Pub Date : 2021-05-01DOI: 10.7763/IJMO.2021.V11.774
Séamus Lankford, Diarmuid Grimes
The training and optimization of neural networks, using pre-trained, super learner and ensemble approaches is explored. Neural networks, and in particular Convolutional Neural Networks (CNNs), are often optimized using default parameters. Neural Architecture Search (NAS) enables multiple architectures to be evaluated prior to selection of the optimal architecture. Our contribution is to develop, and make available to the community, a system that integrates open source tools for the neural architecture search (OpenNAS) of image classification models. OpenNAS takes any dataset of grayscale, or RGB images, and generates the optimal CNN architecture. Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and pre-trained models serve as base learners for ensembles. Meta learner algorithms are subsequently applied to these base learners and the ensemble performance on image classification problems is evaluated. Our results show that a stacked generalization ensemble of heterogeneous models is the most effective approach to image classification within OpenNAS.
{"title":"Open-Source Neural Architecture Search with Ensemble and Pre-trained Networks","authors":"Séamus Lankford, Diarmuid Grimes","doi":"10.7763/IJMO.2021.V11.774","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.774","url":null,"abstract":"The training and optimization of neural networks, using pre-trained, super learner and ensemble approaches is explored. Neural networks, and in particular Convolutional Neural Networks (CNNs), are often optimized using default parameters. Neural Architecture Search (NAS) enables multiple architectures to be evaluated prior to selection of the optimal architecture. Our contribution is to develop, and make available to the community, a system that integrates open source tools for the neural architecture search (OpenNAS) of image classification models. OpenNAS takes any dataset of grayscale, or RGB images, and generates the optimal CNN architecture. Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO) and pre-trained models serve as base learners for ensembles. Meta learner algorithms are subsequently applied to these base learners and the ensemble performance on image classification problems is evaluated. Our results show that a stacked generalization ensemble of heterogeneous models is the most effective approach to image classification within OpenNAS.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"185 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131916777","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-05-01DOI: 10.7763/IJMO.2021.V11.778
M. Hulea
High accuracy in modelling the behavior of human hand and fingers is obtained using control devices of high biological plausibility. Such devices are typically based on neural networks and are able to control in parallel multiple artificial muscles. This paper presents the structure of an electronic spiking neural network that was implemented to control the force of two opposing fingers of an anthropomorphic hand. In order to increase the level of bio-inspiration, the artificial muscles are implemented using shape memory alloy wires which actuates by contraction as the natural muscles. Moreover, the contraction force of the SMA actuators is directly related to the spiking frequency that is generated by the artificial neurons. The results show that using few excitatory and inhibitory neurons the neural network is able to set and regulate the contraction force of the SMA actuators.
{"title":"Force Control for Anthropomorphic Fingers Actuated by Shape Memory Alloy Wires","authors":"M. Hulea","doi":"10.7763/IJMO.2021.V11.778","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.778","url":null,"abstract":"High accuracy in modelling the behavior of human hand and fingers is obtained using control devices of high biological plausibility. Such devices are typically based on neural networks and are able to control in parallel multiple artificial muscles. This paper presents the structure of an electronic spiking neural network that was implemented to control the force of two opposing fingers of an anthropomorphic hand. In order to increase the level of bio-inspiration, the artificial muscles are implemented using shape memory alloy wires which actuates by contraction as the natural muscles. Moreover, the contraction force of the SMA actuators is directly related to the spiking frequency that is generated by the artificial neurons. The results show that using few excitatory and inhibitory neurons the neural network is able to set and regulate the contraction force of the SMA actuators.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133187388","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-05-01DOI: 10.7763/IJMO.2021.V11.775
N. Yoshida
In this paper, the expectation of the reciprocal of first-degree polynomials of non-negative valued random variables is calculated. This is motivated to compute the Kelly criterion, which is the optimal solution of the maximization of the expected logarithm of the investment return. As soon as the expectation of the reciprocal of first-degree polynomials of asset returns is calculated, which is our main interest, the Kelly criterion can be obtained by using the ordinary optimization technique or applying the appropriate algorithm.
{"title":"On Calculating Method of the Kelly Criterion for Financial Investment in Single Risky Asset with Various Distributions of Returns","authors":"N. Yoshida","doi":"10.7763/IJMO.2021.V11.775","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.775","url":null,"abstract":"In this paper, the expectation of the reciprocal of first-degree polynomials of non-negative valued random variables is calculated. This is motivated to compute the Kelly criterion, which is the optimal solution of the maximization of the expected logarithm of the investment return. As soon as the expectation of the reciprocal of first-degree polynomials of asset returns is calculated, which is our main interest, the Kelly criterion can be obtained by using the ordinary optimization technique or applying the appropriate algorithm.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"292 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132198802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-05-01DOI: 10.7763/IJMO.2021.V11.776
A. Olaru, T. Dobrescu, S. Olaru, I. Mihai
The paper shown one new LabVIEW software platform for the Kinematics analyse in Robotics. This platform contents some more important type of robots and the positions, velocities and accelerations assisted analyse. The program contains a case-type structure with the various types of analysed robots, which also include related Cartesian systems applied in all joints. The front panel of the program contains a twodimensional table with the input data of all relative position vectors between all joints, clusters for defining all robot modules and clusters for defining all parameters of the trapezoidal characteristics of relative motion in all robot’s joints. The clusters that define the robot modules contain information on the translation or rotation couple, the angular or linear home position and respectively the axes of movement by rotation or translation. The results are shown by 3D graphics of space trajectory, of space movement of the velocities and acceleration vectors. With this platform will be possible to quickly analyse some different variants of the movement like simultaneously, successive and complex combination between them and choose the best variant for one good dynamic behaviour without vibration, without pick of moments and forces. This software platform solves one small part of the complex problems of the robot’s kinematics.
{"title":"LabVIEW Software Platform for Kinematics Analyse in Robotics","authors":"A. Olaru, T. Dobrescu, S. Olaru, I. Mihai","doi":"10.7763/IJMO.2021.V11.776","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.776","url":null,"abstract":"The paper shown one new LabVIEW software platform for the Kinematics analyse in Robotics. This platform contents some more important type of robots and the positions, velocities and accelerations assisted analyse. The program contains a case-type structure with the various types of analysed robots, which also include related Cartesian systems applied in all joints. The front panel of the program contains a twodimensional table with the input data of all relative position vectors between all joints, clusters for defining all robot modules and clusters for defining all parameters of the trapezoidal characteristics of relative motion in all robot’s joints. The clusters that define the robot modules contain information on the translation or rotation couple, the angular or linear home position and respectively the axes of movement by rotation or translation. The results are shown by 3D graphics of space trajectory, of space movement of the velocities and acceleration vectors. With this platform will be possible to quickly analyse some different variants of the movement like simultaneously, successive and complex combination between them and choose the best variant for one good dynamic behaviour without vibration, without pick of moments and forces. This software platform solves one small part of the complex problems of the robot’s kinematics.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134604324","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-02-01DOI: 10.7763/IJMO.2021.V11.772
C. Deac, G. Deac, R. Parpală, C. Popa, C. E. Cotet
Identifying the “health state” of the equipment is the domain of condition monitoring. The paper proposes a study of two models: DNN (Deep Neural Network) and CNN (Convolutional Neural Network) over an existent dataset provided by Case Western Reserve University for analyzing vibrations in fault diagnosis. After the model is trained on the windowed dataset using an optimal learning rate, minimizing the cost function, and is tested by computing the loss, accuracy and precision across the results, the weights are saved, and the models can be tested on other real data. The trained model recognizes raw time series data collected by micro electromechanical accelerometer sensors and detects anomalies based on former times series entries.
{"title":"Vibration Anomaly Detection using Deep Neural Network and Convolutional Neural Network","authors":"C. Deac, G. Deac, R. Parpală, C. Popa, C. E. Cotet","doi":"10.7763/IJMO.2021.V11.772","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.772","url":null,"abstract":"Identifying the “health state” of the equipment is the domain of condition monitoring. The paper proposes a study of two models: DNN (Deep Neural Network) and CNN (Convolutional Neural Network) over an existent dataset provided by Case Western Reserve University for analyzing vibrations in fault diagnosis. After the model is trained on the windowed dataset using an optimal learning rate, minimizing the cost function, and is tested by computing the loss, accuracy and precision across the results, the weights are saved, and the models can be tested on other real data. The trained model recognizes raw time series data collected by micro electromechanical accelerometer sensors and detects anomalies based on former times series entries.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132337156","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-02-01DOI: 10.7763/IJMO.2021.V11.770
Dineu Assis, M. Neto, M. Motta
There are many different reasons that can lead a tourist to decide which destination will be chosen on his/her next trip. Besides knowing what are the attractions that must be visited, it is also common to look for more information regarding the overall safety and well-being conditions of travel destinations. Usually shared by local authorities, this kind of information can also be found in a less structured form through public sources, such as web sites and social platforms. However, there are a couple of challenges to be considered: the predominance of unstructured data; the lack of a common standard to distinguish safe and unsafe places; the distinct period needed to update the collected data. In this study, the proposed model combines official census data with open data, social platforms and other online sources, allowing the definition of a score for touristic spots in Lisbon. The resulting score should be able to quantify the community safety and well-being, as well as to identify threats and opportunities for the local tourism industry. Furthermore, it would not only help tourists in their traveling decisions but also, allow decision-makers to track socioeconomic issues and to support public management through a data-driven approach.
{"title":"Community Safety and Well-being in Touristic Spots Using Open Data","authors":"Dineu Assis, M. Neto, M. Motta","doi":"10.7763/IJMO.2021.V11.770","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.770","url":null,"abstract":"There are many different reasons that can lead a tourist to decide which destination will be chosen on his/her next trip. Besides knowing what are the attractions that must be visited, it is also common to look for more information regarding the overall safety and well-being conditions of travel destinations. Usually shared by local authorities, this kind of information can also be found in a less structured form through public sources, such as web sites and social platforms. However, there are a couple of challenges to be considered: the predominance of unstructured data; the lack of a common standard to distinguish safe and unsafe places; the distinct period needed to update the collected data. In this study, the proposed model combines official census data with open data, social platforms and other online sources, allowing the definition of a score for touristic spots in Lisbon. The resulting score should be able to quantify the community safety and well-being, as well as to identify threats and opportunities for the local tourism industry. Furthermore, it would not only help tourists in their traveling decisions but also, allow decision-makers to track socioeconomic issues and to support public management through a data-driven approach.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"26 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122863297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-02-01DOI: 10.7763/IJMO.2021.V11.773
I. Clitan, V. Muresan, M. Abrudean, A. Clitan
This paper presents a home automation plant, consisting of a distributed heating system. It is a system implemented on a residential home, however it could be extended and used for other buildings as well. The paper presents the distributed heating system’s structure, extended from a classical heating system, and the authors also describe the equipment used for the designing and implementation of such a system. The way the system works is depicted, and the authors enfold all the benefits of using such a distributed heating system, such as, increasing the user’s thermal comfort on different living areas and reducing the costs of thermal heating.
{"title":"Distributed Heating System for Residential Homes","authors":"I. Clitan, V. Muresan, M. Abrudean, A. Clitan","doi":"10.7763/IJMO.2021.V11.773","DOIUrl":"https://doi.org/10.7763/IJMO.2021.V11.773","url":null,"abstract":"This paper presents a home automation plant, consisting of a distributed heating system. It is a system implemented on a residential home, however it could be extended and used for other buildings as well. The paper presents the distributed heating system’s structure, extended from a classical heating system, and the authors also describe the equipment used for the designing and implementation of such a system. The way the system works is depicted, and the authors enfold all the benefits of using such a distributed heating system, such as, increasing the user’s thermal comfort on different living areas and reducing the costs of thermal heating.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130434780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-12-01DOI: 10.7763/ijmo.2020.v10.768
Toshitake Araie, Ikeda Tomozumi, A. Kakimoto, S. Adachi
Agricultural tasks result in significant strain on the arms, thereby necessitating posture support. One such task is measuring the photosynthetic capacity of individual leaves. This task requires the operator to hold a measuring device for long periods, which is physically demanding. This study aims to develop an assist suit to reduce the physical load involved in photosynthesis measurement work. We used work posture evaluation methods to quantify the workload of this task and identified the parts of the body at high-risk of injury. Then, we designed an assist suit based on the required specifications and verified its effectiveness.
{"title":"Development of Upper-Limb Assist Suit for Reduction Physical Load in Leaf Photosynthesis Measurement","authors":"Toshitake Araie, Ikeda Tomozumi, A. Kakimoto, S. Adachi","doi":"10.7763/ijmo.2020.v10.768","DOIUrl":"https://doi.org/10.7763/ijmo.2020.v10.768","url":null,"abstract":"Agricultural tasks result in significant strain on the arms, thereby necessitating posture support. One such task is measuring the photosynthetic capacity of individual leaves. This task requires the operator to hold a measuring device for long periods, which is physically demanding. This study aims to develop an assist suit to reduce the physical load involved in photosynthesis measurement work. We used work posture evaluation methods to quantify the workload of this task and identified the parts of the body at high-risk of injury. Then, we designed an assist suit based on the required specifications and verified its effectiveness.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124942396","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-12-01DOI: 10.7763/ijmo.2020.v10.769
G. Deac, C. Georgescu, C. Popa, C. E. Cotet
This paper describes authors’ research in developing collaborative virtual reality applications as an interface for monitoring big data by creating a digital twin of the factory and sync the movement of virtual machines with the real ones. The platform allows an interactive reading of the sensor telemetry and processes data, maintenance information and access to a large technical library. For data acquisition and reports, a novel image data method was used. The data values that are encoded as pixel colors of images, using different encoding methods for each data type (text, integer, float, Boolean) are also encrypted using an image as a symmetric encryption key and are stored in the cloud in a time base folder structure, assuring a better data compression, security and speed, compared with the existing solutions based on JSON and NoSQL databases. The platform allows the remote access from the VR environment to the machines consoles and allows parametrization and remote commands.
{"title":"Virtual Reality Digital Twin for a Smart Factory","authors":"G. Deac, C. Georgescu, C. Popa, C. E. Cotet","doi":"10.7763/ijmo.2020.v10.769","DOIUrl":"https://doi.org/10.7763/ijmo.2020.v10.769","url":null,"abstract":"This paper describes authors’ research in developing collaborative virtual reality applications as an interface for monitoring big data by creating a digital twin of the factory and sync the movement of virtual machines with the real ones. The platform allows an interactive reading of the sensor telemetry and processes data, maintenance information and access to a large technical library. For data acquisition and reports, a novel image data method was used. The data values that are encoded as pixel colors of images, using different encoding methods for each data type (text, integer, float, Boolean) are also encrypted using an image as a symmetric encryption key and are stored in the cloud in a time base folder structure, assuring a better data compression, security and speed, compared with the existing solutions based on JSON and NoSQL databases. The platform allows the remote access from the VR environment to the machines consoles and allows parametrization and remote commands.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123744663","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-12-01DOI: 10.7763/ijmo.2020.v10.766
U. Schmitt
The predicted embracing of thriving knowledge societies is increasingly compromised by threatening perceptions of information overload and attention poverty, opportunity divides and career uncertainties. By integrating system dynamics, discrete-event, and agent-based modeling, this paper traces the roots of these symptoms back to their causes of information entropy and structural holes, invisible private and undiscoverable public knowledge which together characterize the sad state of our current knowledge management (KM) and creation practices. Looking forward, it proposes a decentralized generative KM approach that prioritizes the capacity development of autonomous individual knowledge workers not at the expense but as a viable means to foster a fruitful co-evolution with traditional organizational KM systems. As part of an ongoing design science research and prototyping project, this systems thinking and hybrid model perspective complements a succession of prior multidisciplinary publications on the subject.
{"title":"Systems Dynamics and Activity-Based Modeling to Blueprint Generative Knowledge Management Systems","authors":"U. Schmitt","doi":"10.7763/ijmo.2020.v10.766","DOIUrl":"https://doi.org/10.7763/ijmo.2020.v10.766","url":null,"abstract":"The predicted embracing of thriving knowledge societies is increasingly compromised by threatening perceptions of information overload and attention poverty, opportunity divides and career uncertainties. By integrating system dynamics, discrete-event, and agent-based modeling, this paper traces the roots of these symptoms back to their causes of information entropy and structural holes, invisible private and undiscoverable public knowledge which together characterize the sad state of our current knowledge management (KM) and creation practices. Looking forward, it proposes a decentralized generative KM approach that prioritizes the capacity development of autonomous individual knowledge workers not at the expense but as a viable means to foster a fruitful co-evolution with traditional organizational KM systems. As part of an ongoing design science research and prototyping project, this systems thinking and hybrid model perspective complements a succession of prior multidisciplinary publications on the subject.","PeriodicalId":134487,"journal":{"name":"International Journal of Modeling and Optimization","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129529857","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：