Inference System Research Articles

Multi-cloud resource scheduling intelligent system with endogenous security

<abstract> <p>A secure and reliable intelligent multi-cloud resource scheduling system in cyberspace is especially important in some industry applications. However, this task has become exceedingly challenging due to the intricate nature of information, the variety of knowledge representations, the compatibility of diverse knowledge reasoning engines, and the numerous security threats found in cloud networks. In this paper, we applied the endogenous security theory to the multi-cloud resource scheduling intelligent system and presented a novel model of the system. The proposed model incorporates various knowledge representations and inference engines, resulting in a multi-cloud resource scheduling intelligent system that ensures endogenous security. In addition, we have devised a scheme for an intelligent system that schedules multi-cloud resources using dual-channels and has an endogenous security mechanism, which we have named Dynamic, Heterogeneous, and Redundant (DHR). Finally, we have used the multi-cloud resource scheduling intelligent run log database to carry out numerous experiments to validate the efficiency of the dual-channel redundant reasoning system with the endogenous security mechanism's DHR property. The results of the experiment demonstrated that the multi-cloud resource intelligent scheduling system model with an endogenous security mechanism was superior to the current single-channel inference system scheme in regards to security and reliability.</p> </abstract>

Heat and Mass Transfer of Boundary Layer Jeffrey Ternary Hybrid Nanofluid Flow Over Porous Wedge With Surface‐Catalysed Chemical Reactions: An ANFIS Model

The current study employs machine learning (ML) techniques to investigate the heat and mass transport of Jeffrey ternary hybrid nanofluid (THNF). Various influential factors, including magnetic field, thermal radiation, viscous dissipation, activation energy, thermophoresis, Brownian motion and surface‐catalysed chemical reactions, are considered in the analysis. The porous moving wedge supports and influences the flow pattern. For thermal enhancement, three different nanoparticles, namely, Ag (silver), CuO (copper oxide) and SWCNT (single − walled carbon nanotubes), diluted in the regular fluid (blood). Initially, the mathematically modelled partial differential equations (PDEs) are solved numerically by the help of the shooting method. The suitable nondimensional similarity transformations are implied to convert the dimensional PDEs to nondimensional ordinary differential equations (ODEs). The reduced dimensionless nonlinear coupled ODEs are solved using ode‐45 in MATLAB. The impact of φAg, φCuO, φSWCNT in Cfx, Nux, Shx, is displayed in cone plots. It is found that, on increasing φAg, φCuO, φSWCNT, the heat transfer and mass transfer rate gets enhanced in THNF than the hybrid nanofluid (HNF) and nanofluid (NF). The influence of nondimensional parameters over f″(0), θ′(0), ϕ′(0), G′(0) and H′(0) is displayed in 3D contour plots for THNF and HNF. The radiation parameter (R), Brownian motion parameter (Nb), thermophoresis parameter (Nt) and volume fraction parameters φAg, φCuO, φSWCNT boost the energy transfer rate. Finally, the obtained numerical results are split into training and validation datasets that are used in designing the adaptive neuro‐fuzzy inference system (ANFIS) ML model. Five different ANFIS models are developed with the help of nondimensional parameters affecting f″(0), θ′(0), ϕ′(0), G′(0) and H′(0) to predict the physical quantities of dragging force (Cfx), energy transfer rate (Nux), the rate of mass transport (Shx) and mass fluxes for chemical species and , respectively. The training and checking errors attained the convergence less than 2 × 10−4 and 0.2, respectively, for all the five factors. The present ANFIS models have good balance between fitting the training data and generalising to new data, which can make accurate predictions irrespective of variations. The numerical and predicted ANFIS f″(0), θ′(0), ϕ′(0), G′(0) and H′(0) for training data, checking data and results are demonstrated in regression plots. From regression plots, we can observe that the Pearson correlation coefficient attains the positive correlation with R value closer to one. Hence, the developed ANFIS model shows the minimal error in predictions with high degree of accuracy of forecast in THNF flow.

Hybrid ANFIS‐PI‐Based Robust Control of Wind Turbine Power Generation System

This paper introduces a novel hybrid controller designed for a wind turbine power generation system (WTPGS) that utilizes a permanent magnet synchronous generator (PMSG). This hybrid controller combines the adaptability of an adaptive neuro‐fuzzy inference system (ANFIS) with the simplicity of a proportional‐integral (PI) controller. The PI controllers are traditionally used for stability and noise handling. ANFIS adds adaptability, making it more suitable to cope with the variable nature of wind energy. The primary objective of this hybrid strategy is to augment the overall control performance and reliability of PMSG‐based WTPGS when encountered with continuous variable wind conditions. However, implementing the PI controller alone with the WTPGS often suffers from high overshoot and sluggish response in nonlinear systems like WTPGS. In contrast, ANFIS controllers offer superior performance to PI and other artificial intelligence controllers but are still susceptible to noise issues. In this paper, the proposed WTPGS system is designed in MATLAB/Simulink software where a hybrid controller (ANFIS‐PI) is implemented in the machine‐side converter (MSC) and grid‐side converter (GSC) of a variable speed PMSG‐based wind turbine to enhance its performance subjected to wind variations. The hybrid controller is implemented in such a way that the ANFIS controller is implemented in the outer layers while the PI controller is applied in the inner layers of both MSC and GSC. The simulation results for this hybrid controller in the MSC outperform those of the conventional PI controller. They demonstrate minimal overshooting and settling time, maintaining consistent stability even when subjected to various test signals at different intervals. Similarly, the GSC also surpasses conventional PI controllers, achieving a significant 6.4% reduction in maximum overshoot and a decrease of 4.36 seconds in settling time. This highlights its strong suitability for wind turbine applications.

The uneven deployment of algorithms as tools in government: evidence from the use of an expert system

Abstract While governments have long discussed the promise of delegating important decisions to machines, actual use often lags. Consequently, we know little about the variation in the deployment of such delegations in large numbers of similar governmental organizations. Using data from crime laboratories in the United States, we examine the uneven distribution over time of a specific, well-known expert system for ballistics imaging for a large sample of local and regional public agencies; an expert system is an inference engine joined with a knowledge base. Our statistical model is informed by the push-pull-capability theory of innovation in the public sector. We test hypotheses about the probability of deployment and provide evidence that the use of this expert system varies with the pull of agency task environments and the enabling support of organizational resources—and that the impacts of those factors have changed over time. Within this context, we also present evidence that general knowledge of the use of expert systems has supported the use of this specific expert system in many agencies. This empirical case and this theory of innovation provide broad evidence about the historical utilization of expert systems as algorithms in public sector applications.

A Hybrid Inference Engine to Predict Uncertain Sugar Content for Sugar Industry

A Hybrid Inference Engine to Predict Uncertain Sugar Content for Sugar Industry

Paving the way for automated factory planning – applying rule-based expert systems to capacity planning

Today, numerous different software systems aid factory planners in their tasks. Nevertheless, due to their lacking interoperability and project-specific approaches, generic support for automated decision-making is still missing. Investigating the state of the art, we conclude that knowledge-based information modeling is needed for decision-making support. However, as the identified approaches of previous works propose no general automation concepts. Therefore, we define the guiding research question as how to model processual domain knowledge for automating factory planning processes. In this paper, we propose a planning assistance on rule-based expert systems. The planning assistance is composed of an ontology-based information model, a planning model consisting of individual planning functions, and a domain-specific inference engine. We implement the planning assistance with Semantic Web technologies and validate the solution using an application example from capacity planning. Thereby, we demonstrate the applicability of rule-based expert systems for automated factory planning. Finally, implications for future research are drawn for exploring further application areas and developing anticipated hybrid solution concepts.

Energy-Efficient and High-Throughput CNN Inference Engine Based on Memory-Sharing and Data-Reusing for Edge Applications

Energy-Efficient and High-Throughput CNN Inference Engine Based on Memory-Sharing and Data-Reusing for Edge Applications

Low-code development using requirements and knowledge representation models

Low-Code Development is a new software development paradigm which is typically used to generate web applications from high-level visual notations. These notations allow to express the User Interface and Application Logic (usersystem interactions) in a way which is understandable by the end-users. Using certain model-driven approaches, the low-code environments allow for generating the front-end layer and basic CRUD operations in the back-end. Yet still, non-standard domain logic (data processing) operations still necessitate the use of traditional programming. In this article we present a visual language, called RSL-DL that can be used to represent such non-standard domain logic in a visual form. It allows to capture domain knowledge with complex domain rules aligned with requirements models. The language synthesises and extends approaches found in knowledge representation (ontologies) and software modelling language engineering. The development environment of RSL-DL enables fully automatic generation of domain logic code by reasoning over and reusing domain knowledge. The environment includes a dedicated model editor and a transformation engine. The language?s abstract syntax is defined using a meta-model expressed in MOF. Its semantics is expressed with several translational rules that map RSL-DL models onto typical programming language constructs. The article presents a list of these rules in an informal way, and then introduces their formalisation using a graphical transformation notation. The RSL-DL environment includes an inference engine that enables processing queries to domain models and selecting appropriate invocations to generated code. It was also initially validated through studies that involve understandability, operability and complexity assessment. Based on these results, we conclude that declarative knowledge representations can be successfully used to produce imperative backend code with non-trivial logic.

INTEGRATED ARTIFICIAL INTELLIGENCE TECHNIQUES APPLICATION FOR WASTEWATER pH CONTROL

The control of the pH neutralization process has a decisive role in many industrial branches, as: wastewater treatment, biotechnology, pharmaceutical industry, chemical processing, etc. However, the pH neutralization process is a difficult one to control, due to time-varying process parameters, its high non-linearity and to the sudden variation of the pH around the equivalence point. The paper presents a device and an integrated application that uses artificial intelligence (IA) developed for dosage of the necessary reactants in wastewater pH control. The AI techniques were integrated into the application by developing an expert system (ES). For this purpose, were used the necessary if-then rules (rules that make up the inference engine (IE) of the ES) related to the flowrates of the reactants necessary for wastewater pH neutralization. The ES knowledge base (KB) was built based on the recorded pH values. The hardware component (the device) was created using the Arduino equipment.

Intelligent System in Education: Requirements and Design Method

Intelligent systems in education have proven to be highly beneficial in supporting self-learning among students, particularly in remote learning scenarios. This study proposes the necessary requirements for an intelligent educational system that serves two primary functions: querying course knowledge and evaluating learner proficiency through multiple-choice testing. Furthermore, this study builds a solution to design the knowledge base, inference engine, and tracing system based on a knowledge model that integrates ontology and knowledge graph.

Machine learning techniques for the prediction of polymerization kinetics and polymer properties

AbstractIn the current study, the ability of two data‐driven machine learning tools, the extreme learning machine (ELM) and the adaptive neuro‐fuzzy inference system (ANFIS), to predict the polymerization rate and melt flow index of linear low‐density polyethylene produced in a gas phase process was investigated. The level of interaction between the input variables (ethylene, 1‐butene, isopentane pressures, and reaction temperature) on the outputs (melt flow index and activity) was also examined. It was found that both outputs are impacted by the presence of isopentane as an induced condensing agent. Various statistical indicators, including the coefficient of correlation (R2) and root mean square error (RMSE), were used to quantitatively evaluate both developed models. The ANFIS model outperformed the developed ELM model in terms of predicting the MFI and the catalyst activity. A sensitivity analysis of the ANFIS and ELM models showed that all the input variables under investigation had a sizable impact on the responses and none of them could have been discarded. The present study showed that machine learning tools could be employed to adequately develop empirical models to predict polymerization kinetics as well as the final polymer properties.

Framework of Maximum Solar Energy Conversion Using Lyapunov‐Based Model Reference Adaptive Control Technique in Stochastic Weather Conditions

The solar photovoltaic (PV) energy generation technology has grown rapidly. Herein, a novel continuous‐time Lyapunov‐based model reference adaptive control (LB‐MRAC) technique is suggested for determining the duty cycle of boost converter to maintain maximum power point (MPP) of the solar PV panel. For ensuring the fast‐tracking speed, the reference voltage of, respectively, MPP is calculated using perturb and observe (PO) and input to the MRAC technique. Using MATLAB/Simulink, the performances of the LB‐MRAC maximum power point tracking are comparatively analyzed with PO, incremental conductance, and adaptive neuro‐fuzzy inference system under diverse testing modes such as stand‐alone, partial shading, grid integrated, and real time. The proposed technique tracks MPP in just 3.7 ms having tracking efficacy between 99.15% and 99.59% under highly fluctuating irradiance and temperature with load uncertainties. In partial shading mode, four patterns (3 × 1 PV string) are devised and determined where that the suggested approach monitors the global MPP in merely 0.04 s with the highest power tracked and the least amount of shading loss. Next, the efficacy is checked in a three‐phase 50 kW grid‐integrated mode under realistic weather. Finally, the performance of the suggested continuous‐time LB‐MRAC technique is experimentally corroborated employing OPAL‐RT(OP4510) in the real‐time mode.

Expert System For Diagnosing Alzheimer's Using Certainty Factor Method

This expert system for diagnosing Alzheimer's disease is a computer-based stem system used as a tool to diagnose Alzheimer's disease based on a dynamic knowledge base. This knowledge base contains knowledge obtained from various sources including from the experience of experts who are proficient in their fields and also books related to diagnosis of Alzheimer's disease and then collected into a database system needed for making conclusions. In this system, the certainty factor method will be used using a forward chaining inference engine. The purpose of making this root stem p can be useful for many people and can know clearly about Alzheimer's disease from its symptoms and solutions, which will be displayed in the form of a website using PHP programming with MySQL database.

A Rule-Based Expert System for Efficient Food Planning and Consumption

Proper food planning is essential to consume nutritious food that keeps people healthy. Hygienic food offers a favorable existence that helps us to cut the chance of illness and to save time in selecting food. Food planning may ensure a new way to change our buying nature as well as bring variation in our food consumption. Selecting the right food at the right time is necessary. This paper presents an expert system that is intended to select the right food for the right meal for any age people. The expert system also suggests the amount of energy in Kcal required per day. We use rule rule-based expert system to carry out the system. The knowledge base, inference engine, and user interface are combined to give the service. This system helps us to cut the amount of food wastage as well as gives the proper nutritious diet.

Dynamic and Real-Time Object Detection Based on Deep Learning for Home Service Robots

Home service robots operating indoors, such as inside houses and offices, require the real-time and accurate identification and location of target objects to perform service tasks efficiently. However, images captured by visual sensors while in motion states usually contain varying degrees of blurriness, presenting a significant challenge for object detection. In particular, daily life scenes contain small objects like fruits and tableware, which are often occluded, further complicating object recognition and positioning. A dynamic and real-time object detection algorithm is proposed for home service robots. This is composed of an image deblurring algorithm and an object detection algorithm. To improve the clarity of motion-blurred images, the DA-Multi-DCGAN algorithm is proposed. It comprises an embedded dynamic adjustment mechanism and a multimodal multiscale fusion structure based on robot motion and surrounding environmental information, enabling the deblurring processing of images that are captured under different motion states. Compared with DeblurGAN, DA-Multi-DCGAN had a 5.07 improvement in Peak Signal-to-Noise Ratio (PSNR) and a 0.022 improvement in Structural Similarity (SSIM). An AT-LI-YOLO method is proposed for small and occluded object detection. Based on depthwise separable convolution, this method highlights key areas and integrates salient features by embedding the attention module in the AT-Resblock to improve the sensitivity and detection precision of small objects and partially occluded objects. It also employs a lightweight network unit Lightblock to reduce the network’s parameters and computational complexity, which improves its computational efficiency. Compared with YOLOv3, the mean average precision (mAP) of AT-LI-YOLO increased by 3.19%, and the detection precision of small objects, such as apples and oranges and partially occluded objects, increased by 19.12% and 29.52%, respectively. Moreover, the model inference efficiency had a 7 ms reduction in processing time. Based on the typical home activities of older people and children, the dataset Grasp-17 was established for the training and testing of the proposed method. Using the TensorRT neural network inference engine of the developed service robot prototype, the proposed dynamic and real-time object detection algorithm required 29 ms, which meets the real-time requirement of smooth vision.

Binomial distribution based grey wolf optimization algorithm for channel estimation in wireless communication system

SummarySeveral input high‐data‐rate transmissions over broadband wireless channels are possible using multiple input multiple output (MIMO) systems paired with orthogonal frequency division multiplexing (OFDM) technology. Channel estimation is an essential technique and a necessary component of MIMO‐OFDM systems. However, the noise will be there in MIMO‐OFDM due to the environment. As a result, the wireless system performs degrades in terms of bit error rate (BER). The suggested method offers a better pilot pattern strategy for MIMO‐OFDM and an efficient power allocation to address this issue. The binomial distribution‐based grey wolf optimization (BDGWO) algorithm is proposed to identify the optimal pilot patterns. The power is then adaptively distributed to each transmit antenna to increase the spectral efficiency and maximum channel capacity through an adaptive neuro‐fuzzy inference system with a sigmoid membership function (SMFANFIS). The best pilot patterns in PDGSIP (pilot design with generalized shift invariant property) were determined using the BDGWO algorithm based on the binomial distribution. According to the simulation results, the proposed BDGWO established pilot design with generalized shift invariant property (BDGWO‐DGSIP) achieves higher performance compared other existing approaches such as PDGSIP, TPDGSIP, and LS in terms of NMSE, BER, and SER. Compared to the PDGSIP technique, the proposed PDGSIP‐BDGWO system minimizes NMSE at 10%, BER at about 12%, and SER at 15%.

Assessment of governmental strategies for sustainable environment regarding greenhouse gas emission reduction under uncertainty

Since greenhouse gas emissions (GHGE) directly impact climate change that affects the environment, human health, society, and ecosystems, the reduction of GHGE is one of the essential actions for the sustainability of the environment. To reduce global GHGE, the United Nations has defined strategies at three levels: government, private, and public. Choosing between these strategies is a difficult process since there are relationships and contradictions among them. The process also includes uncertainties due to some reasons, such as lack of information, social structure, decision makers’ hesitancy, and imprecision in the collected data. In this paper, a hierarchically structured methodology based on a decision-making procedure is proposed for the evaluation of the governmental strategies determined to decide the best strategies by integrating expert knowledge and a literature review. For this aim, interpretative structural modeling, cognitive mapping, and inference systems are integrated as a two-stage decision-making methodology based on fuzzy sets to address uncertainties and imprecision in the evaluation of government strategies for GHGE reduction in Türkiye. Based on the results of the first stage, “Transportation” is determined as the most influential sector for GHGE mitigation. In the second stage, strategies of the transportation sector are assessed and ranked. “Promoting the significant public health benefits of low-carbon policies, including increased public transportation and non-motorized mobility” is determined as the most appropriate transportation strategy for the governmental action plan regarding the climate change reduction objective. This paper contributes to applying knowledge and experiences from the current environmental characteristics and social fields to the strategic decision of the GHGE reduction area, to streamline its assessment process, provide human-centered solutions, and accelerate governmental actions.

A double inference engine belief rule base for oil pipeline leakage

Accurate prediction of oil pipeline leakage is important for energy security and environmental protection. The belief rule base (BRB) is a rule-based modeling approach that can make use of information with uncertainty to describe causal relationships in predicting oil pipeline leaks. Due to the complexity and uncertainty of oil pipeline systems, traditional BRB models produce a large rule base, which reduces the modeling capability and interpretability of BRB. Therefore, this paper proposes a double inference engine BRB (BRB-DI) model. Compared with the traditional BRB models, the new model diminishes the number of the rules from 56 to 25, while maintaining the accuracy. In the BRB-DI model, firstly, rule reduction is used to reduce the complexity of the model by comprehensively analyzing the importance of rules. Secondly, to ensure the completeness of the model rule base, a double inference engine consisting of the evidence reasoning (ER) algorithm and ER rule is proposed, and a new reasoning computation process is designed. Finally, an optimization algorithm based on projection covariance matrix adaptation evolution strategy (P-CMA-ES) is proposed to prevent the diminishing interpretability of the optimized model. In order to verify the effectiveness of the proposed method, an oil pipeline leakage prediction problem is studied as a numerical example.

Machine learning and knowledge engineering for cognitive memory assessment of age groups by anomalies in a serious game

In psychology, clinical instruments are often used for assessing the cognitive domains and detecting mental alterations. Nevertheless, after the screening stage, the recommendation of activities for training or new assessments could take a long time. Other instruments such digital serious games are commonly used to train cognitive domains, unlike classical instruments, can collect extra data on players' behavior without altering a person's state. In the other hand, decision-making tools are based mainly on two techniques, machine learning for revealing hidden data patterns, and knowledge engineering for building knowledge-based systems. However, the synergy among these both techniques is not completely leveraged, otherwise, when it is, this intelligent data analysis allows the creation of hypotheses and validate contradictions. In this context, this paper presents a tool for supporting the cognitive memory assessment of people based on machine learning and knowledge engineering (M&K-CogMem). The source of data is a matching game for memory training, and after that, this tool determines the memory status of players among their respective age groups, it based on the scores and times in the game. In case of detecting a cognitive memory alteration, it recommends activities for memory improvement through its inference engine. Testing its adoption in industry, an empirical evaluation carried out by psychologists through a technology acceptance model showed positive indicators.

Driving Environment Inference from POI of Navigation Map: Fuzzy Logic and Machine Learning Approaches.

To adapt vehicle control and plan strategies in a predictive manner, it is usually desired to know the context of a driving environment. This paper aims at efficiently inferring the following five driving environments around vehicle's vicinity: shopping zone, tourist zone, public station, motor service area, and security zone, whose existences are not necessarily mutually exclusive. To achieve that, we utilize the Point of Interest (POI) data from a navigation map as the semantic clue, and solve the inference task as a multilabel classification problem. Specifically, we first extract all relevant POI objects from a map, then transform these discrete POI objects into numerical POI features. Based on these POI features, we finally predict the occurrence of each driving environment via an inference engine. To calculate representative POI features, a statistical approach is introduced. To composite an inference engine, three inference systems are investigated: fuzzy inference system (FIS), support vector machine (SVM), and multilayer perceptron (MLP). In total, we implement 11 variants of inference engine following two inference strategies: independent and unified inference strategies, and conduct comprehensive evaluation on a manually collected dataset. The result shows that the proposed inference framework generalizes well on different inference systems, where the best overall F1 score 0.8699 is achieved by the MLP-based inference engine following the unified inference strategy, along with the fastest inference time of 0.0002 millisecond per sample. Hence, the generalization ability and efficiency of the proposed inference framework are proved.