the images and combining them with the patient clinical information. We present a comprehensive study to show the impact of the clinical data on the final predictions. The results obtained by combining both sets of information show a general improvement of around 7% in the balanced accuracy for all models. In addition, the statistical test indicates significant differences between the models with and without considering both data. The improvement achieved shows the potential of using patient clinical information in skin cancer detection and indicates that this piece of information is important to leverage skin cancer detection systems. Keywords Skin cancer detection • deep learning • data aggregation • clinical images • clinical information 1 Introduction The skin cancer occurs when skin cells are damaged, for example, by overexposure to ultraviolet (UV) radiation from the sun [1]. Although its incidence data are not required to be reported by most cancer registries [2], the World Health Organization (WHO) estimates that one in every three cancers diagnosed is a skin cancer [3]. In countries such as USA, Canada, and Australia, the number of people diagnosed with skin cancer has been increasing at a fairly constant rate over the past decades [1, 4, 5]. In Brazil, according to the Brazilian Cancer Institute (INCA), the skin cancer accounts for 33% of all cancer diagnoses in the country. This is the highest diagnosis rate among all kind of cancer and for 2018-2019 it is expected 180 thousand new cases in the whole country [6]. There are three main types of skin cancer: basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and melanoma. The melanoma is the rarest type of skin cancer, however, due to the high level of metastasis 1 , it is the most lethal one. On the other hand, BCC and SCC, which is known as non-melanoma skin cancer (NMSC), represent the major skin cancer occurrence. As they rarely metastasize, they have low lethality risk [1]. In order to diagnose the skin cancer, dermatologists screen the suspicious skin lesion using their experience to diagnose it. Moreover, they also take into account clinical information such as patient's age, wherein the lesion is located, if the lesion bleeds, among others 1 when damaged cells invade other parts of the body via blood vessels and lymph vessels
, declaro que este trabalho foi escrito e desenvolvido por mim. Toda e qualquer ajuda recebida du... more , declaro que este trabalho foi escrito e desenvolvido por mim. Toda e qualquer ajuda recebida durante todo o processo de desenvolvimento foi devidamente reconhecida. Além disso, certifico que todas as fontes de informação e literatura utilizadas são indicadas na dissertação. Vitória 2016 Agradecimentos Agradeço primeiramente aos meus pais que foram o suporte fundamental por toda minha caminhada. Sem eles, nada disso seria possível. À minha namorada, Mariane, por toda paciência e compreensão desde os tempos de graduação. Ao professor orientador Renato Krohling, fonte de estímulo e enorme conhecimento. Obrigado pela oportunidade, a mim concedida, de trabalhar ao seu lado. Aos professores da banca, Maria Claudia Silva Boeres e João Paulo Papa, por aceitarem o convite de participar da mesma. Aos amigos de laboratório por todas as discussões de ideias e ajuda prestada ao longo deste trabalho. Palavras-chaves: Classificação de dados. Elencos de classificadores. Integral de Choquet. Medida fuzzy. Aprendizado profundo.
Abstract: In this paper we present a new PID controllers design method for disturbance rejection.... more Abstract: In this paper we present a new PID controllers design method for disturbance rejection. The method is based on the minimization of the integral of time multiplied-squared error (ITSE) subject to a disturbance rejection constraint of the type H, norm. A design ...
Bare bones particle swarm optimization (BBPSO) is a well-known swarm algorithm which has shown po... more Bare bones particle swarm optimization (BBPSO) is a well-known swarm algorithm which has shown potential for solving single-objective constrained optimization problems in static environments. In this paper, a generalized BBPSO for dynamic single-objective constrained optimization problems is proposed. An empirical study was carried out to evaluate the performance of the proposed approach. Experimental results show the suitability of the proposed algorithm in terms of effectiveness to find good solutions for all benchmark problems investigated. For comparison purposes, experimental results found by other algorithms are also presented.
In this article, the use of some well-known versions of particle swarm optimization (PSO) namely ... more In this article, the use of some well-known versions of particle swarm optimization (PSO) namely the canonical PSO, the bare bones PSO (BBPSO) and the fully informed particle swarm (FIPS) is investigated on multimodal optimization problems. A hybrid approach which consists of swarm algorithms combined with a jump strategy in order to escape from local optima is developed and tested.
... In Section V the experimental results is shown and discuss, which we compare performance ... ... more ... In Section V the experimental results is shown and discuss, which we compare performance ... multiply vi by (-1) if the particle goes beyond its boundary to search back ... the accelerated co-evolutionary PSO to solve min-max problem, two constraint optimization benchmarks from ...
Classical control methods are still employed in the industry. Proper design and tuning of the con... more Classical control methods are still employed in the industry. Proper design and tuning of the controllers are essential for the acceptable performance of the controlled system. In the past, a considerable number of methods has been developed for this purpose. This paper presents a novel, alternative technique of controller design based on Genetic Algorithms and a laboratory-scale direct current drive as application example. Experimental results show the potential of the method.
This paper presents a co-evolutionary particle swarm optimization (CPSO) algorithm to solve globa... more This paper presents a co-evolutionary particle swarm optimization (CPSO) algorithm to solve global nonlinear optimization problems. A new co-evolutionary PSO (CPSO) is constructed. In the algorithm, a deterministic selection strategy is proposed to ensure the diversity of population. Meanwhile, based on the theory of extrapolation, the induction of evolving direction is enhanced by adding a co-evolutionary strategy, in which the particles make full use of the information each other by using gene-adjusting and adaptive focus-varied tuning operator. Infeasible degree selection mechanism is used to handle the constraints. A new selection criterion is adopted as tournament rules to select individuals. Also, the infeasible solution is properly accepted as the feasible solution based on a defined threshold of the infeasible degree. This diversity mechanism is helpful to guide the search direction towards the feasible region. Our approach was tested on six problems commonly used in the literature. The results obtained are repeatedly closer to the true optimum solution than the other techniques.
In this paper, we propose an alternative novel method based on the Technique for Order Preference... more In this paper, we propose an alternative novel method based on the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to solve the problem of ranking and comparing algorithms. In evolutionary computation, algorithms are executed several times and then a statistic in terms of mean values and standard deviations are calculated. In order to compare algorithms performance it is very common to handle such issue by means of statistical tests. Ranking algorithms, e.g., by means of Friedman test may also present limitations since they consider only the mean value and not the standard deviation of the results. Since the TOPSIS is not able to handle directly this kind of data, we develop an approach based on TOPSIS for algorithm ranking named as A-TOPSIS. In this case, the alternatives consist of the algorithms and the criteria are the benchmarks. The rating of the alternatives with respect to the criteria are expressed by means of a decision matrix in terms of mean values and standard deviations. A case study is used to illustrate the method for evolutionary algorithms. The simulation results show the feasibility of the A-TOPSIS to find out the ranking of algorithms under evaluation.
The dataset constructed contains images of part of Arabica coffee trees affected by biotic stress... more The dataset constructed contains images of part of Arabica coffee trees affected by biotic stresses. The images were obtained using a smartphone Galaxy S8. The images were collected in September 2019 and March 2020 in the mountain regions of Santa Maria, Marechal Floriano, state of Espirito Santo, Brazil. The photos were taken in a coffee plantation. A total of 300 images of part of Arabica coffee trees were collected, including healthy leaves and diseased leaves, affected by one or more types of biotic stresses: leaf miner, rust, brown leaf spot, and cercospora leaf spot.
The dataset was developed with the purpose to evaluate deep learning algorithms for segmentation ... more The dataset was developed with the purpose to evaluate deep learning algorithms for segmentation and classification. it contains images of arabica coffee leaves affected by the main biotic stresses that affect the coffee tree: leaf miner, leaf rust, brown leaf spot, and cercospora leaf spot. The images were obtained using different smartphones (ASUS Zenfone 2, Xiaomi Redmi 5A, Xiaomi S2, Galaxy S8, and iPhone 6S). The leaves were collected at different times of the year in Santa Maria of Marechal Floreano in the mountains regions of the state of Espirito Santo, Brazil. The photos were taken from the abaxial (lower) side of the leaves under partially controlled conditions and placed on a white background. The acquisition of the images was done without much criterion to make the dataset more heterogeneous. A total of 1747 images of arabica coffee leaves were collected, including healthy leaves and diseased leaves, affected by one or more types of biotic stresses. The process of biotic stresses recognition for dataset labeling was assisted by an expert and performed with the captured images. From the obtained photos were generated two datasets. A dataset with the original images of the entire leaves and a second one containing only symptoms images. Details of each dataset are described in the following. Leaf dataset: It consists of the original images of the entire leaves. The images were labeled in relation to the predominant biotic stress of each leaf and its severity. Stress severity was calculated using the symptom and leaf segmentation mask using automatic image processing methods presented in Manso et al. (2019). For certain severity ranges, labels were assigned as follows: healthy (< 0:1%), very low (0.1% - 5%), low (5% - 10%), high (10% - 15%) and very high (> 15%). Symptom dataset: This dataset was created by cropping the isolated symptoms from the original images in a way that only single stress was present in each image. A total of 2147 symptom images were cropped. Each dataset is divided in trainin [...]
Over the past few years, different computer-aided diagnosis (CAD) systems have been proposed to t... more Over the past few years, different computer-aided diagnosis (CAD) systems have been proposed to tackle skin lesion analysis. Most of these systems work only for dermoscopy images since there is a strong lack of public clinical images archive available to design them. To fill this gap, we release a skin lesion benchmark composed of clinical images collected from smartphone devices and a set of patient clinical data containing up to 22 features. The dataset consists of 1,373 patients, 1,641 skin lesions, and 2,298 images for six different diagnostics: three skin diseases and three skin cancers. In total, 58.4% of the skin lesions are biopsy-proven, including 100% of the skin cancers. By releasing this benchmark, we aim to aid future research and the development of new tools to assist clinicians to detect skin cancer.
Abstract This paper describes how to solve numerical equations of hydraulic problems that involve... more Abstract This paper describes how to solve numerical equations of hydraulic problems that involve the calculation of free and forced channels. The problem is modeled by using the Manning equation. This equation allows the calculation of outflows, inclination of the ...
Machine and deep learning algorithms have increasingly been applied to solve problems in various ... more Machine and deep learning algorithms have increasingly been applied to solve problems in various areas of knowledge. Among these areas, Chemometrics has been benefited from the application of these algorithms in spectral data analysis. Commonly, algorithms such as Support Vector Machines and Partial Least Squares are applied to spectral datasets to perform classification and regression tasks. In this paper, we present a 1D convolutional neural networks (1D-CNN) to evaluate the effectiveness on spectral data obtained from spectroscopy. In most cases, the spectrum signals are noisy and present overlap among classes. Firstly, we perform extensive experiments including 1D-CNN compared to machine learning algorithms and standard algorithms used in Chemometrics on spectral data classification for the most known datasets available in the literature. Next, spectral samples of the SARS-COV2 virus, which causes the COVID-19, have recently been collected via spectroscopy was used as a case study. Experimental results indicate superior performance of 1D-CNN over machine learning algorithms and standard algorithms, obtaining an average accuracy of 96.5%, specificity of 98%, and sensitivity of 94%. The promissing obtained results indicate the feasibility to use 1D-CNN in automated systems to diagnose COVID-19 and other viral diseases in the future.
Convolutional neural networks have been attracted great attention in the realm of complex tasks, ... more Convolutional neural networks have been attracted great attention in the realm of complex tasks, mainly in image recognition. They were specifically designed to process images as inputs, as they act in local receptive fields, performing a convolution process. However, understanding the work principle of this network may not be an easy task, especially for beginners in the area of computational intelligence. Therefore, the objective of this work is to present in a didactic and intuitive way the convolutional neural networks. A case study involving alphabet character recognition is presented in order to ilustrate the feasibility of the approach.
Muitos métodos de tomada de decisão multicritério (do inglês, multi-criteria decision making, abr... more Muitos métodos de tomada de decisão multicritério (do inglês, multi-criteria decision making, abreviada por MCDM) têm sido propostos para lidar com problemas de tomada de decisão incertos. A maioria deles tem como base números nebulosos e não são capazes de lidar com risco no processo de tomada de decisão. Nos últimos anos, alguns métodos MCDM baseados na teoria da propensão para lidar com problemas MCDM têm sido desenvolvidos. Neste artigo, nós estendemos o Fuzzy TODIM para tomada de decisão em grupo, para que seja possível abordar o problema que envolve um grupo de tomadores de decisão. Um estudo de caso envolvendo derramamento de óleo no mar ilustra a aplicação do novo método. Os resultados mostram a viabilidade do método. Palavras chave: Tomadores de decisão multicritério, tomada de decisão em grupo, lógica nebulosa
The Extreme Learning Machine (ELM) is a fast and efficient learning algorithm for single-hidden l... more The Extreme Learning Machine (ELM) is a fast and efficient learning algorithm for single-hidden layer feedforward neural networks (SLFN). Recently,with the increase in data volume in real-world applications, and the need to process data from streaming, two problems have become recurrent in data classification: it is not possible to gather all the necessary data before training the algorithms, and it is difficult to manually label the data for the classification tasks. To address these problems, many variations of ELM have been proposed to allow semi-supervised learning, online sequential learning, or both. In this paper, we propose a variation of ELM called Semi-Supervised Online Elastic Extreme Learning Machine (SSOE-ELM), an algorithm that uses both labeled and unlabeled data to learn in an online sequential way (chunk-by-chunk). We compare our approach to the SOS-ELM in several benchmarks. Our experimental results show that SSOE-ELM outperforms SOS-ELM in accuracy, generalization ability and in training speed.
The Extreme Learning Machine (ELM) is a single-hidden layer feedforward neural network (SLFN) lea... more The Extreme Learning Machine (ELM) is a single-hidden layer feedforward neural network (SLFN) learning algorithm that can learn effectively and quickly. The ELM training phase assigns the input weights and bias randomly and do not change them in the whole process. Although the network works well, the random weights in the input layer can make the algorithm less effective and impact on its performance. Therefore, we propose a new approach to determine the input weights and bias for the ELM using the restricted Boltzmann machine (RBM), which we call RBM-ELM. We compare our new approach with a well-known approach to improve the ELM and a state of the art algorithm to select the weights for the ELM. The results show that the RBM-ELM outperforms both methodologies and achieve a better performance than the ELM.
Skin cancer is one of the most common types of cancer around the world. For this reason, over the... more Skin cancer is one of the most common types of cancer around the world. For this reason, over the past years, different approaches have been proposed to assist detect it. Nonetheless, most of them are based only on dermoscopy images and do not take into account the patient clinical information. In this work, first, we present a new dataset that contains clinical images, acquired from smartphones, and patient clinical information of the skin lesions. Next, we introduce a straightforward approach to combine the clinical data and the images using different well-known deep learning models. These models are applied to the presented dataset using only the images and combining them with the patient clinical information. We present a comprehensive study to show the impact of the clinical data on the final predictions. The results obtained by combining both sets of information show a general improvement of around 7% in the balanced accuracy for all models. In addition, the statistical test indicates significant differences between the models with and without considering both data. The improvement achieved shows the potential of using patient clinical information in skin cancer detection and indicates that this piece of information is important to leverage skin cancer detection systems. Keywords Skin cancer detection • deep learning • data aggregation • clinical images • clinical information 1 Introduction The skin cancer occurs when skin cells are damaged, for example, by overexposure to ultraviolet (UV) radiation from the sun [1]. Although its incidence data are not required to be reported by most cancer registries [2], the World Health Organization (WHO) estimates that one in every three cancers diagnosed is a skin cancer [3]. In countries such as USA, Canada, and Australia, the number of people diagnosed with skin cancer has been increasing at a fairly constant rate over the past decades [1, 4, 5]. In Brazil, according to the Brazilian Cancer Institute (INCA), the skin cancer accounts for 33% of all cancer diagnoses in the country. This is the highest diagnosis rate among all kind of cancer and for 2018-2019 it is expected 180 thousand new cases in the whole country [6]. There are three main types of skin cancer: basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and melanoma. The melanoma is the rarest type of skin cancer, however, due to the high level of metastasis 1 , it is the most lethal one. On the other hand, BCC and SCC, which is known as non-melanoma skin cancer (NMSC), represent the major skin cancer occurrence. As they rarely metastasize, they have low lethality risk [1]. In order to diagnose the skin cancer, dermatologists screen the suspicious skin lesion using their experience to diagnose it. Moreover, they also take into account clinical information such as patient's age, wherein the lesion is located, if the lesion bleeds, among others 1 when damaged cells invade other parts of the body via blood vessels and lymph vessels
, declaro que este trabalho foi escrito e desenvolvido por mim. Toda e qualquer ajuda recebida du... more , declaro que este trabalho foi escrito e desenvolvido por mim. Toda e qualquer ajuda recebida durante todo o processo de desenvolvimento foi devidamente reconhecida. Além disso, certifico que todas as fontes de informação e literatura utilizadas são indicadas na dissertação. Vitória 2016 Agradecimentos Agradeço primeiramente aos meus pais que foram o suporte fundamental por toda minha caminhada. Sem eles, nada disso seria possível. À minha namorada, Mariane, por toda paciência e compreensão desde os tempos de graduação. Ao professor orientador Renato Krohling, fonte de estímulo e enorme conhecimento. Obrigado pela oportunidade, a mim concedida, de trabalhar ao seu lado. Aos professores da banca, Maria Claudia Silva Boeres e João Paulo Papa, por aceitarem o convite de participar da mesma. Aos amigos de laboratório por todas as discussões de ideias e ajuda prestada ao longo deste trabalho. Palavras-chaves: Classificação de dados. Elencos de classificadores. Integral de Choquet. Medida fuzzy. Aprendizado profundo.
Abstract: In this paper we present a new PID controllers design method for disturbance rejection.... more Abstract: In this paper we present a new PID controllers design method for disturbance rejection. The method is based on the minimization of the integral of time multiplied-squared error (ITSE) subject to a disturbance rejection constraint of the type H, norm. A design ...
Bare bones particle swarm optimization (BBPSO) is a well-known swarm algorithm which has shown po... more Bare bones particle swarm optimization (BBPSO) is a well-known swarm algorithm which has shown potential for solving single-objective constrained optimization problems in static environments. In this paper, a generalized BBPSO for dynamic single-objective constrained optimization problems is proposed. An empirical study was carried out to evaluate the performance of the proposed approach. Experimental results show the suitability of the proposed algorithm in terms of effectiveness to find good solutions for all benchmark problems investigated. For comparison purposes, experimental results found by other algorithms are also presented.
In this article, the use of some well-known versions of particle swarm optimization (PSO) namely ... more In this article, the use of some well-known versions of particle swarm optimization (PSO) namely the canonical PSO, the bare bones PSO (BBPSO) and the fully informed particle swarm (FIPS) is investigated on multimodal optimization problems. A hybrid approach which consists of swarm algorithms combined with a jump strategy in order to escape from local optima is developed and tested.
... In Section V the experimental results is shown and discuss, which we compare performance ... ... more ... In Section V the experimental results is shown and discuss, which we compare performance ... multiply vi by (-1) if the particle goes beyond its boundary to search back ... the accelerated co-evolutionary PSO to solve min-max problem, two constraint optimization benchmarks from ...
Classical control methods are still employed in the industry. Proper design and tuning of the con... more Classical control methods are still employed in the industry. Proper design and tuning of the controllers are essential for the acceptable performance of the controlled system. In the past, a considerable number of methods has been developed for this purpose. This paper presents a novel, alternative technique of controller design based on Genetic Algorithms and a laboratory-scale direct current drive as application example. Experimental results show the potential of the method.
This paper presents a co-evolutionary particle swarm optimization (CPSO) algorithm to solve globa... more This paper presents a co-evolutionary particle swarm optimization (CPSO) algorithm to solve global nonlinear optimization problems. A new co-evolutionary PSO (CPSO) is constructed. In the algorithm, a deterministic selection strategy is proposed to ensure the diversity of population. Meanwhile, based on the theory of extrapolation, the induction of evolving direction is enhanced by adding a co-evolutionary strategy, in which the particles make full use of the information each other by using gene-adjusting and adaptive focus-varied tuning operator. Infeasible degree selection mechanism is used to handle the constraints. A new selection criterion is adopted as tournament rules to select individuals. Also, the infeasible solution is properly accepted as the feasible solution based on a defined threshold of the infeasible degree. This diversity mechanism is helpful to guide the search direction towards the feasible region. Our approach was tested on six problems commonly used in the literature. The results obtained are repeatedly closer to the true optimum solution than the other techniques.
In this paper, we propose an alternative novel method based on the Technique for Order Preference... more In this paper, we propose an alternative novel method based on the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to solve the problem of ranking and comparing algorithms. In evolutionary computation, algorithms are executed several times and then a statistic in terms of mean values and standard deviations are calculated. In order to compare algorithms performance it is very common to handle such issue by means of statistical tests. Ranking algorithms, e.g., by means of Friedman test may also present limitations since they consider only the mean value and not the standard deviation of the results. Since the TOPSIS is not able to handle directly this kind of data, we develop an approach based on TOPSIS for algorithm ranking named as A-TOPSIS. In this case, the alternatives consist of the algorithms and the criteria are the benchmarks. The rating of the alternatives with respect to the criteria are expressed by means of a decision matrix in terms of mean values and standard deviations. A case study is used to illustrate the method for evolutionary algorithms. The simulation results show the feasibility of the A-TOPSIS to find out the ranking of algorithms under evaluation.
Uploads
Papers by Renato Krohling