Search

This paper presents the modeling, design, and experimental analysis of an Automatic Generation Control (AGC) for a hydropower plant using Adaptive-Neuro-Fuzzy Inference system (ANFIS). This was aimed at reducing the frequency deviations which occur during power generation. The Adaptive-Neuro-Fuzzy Inference system (ANFIS) was used to intelligently control the selection of parameters for the effective control of power in the hydropower plant. The proposed ANFIS was trained with input–output data of the fuzzy logic controller (FLC). The ANFIS model is used as a hybrid learning model which includes the Least Square Estimate (LSE) and back propagation algorithm (BPA). The conventional PID, FLC, and ANFIS controllers were investigated using MATLAB. In order to determine the best controller, the controllers were experimented and compared to determine the controller with the best performance. The results show that frequency deviations occur as a result of a continuous variation of loads, which make the deviations difficult to control when a governor is not applied. Furthermore, the response of the AGC of the hydropower plant (in the single and double area) with step load changes was studied. The simulation results show that the ANFIS controller performs better compared to the PID as well as the FLC. Further results indicate that the proposed ANFIS controller helps to speed up the performance of the AGC of the hydropower plant. The ANFIS controller not only improved the performance but also made the fuzzy inference system (FIS) less dependent on the expert system.

The objective of this work is to propose a machine-learning assisted multi-objective optimization protocol to design and optimize alkali–surfactant–polymer flooding processes in the presence of multiple technical and economic objective functions. Several universal multi-layer neural networks are trained, and they act as universal surrogate models of high-fidelity numerical simulator to evaluate the objective functions involved in the optimization workflow. The HLD-NAC equation is employed to model the microemulsion phase behavior of the crude oil/brine/surfactant system with the presence of alkali agents. The validity of the expert systems is confirmed via extensive blind testing applications with error margins of 9%, which includes the predictions of fluid production and pressure responses. A multi-objective optimization workflow is structured by coupling the expert systems with particle swarm optimizer, which employs Pareto optimum theory to carry comprehensive ASP injection assessments considering various technical and economic objective functions. Moreover, the proposed workflow enables the decision makers to comprehend the project risks by taking the inherent uncertainties from crude oil market into accounts. The robustness of the optimization protocol is verified by introducing a synthetic field case study, which validates the competence of optimizing ASP injection design considering project net present value as an objective function. A series of Pareto front solutions are generated when more objective functions such as chemical efficacies and water cut reductions are included. Subsequently, the workflow evaluates the project economic uncertainties from the fluctuation of oil price, which helps the decision makers investigate the project risks from technical and economic perspectives.

The Vietnamese stock market provides an interesting and enriching test field for the application of trading expert systems as its economy is opening up, has high growth rate and may offer risk diversification opportunities. This paper examines the question of whether this frontier emerging market offers possibilities for statistical arbitrage through a financial expert system. Based on a sample of the most liquid stocks in the VN30 benchmark index, our results indicate that the index itself and some of its components offer moderate opportunities for statistical arbitrage even after considering transaction costs. It is also found that the purely momentum-based models already work satisfactorily for specific stocks, while the long-short strategies do not work more robustly than the long-only strategies. Overall, our findings hint into the direction of some exploitable inefficiencies, but the magnitude of the tradable volume is such that only comparatively small amounts can be traded.

Manual analysis of electrocardiogram (ECG) signals is a laborious and prone-to-error task, even for a specialist with many hours of experience. For this reason, research on automatic ECG diagnosis is widespread in the literature and continues to grow each year. The present paper describes a novel and fully functional expert system for automatic diagnosis of 13 different diseases using standard 12-lead ECGs. This system makes three significant contributions to the state of the art: (a) the large number of different diseases diagnosed; (b) the use of 5 leads for a more precise identification and measurement of the ECG waves; and (c) a novel noise indicator that measures the quality of the acquired ECG signal. The kernel of the system consists of a set of rules that replicate a specialist’s diagnostic process but with the speed of an automatic system. The rules use a set of parameters generated after a noise-filtering process of the ECG signal and subsequent identification of its different waves (P, QRS complex, T, and Delta). The design of the rules was carried out with the collaboration of a specialist with more than 20 years of experience in ECG diagnosis and using a database of 284,000 ECGs as support. The system was validated by the specialist, obtaining a reliability of 80.8%. Given the complexity of the problem and the number of diagnoses covered, the results are considered satisfactory and make the system a useful support tool for diagnosis.

Several techniques have been proposed to predict the fault proneness of software modules in the absence of fault data. However, the application of these techniques requires an expert assistant and is based on fixed thresholds and rules, which potentially prevents obtaining optimal prediction results. In this study, the development of a fuzzy logic expert system for predicting the fault proneness of software modules is demonstrated in the absence of fault data. The problem of strong dependability with the prediction model for expert assistance as well as deciding on the module fault proneness based on fixed thresholds and fixed rules have been solved in this study. In fact, involvement of experts is more relaxed or provides more support now. Two methods have been proposed and implemented using the fuzzy logic system. In the first method, the Takagi and Sugeno-based fuzzy logic system is developed manually. In the second method, the rule-base and data-base of the fuzzy logic system are adjusted using a genetic algorithm. The second method can determine the optimal values of the thresholds while recommending the most appropriate rules to guide the testing of activities by prioritizing the module’s defects to improve the quality of software testing with a limited budget and limited time. Two datasets from NASA and the Turkish white-goods manufacturer that develops embedded controller software are used for evaluation. The results based on the second method show improvement in the false negative rate, f-measure, and overall error rate. To obtain optimal prediction results, developers and practitioners are recommended to apply the proposed fuzzy logic expert system for predicting the fault proneness of software modules in the absence of fault data.

We present a new release of the expert systems prototyping software Personal Knowledge Base Designer. The software provides the use of a special notation for rule modeling, namely, Rule Visual Modeling Language (RVML); wizards for creating and editing knowledge base elements; conceptual models, and canonical spreadsheet tables as main sources of domain knowledge. The new version presents enhancements of existing tools and introduces new abilities: using decision tables as a new source of domain knowledge; implementation of an RVML extension for modeling incompleteness and inaccuracy in logical rules, namely, FuzzyRVML; generation source codes for PHP (Hypertext Preprocessor) and FuzzyCLIPS.

In Yangtze River traffic management, to manage ship sequencing and scheduling effectively and efficiently in restricted waterways has long been a challenging issue. This paper proposes a Sliding Window based Online ship Sequencing and Scheduling algorithm (OSS-SW) to tackle this problem. The OSS algorithm is capable of generating a more efficient ship sequence by introducing the ‘position shift’ concept which takes advantages of ships’ sailing time differences in the restricted waterway. The sliding window mechanism on the other hand is introduced to handle the traffic uncertainties and to reduce the computational complexity. The impact of the restricted waterway congestion on the performance of OSS-SW is also investigated. Further, the parameter settings of the proposed OSS-SW are investigated in detail. Both simulation studies and the real data applications confirm that the proposed OSS-SW algorithm outperforms the existing widely adopted Traffic Signal Revealing System (TSRS) and the Expert System based algorithm (FAHP-ES) in solving the ship sequencing and scheduling problem.

Nanoimprint lithography (NIL) is a high-throughput and cost-effective technique for fabricating nanoscale features. However, the fine tuning of NIL process parameters is critical in achieving defect-free imprints. Currently, there exists no unified material and process design guidelines to deal with the complex set of process parameters. In this research, an artificial neural network (ANN) algorithm was developed to predict the imprint quality based on a set of input factors collected from experiments and literature. The prediction accuracy was tested in two stages by the use of three ANN models - General Regression Neural Network (GRNN), Back Propagation Neural Network (BPNN), and Probabilistic Neural Network (PNN). Higher prediction accuracy was observed for GRNN in both stages (1st stage = 58.22% and 2nd stage = 92.37%). A significant increase in the prediction accuracy was observed in the second stage after including additional input factors and larger data set. The average prediction accuracies for GRNN, BPNN, and PNN in the second stage were reported as 92.37%, 84.79%, and 91.19%, respectively. The prediction of the GRNN model was validated with experimental results confirming the accurate classification of all the outputs. This research lays the foundation for the development of an expert system for nanoimprint lithography. Keywords: Artificial neural network, Expert systems, Nanoimprint lithography, Process parameters, Prediction accuracy

Group recommender systems are specialized in suggesting preferable products or services to a group of users rather than an individual by aggregating personal preferences of group members. In such expert systems, the initial task is to identify groups of similar users via clustering approaches as user groups are usually not predefined. However, clustering users into groups commonly suffer from sparsity, scalability, and complexity problems as the content in the domain proliferate. Moreover, group homogeneity and size are the critical parameters for organizing group members and enhancing their satisfaction. In this study, we propose novel automatic user grouping approaches by constructing a binary decision tree via bisecting k -means clustering for enhanced group formation and group size restriction. Furthermore, we propose adopting a genre-based mapping of user ratings into a tiny and dense vector to represent users, which both improves computation time for constructing the binary decision tree and enables eliminating adverse effects of sparsity. Finally, since the quality of group formation is not only dependent on conforming preferences but also to the demographic harmony among members, we further introduce utilizing similarities based on demographic characteristics along with the genre-based similarities. We propose applying two distinct strategies for small and large groups by decorating the genre-based similarities with demographic properties, which leads to a more homogeneous automatic group formation. Experiments performed on real-world benchmark datasets demonstrate that each proposed method outperforms its traditional rival significantly, and the final proposed method achieves significantly more qualified ranked recommendation lists than the state-of-the-art algorithm.

Biomedical engineering is the concept of applying fundamental theories and analytical practices to medicine and biology. It can be profitable in the field of healthcare from implementation of medical devices to diagnostic expert systems. These devices and expert systems produce high-dimensional and irregular data. Employing deep learning (DL) algorithms in these devices will be effective for signal analyzing and identification of diseases. DL is a subset of machine learning that employs multiple levels of neural network. It is capable of learning features automatically. The applications of DL in biomedical engineering can be categorized into four fields. They are bio and medical images analysis, brain, body, and machine interface, genomic sequencing and gene expression analysis, and public and medical health management system. This chapter discusses fundamentals of biomedical engineering and DL. It also explores applications of DL in various problems of biomedical field.

Expert systems are a rapidly emerging technology in the field of artificial intelligence (AI), having an immense impact on human healthcare. The main objective of a medical expert system is to help medical professionals in arriving at correct diagnostics. Information gathering is an important part of disease diagnosis, and traditional methods of diagnosis are very time-consuming and require a high level of expertise. Expert systems are computer systems that attempt to imitate the ability of human diagnostic decision-making. They employ knowledge about diseases and facts about patients as data and suggest diagnoses by using machine-learning methods. Expert systems offer suggestions to physicians/area experts in order to improve the expert’s ability and increase the consistency and quality of diagnostics. This type of system is also very helpful for patients who are unable to reach a doctor due to cost, being in a remote area, or feeling too ashamed to discuss their circumstances with a doctor. The expert system also helps to improve decision quality, reduce cost, and maintain consistency, reliability, and speed of diagnosis. There are various learning-based expert systems for different diseases, developed to help doctors and to serve in disease diagnosis. This chapter discusses important existing expert systems for human disease diagnosis in detail. It also provides a brief evaluation of various techniques used in the development of expert systems.

A method is proposed for generating application domain agnostic data for training and evaluating machine learning systems. The proposed method randomly generates an expert system network based upon user specified parameters. This expert system serves as a generic model of an unspecified phenomena. The expert system is run to determine the ideal output from a set of random inputs. These inputs and ideal output are used for training and testing a machine learning system. This allows a machine learning technology to be developed and tested without requiring compatible test data to be collected or before data collection as a proof-of-concept validation of system operations. It also allows systems to be tested without data error noise or with known levels of noise and with other perturbations, to facilitate analysis. It may also facilitate testing system security, adversarial attacks and conducting other types of research into machine learning systems. • Provides an application domain agnostic way to test machine learning technologies and facilitates the generalization of results. • Allows technologies to be tested with data with different characteristics without having to locate datasets that have these characteristics. • Utilizes randomly generated network to represent non-specific phenomena which can be used for training and testing machine learning techniques., Graphical Abstract Image, graphical abstract

Medical expert systems (ESs) aim to apply computer technology to emulate human decision-making and provide computerized clinical decision support to clinicians, patients, and other individuals with suitable information and knowledge at appropriate times to improve the quality and safety of health care. This chapter first presents the significance and a brief history of this field. It then describes a common architecture and major components of medical ESs and introduces different knowledge representation and reasoning techniques. This chapter will also show some examples of medical ESs, including computer-assisted diagnosis systems, medication alert systems, reminder systems, and so on. Lastly, it discusses the advantages and disadvantages of existing approaches as well as major issues and challenges related to system implementation, evaluation, maintenance, and distribution and points out some of the directions for future research and development.

This paper documents published code which can help facilitate researchers with binary classification problems and interpret the results from a number of Machine Learning models. The original paper was published in Expert Systems with Applications and this paper documents the code and work-flow with a special interest being paid to Shapley values as a means to interpret Machine Learning predictions. The Machine Learning models used are, Naive Bayes, Logistic Regression, Random Forest, adaBoost, Classification Tree, Light GBM and XGBoost.

The transportation sector is the second contributor of global GHG emissions, three-quarters of which coming from light and heavy vehicles. As a consequence, sustainable mobility has become a major issue. This chapter focuses on the spread of electric mobility in territories. It expounds an approach aimed at assessing the adoption potential of battery electric mobility in a territory. An expert system evaluates if the conditions required for practicing electromobility are met in the territory being studied, and at what level. This potential is estimated and mapped in 2019 and by 2040 according to two scenarios: “toward all-electric mobility by 2040” and “toward a diversified decarbonized mobility by 2040.” The approach is presented via an application concerning the communes of a French Region, the South Region. The simulations inform not only on the spatial differentiations in the adoption potential and their possible evolutions between the two dates, but also on the communes’ reactivity to change and their adaptability, which is a major factor of resilience.

An expert system is one of the most popular artificial intelligence applications, and replication of human knowledge and expertise to the knowledge-based expert system, which is called knowledge acquisition, is the most difficult task and problematic bottleneck in the course of development and implementation of the system. Although various techniques had been developed for knowledge acquisition such as interview, questionnaire, machine learning techniques, among others, the problem is not fully addressed. This study proposes a new framework termed fuzzy rule-driven data mining framework to knowledge acquisition that addresses the problematic bottleneck of knowledge transfer/acquisition from a human domain expert to the knowledge-based expert system. The framework focuses on extracting useful and novel knowledge from a historical dataset using an improved C4.5 decision tree classification algorithm as a set of crisp rules for subsequent transformation into a set of fuzzy rules with the help of human experts to capture their opinion about the data of the problem domain. The framework was applied in a real-life application, where an expert system for diagnosis of coronary artery diseases (CADs) using a Nigerian-based dataset was developed to validate and evaluate efficiency of the framework. The data mining algorithm used for extracting useful and novel knowledge from the historical dataset has the best accuracy of 86.56% compared to C4.5, which appears to have 83.99%, and random tree with 82.81% accuracy, respectively. In contrast, the expert system developed with the framework called fuzzy rule-driven data mining expert system (FRDDM-ES) for diagnosis of CADs has 93.50% specificity and 90.20% sensitivity with overall accuracy of 91.95%.

This article investigates the domestication of milking robots, most often labelled automatic milking systems (AMS) into dairy farms in Norway. It shows that producers of AMS tend to represent the integration of the technology as a process where their expert systems aid and guide the farmer so that she learns how to be an ‘AMS farmer’. However, farmers' AMS-stories shows us that learning to live with AMS is a process that continues even after the AMS technology seems to have been fully integrated. Furthermore, cows and fellow farmers are central actors, but machines also learn in the process. Hence, we find a extended domestication process where farmers and cows not only adapt to the machine, but indeed that the machine also need to ‘learn’ how to function in particular farms. As such, we target a domestication triangle of machines, animals and farmers where all elements co-evolve. From this we argue that the concept of domestication in studies of agricultural technology needs to take a posthuman turn and focus on a domestication triangle of humans, animals and machines.

The chapter deals with the description of Fussy Logic (FL) principles and its application in power electronics and motor-drive systems. The FL is a discipline under Artificial Intelligence (AI). AI is basically computer emulation of human thinking (called computational intelligence). The goal of AI is to mimic human intelligence. AI also includes expert systems (ESs), artificial neural networks (ANNs), or neural networks (NNWs), and genetic algorithms (GAs). All of the main areas in AI, except ES, are defined as soft computing. The concept of FL and its characteristics emerge from Zadeh's theory propounded in 1965. Any FL application uses a knowledge base that consists of multivalued membership functions (MFs) describing the fuzzy variables and the rule table consisting of “IF… THEN… statements.” The knowledge base is developed on the basis of the behavioral nature of the system. The trial-and-error approach of FL algorithms may be time consuming, but user-friendly computer programs (such as the MATLAB-based Fuzzy Logic Toolbox) help speed the process. The applications of FL principles, include speed control of induction motor vector drives, efficiency optimization of induction motor vector drives by flux programming; and wind generation systems, linearization of the transfer characteristics of thyristor converters at discontinuous conduction, induction motor stator resistance estimations, estimation of distorted waveforms and MRAC slip gain tuning control of vector drives.

Mental workload (MWL) is an imprecise construct, with distinct definitions and no predominant measurement technique. It can be intuitively seen as the amount of mental activity devoted to a certain task over time. Several approaches have been proposed in the literature for the modelling and assessment of MWL. In this paper, data related to two sets of tasks performed by participants under different conditions is reported. This data was gathered from different sets of questionnaires answered by these participants. These questionnaires were aimed at assessing the features believed by domain experts to influence overall mental workload. In total, 872 records are reported, each representing the answers given by a user after performing a task. On the one hand, collected data might support machine learning researchers interested in using predictive analytics for the assessment of mental workload. On the other hand, data, if exploited by a set of rules/arguments (as in [3]), may serve as knowledge-bases for researchers in the field of knowledge-based systems and automated reasoning. Lastly, data might serve as a source of information for mental workload designers interested in investigating the features reported here for mental workload modelling. This article was co-submitted from a research journal "An empirical evaluation of the inferential capacity of defeasible argumentation, non-monotonic fuzzy reasoning and expert systems" [3]. The reader is referred to it for the interpretation of the data.

This paper proposes a new approach to creating Building Information (BIM) models of existing buildings from digitized images. This automatic approach is based on three main steps. The first involves extracting the useful information automatically from rasterized plans by using image processing techniques that include segmentation, filtering, dilation, erosion, and contour detection. This information feeds the knowledge base of an expert system for BIM model generation. In the second step, using the knowledge base of the expert system, the information required to inform the BIM model can be deduced. The range of information thus obtainable can be extended beyond the examples given. The paper concludes with a discussion of the final stage: the automatic generation of an Industry Foundation Classes (IFC) information model with all the desired geometric, physical and technical information. This can be accomplished by using one of the available open-source application program interfaces (APIs). This stage is currently work-in-progress and will be the subject of a future publication.

Artificial Intelligence (AI) is a broad conceptual category of computer software that is designed to mimic, emulate, or improve human decision-making. There are two broad categories of AI: rules-based expert systems (ES), and machine-learning (ML) systems.

Emergency services are most crowded, busy, and fastest serving units of hospitals. Decision-making should be accurate and swift in emergency medicine. An increase in emergency department admissions in recent years has resulted in hastened diagnostic, decision-making, and treatment processes, increased workload, and shortened duration of care. Fortunately, efforts have been made to overcome these newly developed difficulties with technological advances. Topics like machine learning, data mining, expert systems, and fuzzy logic where artificial intelligence cooperates have been largely introduced into emergency medicine practice. This chapter discusses topics about artificial intelligence that are used in emergency medicine.

With rapid progress of computational technology, especially artificial intelligence, there is a paradigm shift to automated expert system in crops' disease and pest management; and precise and localized crop advisory service to the farmers. The advances in the use of different information and communication technology tools viz. wireless sensor network, drones, GPS system, mobile apps, mobile SMS (USSD and voice), website and portal, and RSS feed have altogether changed the way of crop disease and pest monitoring and advisory support to the farmers. Social media connected farmers and experts in real-time basis in crop advisory and feedback console. Major IT companies are showing more interest in commercial basis to crop advisory to the farmers by wise use of artificial intelligence in agriculture. In nutshell, aforesaid efforts reduce the cost of cultivation of crops and reduce the environmental problem due to less use of pesticide through the timely advisory to the farmers.

Due to the extension of road networks and the complexity of problems in various parts of the pavement management system (PMS), the traditional system and classical techniques cannot create an effective system. Accordingly, the application of automatic and expert systems in PMS is unavoidable to create a more efficient and optimal system. The application of computational intelligence (CI) in PMS leads to solve a complex problem and create more expert and optimal systems. This chapter, first, presents a brief explanation of CI frameworks (artificial neural network, fuzzy logic, evolutionary computation, swarm intelligence, and hybrid method), then provides a big-picture from CI frameworks and techniques. Also, it describes the methodology of the newest and more efficient techniques in the different applications of CI, such as learning from the data, solving the problem with uncertainty, and optimization. Finally, it provides a comprehensive view of the CI applications in a different part of PMS.

Machine learning has been of great use for drug discovery and drug design recently. It is hard to find a general algorithm or simple principles for designing drugs to specific cellular processes since they are strongly depending on the many molecules and their structure participating in the particular target processes. Designing new drugs has mostly been a matter of preclinical studies where drug molecules are tested in vitro and in vivo experiments in animals and cellular tissues and with a subsequent data selection and interpretation study. In recent years, these preclinical studies and data selection processes have been aided and, in some cases, even replaced by machine learning techniques such as artificial neural networks (ANNs) and other artificial intelligence (AI) methods, for example, expert systems. These techniques are making use of data driven methods due to the large available databases of drug experiment results. In the present chapter, various AI methods for drug discovery are reviewed, and especially the application of ANNs is described as an approach to design drug molecules with respect to their chemical structure. Also, other AI methods such as purely computational methods are existing as computer software systems where the chemical and atomic composition of a drug molecule are inputs and structural geometry is output data for that particular molecule. In the end, perspectives on possible future developments in drug design and discovery are presented.

For rating of a given network, a mathematical model with its explicit analytical solution is presented for the calculation of target temperatures, hot and cold utilities, and total annual cost (TAC). One example is calculated; the program is attached. For sizing problems, two methods are described, the matrix formulation and the nonlinear programming. For synthesis, first, the Pinch technology using composite curves is explained and demonstrated with two examples. Then, different mathematical programming methods are described. Finally, stochastic and heuristic optimization algorithms are presented, such as genetic, simulated annealing, particle swarm optimization algorithms, and knowledge-based expert system. Twenty-seven synthesis examples with TAC optimization from the literature using different optimization methods are summarized and verified.