Online energy management strategy of fuel cell hybrid electric vehicles based on rule learning
استراتژی مدیریت انرژی آنلاین از وسایل نقلیه برقی هیبریدی سلول سوختی بر اساس یادگیری قانون-2020
In this paper, a rule learning based energy management strategy is proposed to achieve preferable energy consumption economy for fuel cell hybrid electric vehicles. Firstly, the optimal control sequence of fuel cell power and the state of charge trajectory of lithium-ion battery pack during driving are derived offline by the Pontryagin’s minimum principle. Next, the K-means algorithm is employed to hierarchically cluster the optimal solution into the simplified data set. Then, the repeated incremental pruning to produce error reduction algorithm, as a propositional rule learning strategy, is leveraged to learn and classify the underlying rules. Finally, the multiple linear regression algorithm is applied to fit the abstracted parameters of generated rule set. Simulation results highlight that the proposed strategy can achieve more than 95% savings of energy consumption economy, solved by Pontryagin’s minimum principle, with less calculation intensity and without dependence on prior driving conditions, thereby manifesting the feasibility of online application.
Keywords: Fuel cell hybrid electric vehicle | Energy management strategy | Hierarchical clustering | Rule learning
مروری سیستماتیک بر مدیریت منابع انسانی سبز : پیامدهایی برای پایداری اجتماعی
سال انتشار: 2020 - تعداد صفحات فایل pdf انگلیسی: 65 - تعداد صفحات فایل doc فارسی: 55
این مقاله پیشرفتهای فعلی و شکافهای تحقیقاتی را در زمینه ادبیات مدیریت منابع انسانی سبز را مورد بررسی قرار میدهد و همچنین به بررسی شیوه های سبز در اینده و در مواجه با امور اجتماعی و الزامات پایداری یک سازمان میپردازد. با در نظر گرفتن افزایش اگاهی در مورد مفهوم سبز بودن و پایداری، یک بررسی سیستماتیک در زمینه ادبیات بصورت خاص انجام شده است که در آن از پایگاه داده های اسکوپوسScopus / و گوگل اسکالر/ Google Scholar استفاده شده است که منجر به انجام مجموعه ایی از 174 مطالعه در بین سالهای 1995 تا 2019 شد. در این مطالعه از نرم افزار NVivo Plus نسخه 12 برای پردازش کمی و همچنین تجزیه و تحلیل کیفی داده ها استفاده شد. کدگذاری محتوا و تجزیه و تحلیل خوشه ای انجام شد، که نتایج آن سه خوشه مجزا بنامهای شیوه های مدیریت منابع انسانی سبز، رفتار سبز کارکنان در محل کار و ثبات سازمانی را به نمایش گذاشتند. تجزیه و تحلیل دستی بیشتر نشان داد پایداری اجتماعی بعنوان یکی از حداقل حوزه های فراتر از ارکان اقتصادی و زیست محیطی پایداری است. از اینرو، نویسندگان بصورت مفهومی یک مدل نظری ارایه دادند که نشان دهنده نقش میانجی گری رفتار سبز کارکنان در محل کار در ارتباط با شیوه های مدیریت منابع انسانی سبز و ثبات اجتماع سازمانهایی است که از رویکرد نظریه داده بنیاد استفاده میکنند. این مطالعه اخرین تحقیقات مرتبط به تحقیقات مدیریت منابع انسانی سبز را بمنظور ثبات اجتماعی پیش بینی نشده را در زمینه ثبات اجتماعی مورد بررسی قرار میدهد، که تا کنون مورد ارزیابی قرار گرفته اند. بر اساس محتوای کدگذاری، خوشه بندی، و تجزیه و حلیل های بیشتر، گزاره ها ، مسیرها و پیامدهایی نیز برای آینده در نظر گرفته شده اند.
واژگان کلیدی: شیوه های مدیریت منابع انسانی سبز | رفتار سبز کارکنان در محل کار | ثبات سازمانی | مدیریت منابع انسانی پایدار | ثبات اجتماعی | تجزیه و تحلیل محتوا.
|مقاله ترجمه شده|
The clustering algorithm for efficient energy management in mobile ad-hoc networks
الگوریتم خوشه بندی برای مدیریت انرژی کارآمد در شبکه های ad-hoc تلفن همراه-2020
MANET (Mobile Ad-hoc Network) consumes much energy due to their dynamic capabilities, complex- ities, constraints of design such as lack of a specified communication infrastructure, and their change over time. One strategy to reduce energy consumption is to optimize routing in these networks, which is, in turn, one of the most important challenges in these networks. In addition, optimal routing will in- crease the network lifetime and lead to its stability. Considering the high efficiency of clustering methods among the routing algorithms we present a new clustering method and considering good performance of Evolutionary Algorithms (EAs) in finding proper head clusters, we present a specific EA-based method named ICA (Imperialist Competitive Algorithm) via numerical coding. By thinking of specific conditions of a MANET and estimating the mobility direction of nodes, we prevent from additional reclusterings lead- ing to reducing the overload. We have evaluated our proposed method for: 1) accuracy (including the reproducibility, convergence and stability criteria) through three case studies with different numbers of nodes and ranges and 2) efficiency (by comparing to other methods). Moreover, we applied our proposed method to a case study (multi-robot system) with low velocity.
Keywords: ICA | MANET | Clustering | Energy consumption
Improving response of wind turbines by pitch angle controller based on gain-scheduled recurrent ANFIS type 2 with passive reinforcement learning
بهبود پاسخ توربین های بادی توسط کنترل کننده زاویه گام بر اساس ANFIS تکرار شونده نوع 2 با یادگیری تقویتی غیرفعال-2020
In this paper, passive reinforcement learning (RL) solved by particle swarm optimization policy (PSOeP) is used to handle an adaptive neuro-fuzzy inference system (ANFIS) type-2 structure with unsupervised clustering for controlling the pitch angle of a real wind turbine (WT). The proposed control scheme is based on gain-scheduled reinforcement learning recurrent ANFIS type 2 (GS-RL-RANFIST2) pitch angle controller to maintain the rotor speed at its rated value while smoothing the output power and the performance of the pitch angle system. The practical application of the proposed controller is evaluated by using FAST tool for a real 600 kW WT equipped with a synchronous generator with a full-size power converter (CART3, located at the National Renewable Energy Laboratory, NREL), whose results are compared with those obtained by a gain corrected proportional integral (GC-PI) controller. The results demonstrate that the GS-RL-RANFIST2, which sets the nonlinear characteristics of the system automatically and waves more uncertainties in the windy conditions, allows to increase the energy capture and smooth the output power fluctuation, and therefore, to improve the control and response of theWT.
Keywords: Pith angle controller | Wind turbine | Gain-scheduled | ANFIS type-2 controller | Reinforcement learning (RL) | Unsupervised clustering
A new fast search algorithm for exact k-nearest neighbors based on optimal triangle-inequality-based check strategy
یک الگوریتم جستجوی سریع جدید برای همسایگان دقیق k-مبتنی بر استراتژی بررسی مبتنی بر مثلث-نابرابری بهینه-2020
The k-nearest neighbor (KNN) algorithm has been widely used in pattern recognition, regression, outlier detection and other data mining areas. However, it suffers from the large distance computation cost, especially when dealing with big data applications. In this paper, we propose a new fast search (FS) algorithm for exact k-nearest neighbors based on optimal triangle-inequality-based (OTI) check strategy. During the procedure of searching exact k-nearest neighbors for any query, the OTI check strategy can eliminate more redundant distance computations for the instances located in the marginal area of neighboring clusters compared with the original TI check strategy. Considering the large space complexity and extra time complexity of OTI, we also propose an efficient optimal triangle-inequalitybased (EOTI) check strategy. The experimental results demonstrate that our proposed two algorithms (OTI and EOTI) achieve the best performance compared with other related KNN fast search algorithms, especially in the case of dealing with high-dimensional datasets
Keywords: Exact k-nearest neighbors | Fast search algorithm | Clustering | Triangle inequality | Optimal check strategy
Exploration of the mechanism of traditional Chinese medicine by AI approach using unsupervised machine learning for cellular functional similarity of compounds in heterogeneous networks, XiaoErFuPi granules as an example
کاوش مکانیسم طب سنتی چینی با رویکرد هوش مصنوعی با استفاده از یادگیری ماشین بدون نظارت برای شباهت عملکردی سلولی ترکیبات در شبکه های ناهمگن ، گرانول های XiaoErFuPi به عنوان مثال-2020
‘Polypharmacology’ is usually used to describe the network-wide effect of a single compound, but traditional Chinese medicine (TCM) has a polypharmacological effect naturally based on the ‘multi-components, multitargets and multi-pathways’ principle. It is a challenge to investigate the polypharmacology mechanism of TCM with multiple components. In this study, we used XiaoErFuPi (XEFP) granules as an example to describe an unsupervised learning strategy for polypharmacology research of TCM and to explore the mechanism of XEFP polypharmacology against multifactorial disease function dyspepsia (FD). Unsupervised clustering of compounds based on similarity evaluation of cellular function fingerprints showed that compounds of TCM without similar targets and chemical structure could also exert similar therapeutic effects on the same disease, as different targets participate in the same pathway closely associated with the pathological process. In this study, we proposed an unsupervised machine learning strategy for exploring the polypharmacology-based mechanism of TCM, utilizing hierarchical clustering based on cellular functional similarity, to establish a connection from the chemical clustering module to cellular function. Meanwhile, FDA-approved drugs against FD were used as references for the mechanism of action (MoA) of FD. First, according to the compound-compound network built by the similarity of cellular function of XEFP compounds and FDA-approved FD drugs, the possible therapeutic function of TCM may represent a known mechanism of FDA-approved drugs. Then, as unsupervised learning, hierarchical clustering of TCM compounds based on cellular function fingerprint similarity could help to classify the compounds into several modules with similar therapeutic functions to investigate the polypharmacology effect of TCM. Furthermore, the integration of quantitative omics data of TCM and approved drugs (from LINCS datasets) provides more quantitative evidence for TCM therapeutic function consistency with approved drugs. A spasmolytic activity experiment was launched to confirm vanillic acid activity to repress smooth muscle contraction; vanillic acid was also predicted to be active compound of XEFP, supporting the accuracy of our strategy. In summary, the approach proposed in this study provides a new unsupervised learning strategy for polypharmacological research investigating TCM by establishing a connection between the compound functional module and drug-activated cellular processes shared with FDA-approved drugs, which may elucidate the unique mechanism of traditional medicine using FDA-approved drugs as references, facilitate the discovery of potential active compounds of TCM and provide new insights into complex diseases.
Keywords: Polypharmacology | Traditional Chinese medicine | Unsupervised clustering | Cellular function fingerprints | FDA-approved drugs | Functional dyspepsia
Forecasting across time series databases using recurrent neural networks on groups of similar series: A clustering approach
پیش بینی پیش بینی پایگاه داده های سری زمانی با استفاده از شبکه های عصبی مکرر در گروه های مشابه سری: یک روش خوشه بندی-2020
With the advent of Big Data, nowadays in many applications databases containing large quantities of sim- ilar time series are available. Forecasting time series in these domains with traditional univariate fore- casting procedures leaves great potentials for producing accurate forecasts untapped. Recurrent neural networks (RNNs), and in particular Long Short Term Memory (LSTM) networks, have proven recently that they are able to outperform state-of-the-art univariate time series forecasting methods in this context, when trained across all available time series. However, if the time series database is heterogeneous, ac- curacy may degenerate, so that on the way towards fully automatic forecasting methods in this space, a notion of similarity between the time series needs to be built into the methods. To this end, we present a prediction model that can be used with different types of RNN models on subgroups of similar time series, which are identified by time series clustering techniques. We assess our proposed methodology using LSTM networks, a widely popular RNN variant, together with various clustering algorithms, such as kMeans, DBScan, Partition Around Medoids (PAM), and Snob. Our method achieves competitive results on benchmarking datasets under competition evaluation procedures. In particular, in terms of mean sMAPE accuracy it consistently outperforms the baseline LSTM model, and outperforms all other methods on the CIF2016 forecasting competition dataset.
Keywords: Big data forecasting | RNN | LSTM | Time series clustering | Neural networks
Parallel hierarchical architectures for efficient consensus clustering on big multimedia cluster ensembles
معماری سلسله مراتبی موازی برای خوشه بندی اجماع کارآمد در مجموعه های بزرگ خوشه چندرسانه ای-2020
Consensus clustering is a useful tool for robust or distributed clustering applications. How- ever, given the fact that time complexities of the consensus functions scale linearly or quadratically with the number of combined clusterings, execution can be slow or even impossible when operating on big cluster ensembles, a situation encountered when we pursue robust multimedia data clustering. This work introduces hierarchical consensus ar- chitectures, an inherently parallel approach based on the divide-and-conquer strategy for computationally efficient consensus clustering, in a bid to make faster, more effective con- sensus clustering possible in big multimedia cluster ensemble scenarios. Moreover, we de- fine a specific implementation of hierarchical architectures, including a theoretical analysis of its fully parallel implementation computational complexity. In experiments conducted on unimodal and multimedia data sets involving small and big cluster ensembles, we find parallel hierarchical consensus architectures variants perform faster than traditional flat consensus in 75% of the experiments on small cluster ensembles, a percentage that rises to 100% on unimodal and multimedia big cluster ensembles, achieving an average speedup ratio of 30.5. Moreover, depending on the consensus function employed, the quality of the obtained consensus partitions ensures robust clustering results.
Keywords: Consensus clustering | Big cluster ensembles | Multimedia clustering | Parallelization | Divide-and-conquer
Privacy-preserving clustering for big data in cyber-physical-social systems: Survey and perspectives
خوشه بندی حفظ حریم خصوصی برای داده های بزرگ در سیستم های سایبر-فیزیکی-اجتماعی: بررسی و چشم انداز-2020
Clustering technique plays a critical role in data mining, and has received great success to solve application problems like community analysis, image retrieval, personalized rec- ommendation, activity prediction, etc. This paper first reviews the traditional clustering and the emerging multiple clustering methods, respectively. Although the existing meth- ods have superior performance on some small or certain datasets, they fall short when clustering is performed on CPSS big data because of the high cost of computation and stor- age. With the powerful cloud computing, this challenge can be effectively addressed, but it brings enormous threat to individual or company’s privacy. Currently, privacy preserving data mining has attracted widespread attention in academia. Compared to other reviews, this paper focuses on privacy preserving clustering technique, guiding a detailed overview and discussion. Specifically, we introduce a novel privacy-preserving tensor-based multi- ple clustering, propose a privacy-preserving tensor-based multiple clustering analytic and service framework, and give an illustrated case study on the public transportation dataset. Furthermore, we indicate the remaining challenges of privacy preserving clustering and discuss the future significant research in this area.
Keywords: CPSS | Big data | Cloud computing | Privacy preserving | Clustering
Explainable AI: A Hybrid Approach to Generate Human-Interpretable Explanation for Deep Learning Prediction
هوش مصنوعی قابل توضیح: رویکرد ترکیبی برای ایجاد توضیح قابل تفسیر توسط انسان برای پیش بینی یادگیری عمیق-2020
With massive computing power and data explosion as catalysts, Artificial Intelligence (AI) has finally come out of research labs to become a ground-breaking technology. Businesses are seeing its value in a wide range of applications and therefore looking for ways to make AI an integral part of their decision-making processes. However, to trust an AI model prediction or to take downstream action based on a prediction outcome, one needs to understand the reasons for the prediction. With deep neural networks increasingly becoming the algorithm of choice for models, generation of such reasons has become more challenging. Deep neural networks are highly nested non-linear models that learn patterns in the data through complex combinations of inputs. Their complex architecture makes it very difficult to decipher the exact reasons for their prediction. Due to this lack of transparency, businesses are not able to utilize this technology in many applications. To increase the adoption of deep learning models, explainability is critical in building trust in the solution and in guiding downstream actions in business applications. In this paper we aim to create human-interpretable explanations for predictions from deep learning models. We propose a hybrid of two prior approaches, integrating clustering of the network’s hidden layer representation  with TREPAN decision tree , both of which uniquely deconstruct a neural network. Our aim is to visualize flow of information within the deep neural network using factors that make sense to humans, even if the underlying model uses more complex factors. This enables generation of human interpretable explanations (or, reasons codes) for each model outcome at an individual instance level. We demonstrate the new approach on credit card default prediction given by a deep feed forward neural network model. We compare and contrast this new integrated approach with three different approaches, based on the results we obtained from experimentation.
Keywords: Deep Learning | Neural Network | Explainable AI | TREPAN | Clustering | Reason Code | Comprehensibility | Fidelity | LIME