Search

The application of machine learning to physics problems is widely found in the scientific literature. Both regression and classification problems are addressed by a large array of techniques that involve learning algorithms. Unfortunately, the measurement errors of the data used to train machine learning models are almost always neglected. This leads to estimations of the performance of the models (and thus their generalization power) that is too optimistic since it is always assumed that the target variables (what one wants to predict) are correct. In physics, this is a dramatic deficiency as it can lead to the belief that theories or patterns exist where, in reality, they do not. This paper addresses this deficiency by deriving formulas for commonly used metrics (both for regression and classification problems) that take into account measurement errors of target variables. The new formulas give an estimation of the metrics which is always more pessimistic than what is obtained with the classical ones, not taking into account measurement errors. The formulas given here are of general validity, completely model-independent, and can be applied without limitations. Thus, with statistical confidence, one can analyse the existence of relationships when dealing with measurements with errors of any kind. The formulas have wide applicability outside physics and can be used in all problems where measurement errors are relevant to the conclusions of studies.

Cloud vendor selection (CVS) is a complex decision-making problem, which actively adheres to human behavior/cognition. The complex nature of the problem is due to personal biases/hesitation, trade-offs among attributes, uncertainty in rating, and the nonlinear relationship among cloud vendors and associated attributes. In recent times, researchers started paying more attention to user/expert behavior, which led to non-linear decision -making. Most of the extant decision models for CVS considered the linear form of decision-making, which is not realistic due to expert opinions' complexity and dynamism. Motivated by the claim, in this paper, a non-linear decision approach is put forward for CVS. Likert scale rating is adopted for rating cloud vendors based on some attributes, which are transformed to polynomial space from the linear fuzzy space. After this, weights of attri-butes are determined by using CRITIC in the non-linear space. Following this, cloud vendors are ranked in a personalized fashion using the proposed algorithm that encompasses the WASPAS procedure and rank fusion schemes. Finally, a case study is exemplified to validate the usefulness of the decision approach. Comparison and sensitivity analysis showcases the efficacy and robustness of the developed approach.

In the field of decision making, three-way decision making has been proven more fruitful for providing scopes to make delayed decision along with acceptance and rejection simultaneously. As a result, the decision risk and loss, which could occur due to take rapid decisions in traditional two-way decision making, are effectively reduced. Accordingly, this paper offers a novel three-way multi-attribute decision making model by combining three-way decision making and multi-attribute decision making under an incomplete information system. The incertitude in the information system is illustrated by introducing interval-valued Fermatean connection number based on interval-valued Fermatean fuzzy number and set pair analysis theory. Thereafter, the achievement of this study is five-fold. First, a possibility dominance relation is developed under incomplete information system in the basis of possibility degree measure of interval-valued Fermatean connection number. Second, we put forward a novel procedure to set up the fuzzy state set. Third, the conditional probability is estimated in light of the fuzzy state set and probability dominance relation. Fourth, the relative utility functions are determined in the virtue of regret theory. Lastly, a three-way multi-attribute decision making model is implemented in incomplete information system and exploited to deal with incomplete multi-attribute decision making problems. Eventually, the propriety, stability and superiority of the proposed model is established via conducting the comparative and experimental analysis.

Gaussian conditional random fields (GCRF) are a well-known structured model for continuous outputs that uses multiple unstructured predictors to form its features and at the same time exploits dependence structure among outputs, which is provided by a similarity measure. In this paper, a Gaussian conditional random field model for structured binary classification (GCRFBC) is proposed. The model is applicable to classification problems with undirected graphs, intractable for standard classification CRFs. The model representation of GCRFBC is extended by latent variables which yield some appealing properties. Thanks to the GCRF latent structure, the model becomes tractable, efficient and open to improvements previously applied to GCRF regression models. In addition, the model allows for reduction of noise, that might appear if structures were defined directly between discrete outputs. Two different forms of the algorithm are presented: GCRFBCb (GCRGBC - Bayesian) and GCRFBCnb (GCRFBC - non-Bayesian). The extended method of local variational approximation of sigmoid function is used for solving empirical Bayes in Bayesian GCRFBCb variant, whereas MAP value of latent variables is the basis for learning and inference in the GCRFBCnb variant. The inference in GCRFBCb is solved by Newton-Cotes formulas for one-dimensional integration. Both models are evaluated on synthetic data and real-world data. We show that both models achieve better prediction performance than unstructured predictors. Furthermore, computational and memory complexity is evaluated. Advantages and disadvantages of the proposed GCRFBCb and GCRFBCnb are discussed in detail.

Forecasting climate and the development of the environment have been essential in recent days since there has been a drastic change in nature. Weather forecasting plays a significant role in decision-making in traffic management, tourism planning, crop cultivation in agriculture, and warning the people nearby the seaside about the climate situation. It is used to reduce accidents and congestion, mainly based on climate conditions such as rainfall, air condition, and other environmental factors. Accurate weather prediction models are required by meteorological scientists. The previous studies have shown complexity in terms of model building, and computation, and based on theory-driven and rely on time and space. This drawback can be easily solved using the machine learning technique with the time series data. This paper proposes the state-of-art deep learning model Long Short-Term Memory (LSTM) and the Transductive Long Short-Term Memory (T-LSTM) model. The model is evaluated using the evaluation metrics root mean squared error, loss, and mean absolute error. The experiments are carried out on HHWD and Jena Climate datasets. The dataset comprises 14 weather forecasting features including humidity, temperature, etc. The T-LSTM method performs better than other methodologies, producing 98.2% accuracy in forecasting the weather. This proposed hybrid T-LSTM method provides a robust solution for the hydrological variables.

This paper proposes a novel ship energy systems (SESs) benchmarking model for performance measurement of sustainable transportation based on the extension of q-rung orthopair fuzzy rough sets (q-ROFRS) and multicriteria decision-making (MCDM) methods. The underlying research methodology consists of two main stages: (i) Formulation of the SES decision matrix between SESs and the sustainability, (ii) Development of a q-ROFRS and fuzzy-weighted zero-inconsistency (q-ROFRS–FWZIC) model to determine the weights of each criterion. The integrated model of the q-ROFRS and fuzzy decision by the opinion score method (q-ROFRS-FDOSM) is offered as a tool for benchmarking the SESs. Sixty-two SESs are evaluated and benchmarked according to the three layers of criteria concerning the five design alternatives. The analysis of the proposed q-ROFRS–FWZIC methodology revealed that decision support methods (C2) is the most important criterion with a weight of 0.4174, followed by gas emissions (C1.1.2) and economic criterion (C1.1.1) with weights of 0.1661 and 0.1498, respectively; and energy efficiency design index (C1.2.1) is the least important. Furthermore, the results from q-ROFRS-FDOSM reveal that SES62 is the most suitable SES followed by SES60, whereas SES37 is the least suitable. Finally, the robustness of the proposed method is assessed by conducting a sensitivity analysis.

The risk analysis of human error factors is one of the most significant procedures for preventing and reducing risk in the clinical use of medical devices. Human Factor Analysis and Classification System (HFACS) framework, a systematic human error analysis tool, is widely used for human error factors risk analysis. However, the traditional HFACS framework is insufficient to deal with the scenario with complex and uncertain risk information and conflicting opinions among experts caused by their heterogeneous risk preferences and diverse knowledge. To address these limitations, this paper integrated the prospect theory with the consensus model under Interval Type-2 Fuzzy Sets (IT2FSs) environment for addressing the HFACS-based human error factors risk analysis problem. This integrated method enabled the HFACS to yield highly acceptable risk analysis results considering experts' heterogeneous risk preferences. Specifically, the IT2FSs are utilized to represent highly complex and uncertain risk assessment information of human error factors. Secondly, the prospect theory is applied to model the heterogeneous risk preferences of experts and eliminate their impact on risk evaluation results. After obtaining the risk evaluation matrix by prospect theory, the consensual risk evaluation matrix is yielded by a consensus model that can balance the group aim of reaching a consensus and the individual aim of keeping original risk evaluation information as much as possible. Then, the risk ranking of human error factors is determined based on the distance of IT2FSs. Finally, a case study of the clinical use of ventilators, including sensitivity analysis and comparative analysis, is presented to illustrate the efficiency of the proposed method.

Wind resource is one of the most promising renewable energy, which has become a suitable replacement for fossil fuels. Optimizing the transferring wind energy from a wind turbine is essential to obtain the maximum power output as other variables are uncontrollable. This paper presents four different optimization algorithms, namely ant lion optimization (ALO), whale optimization algorithm (WOA), particle swarm optimization (PSO), and crow search optimization (CSO), considering a hybrid decision-making model to compare the performances of wind energy optimization. In the first phase, the evolutionary algorithms are defined based on several factors to meet the need for wind energy based on volumetric and time reliability, reversibility, and vulnerability as well as evaluate optimized energy to the subscriber from the Gansu region. In the second phase, the ordinal priority approach (OPA) is coupled with VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) method to rank the evolutionary algorithms. Then, the results are compared with the absolute optimal response based on the nonlinear programming method obtained from GAMS software. The results demonstrate that an ALO out-performs other algorithms. The average accuracy of ALO is 92%. CSO is the least accurate with 55% of the absolute optimal response. ALO is found to be faster, more efficient, and achieved economy and reliability as compared to other optimization algorithms for solving the problem under consideration. It is shown that the applied models are robust, effective, and able to save costs.

In the interest of developing an intelligent manufacturing environment with an agile, efficient, and optimally utilized transportation system, mobile robots need to achieve a certain level of autonomy as they play an important role in carrying out transportation tasks. Bearing this in mind, in the paper we propose a novel stereo visual servoing method for nonholonomic mobile robot control based on semantic segmentation. Semantic segmentation provides a rich body of information required for an adequate decision-making process in a clustered, dynamic, and ever-changing manufacturing environment. The innovative idea behind the new visual servoing system is to utilize semantic information of the scene for visual servoing, as well as for other mobile robot tasks, such as obstacle avoidance, scene understanding, and simultaneous localization and mapping. Semantic segmentation is carried out by exploiting fully convolutional neural networks. The new visual servoing algorithm utilizes an intensity-based image registration procedure, which results in the image transformation matrix. The transformation matrix encompasses the relations of images taken at the current and desired pose, and that information is directly used for visual servoing. The developed algorithm is deployed on our own developed wheeled differential drive mobile robot RAICO (Robot with Artificial Intelligence based COgnition). The experimental evaluation is carried out in the 3D simulation environment and in the laboratory model of the real manufacturing environment. The experimental results show that the accuracy of the proposed approach is improved when compared to the state-of-the-art approaches while being robust to the partial occlusions of the scene and illumination changes.

Process planning and scheduling are two of the most important manufacturing functions traditionally performed separately and sequentially. These functions being complementary and interrelated, their integration is essential for the optimal utilization of manufacturing resources. Such integration is also significant for improving the performance of the modern manufacturing system. A variety of alternative manufacturing resources (machine tools, cutting tools, tool access directions, etc.) causes integrated process planning and scheduling (IPPS) problem to be strongly NP-hard (non deterministic polynomial) in terms of combinatorial optimization. Therefore, an optimal solution for the problem is searched in a vast search space. In order to explore the search space comprehensively and avoid being trapped into local optima, this paper focuses on using the method based on the particle swarm optimization algorithm and chaos theory (cPSO). The initial solutions for the IPPS problem are presented in the form of the particles of cPSO algorithm. The particle encoding/decoding scheme is also proposed in this paper. Flexible process and scheduling plans are presented using AND/OR network and five flexibility types: machine, tool, tool access direction (TAD), process, and sequence flexibility. Optimal process plans are obtained by multi objective optimization of production time and production cost. On the other hand, optimal scheduling plans are generated based on three objective functions: makespan, balanced level of machine utilization, and mean flow time. The proposed cPSO algorithm is implemented in Matlab environment and verified extensively using five experimental studies. The experimental results show that the proposed algorithm outperforms genetic algorithm (GA), simulated annealing (SA) based approach, and hybrid algorithm. Moreover, the scheduling plans obtained by the proposed methodology are additionally tested by Khepera II mobile robot using a laboratory model of ma

In this paper a brute force logistic regression (LR) modeling approach is proposed and used to develop predictive credit scoring model for corporate entities. The modeling is based on 5 years of data from end-of-year financial statements of Serbian corporate entities, as well as, default event data. To the best of our knowledge, so far no relevant research about predictive power of financial ratios derived from Serbian financial statements has been published. This is also the first paper that generated 350 financial ratios to represent independent variables for 7590 corporate entities default predictions'. Many of derived financial ratios are new and were not discussed in literature before. Weight of evidence (WOE) method has been applied to transform and prepare financial ratios for brute force LR fitting simulations. Clustering method has been utilized to reduce long list of variables and to remove highly correlated financial ratios from partitioned training and validation datasets. The clustering results have revealed that number of variables can be reduced to short list of 24 financial ratios which are then analyzed in terms of default event predictive power. In this paper we propose the most predictive financial ratios from financial statements of Serbian corporate entities. The obtained short list of financial ratios has been used as a main input for brute force LR model simulations. According to literature, common practice to select variables in final model is to run stepwise, forward or backward LR. However, this research has been conducted in a way that the brute force LR simulations have to obtain all possible combinations of models that comprise of 5-14 independent variables from the short list of 24 financial ratios. The total number of simulated resulting LR models is around 14 million. Each model has been fitted through extensive and time consuming brute force LR simulations using SAS (R) code written by the authors. The total number of 342,016 simulat

Understanding and using cost management is a key to success in any type of organization, because it helps managers to collect important data about the costs assigned to products, services and customers and use it for planning and control function. This paper discusses development and application of an expert system for reasonable cost related decision making and continuous cost management process improvement. The system is called COMEX (cost management expert system). This paper also shows how COMEX can be used for improvement of the billing department business processes. COMEX relies on information generated using the activity-based costing (ABC) method, accounting method based on activities, created to improve cost management system by focusing on the driver costs of individual activities. Still, this paper focuses on design and implementation of the COMEX system, and only briefly surveys the ABC method.

The paper describes JavaDON, an open-source expert systems shell based on the OBOA framework for developing intelligent systems. The central idea of the JavaDON project was to make an easy-to-use and easy-to-extend tool for building practical expert systems. Since JavaDON is rooted in a sound theoretical framework, it is well-suited for building even complex expert system applications, both stand-alone and Web-based ones. JavaDON knowledge representation scheme supports using multimedia elements along with traditional techniques, such as rules and frames. Another important feature of JavaDON is its capability of saving knowledge bases in XML format (in addition to the shell's native format), thus making them potentially easy to interoperate, with other knowledge bases on the Internet. So far, JavaDON has been used to build several practical expert systems, as well as a practical knowledge engineering tool to support both introductory and advanced university courses on expert systems. The paper presents design details of JavaDON, explains its links with the underlying OBOA framework, and shows examples of using JavaDON in expert system development. It also discusses some of the current efforts in extending JavaDON.

The paper describes JessGUI, a graphical user interface developed on top of the Jess expert system shell. The central idea of the JessGUI project was to make building, revising, updating, and testing Jess-based expert systems easier, more flexible, and more user friendly. There are many other expert system building tools providing a rich and comfortable integrated development environment to expert system builders. However, they are all either commercial or proprietary products. Jess and JessGUI are open-source freeware, and yet they are well suited for building even complex expert system applications, both stand-alone and Web-based ones. An important feature of JessGUI is its capability of saving knowledge bases in XML format (in addition to the original Jess format), thus making them potentially easy to interoperate with other knowledge bases on the Internet. Jess and JessGUI are also used as practical knowledge engineering tools to support both introductory and advanced university courses on expert systems. The paper presents design details of JessGUI, explains its links with the underlying Jess knowledge representation and reasoning tools, and shows examples of using JessGUI in expert system development. It also discusses some of the current efforts in extending Jess/JessGUI in order to provide intelligent features originally not supported in Jess.

This paper is a detailed case study of building Code Tutor, a Web-based intelligent tutoring system (ITS) in the domain of radio communications. It is ontologically founded and was built using CLIPS and Java-based expert system tools, latest integrated graphical CASE tools for software analysis and design, and Java servlets. In Code Tutor, Apache HTTP Server stores and serves static HTML pages, and Apache JServ Java package enables dynamic interpretation of user defined servlet classes and generation of active HTML pages. XML technology is used to generate files that Code Tutor uses to provide recommendations to the learners. Such a rich palette of integrated advanced technologies has greatly alleviated the system design and implementation, and has also led to interesting solutions of a number of problems common to many ITSs. The paper describes these solutions and useful design decisions, and discusses several practical issues related to architectures of intelligent Web-based applications.

This paper surveys a number of well-known expert systems in the domain of finance. The idea is to illustrate the span of typical expert systems in finance and to provide an insight into the approaches and techniques they employ. The systems considered in the paper are all working systems, and come from different fields of finance. Specifically, the systems considered are FINEVA (from the field of financial analysis), PORT-MAN (banking management), INVEX (investment advisory) and FAME (financial marketing). The paper also briefly presents DEVEX, an expert system for currency exchange advising in international business transactions. Financial institutions in less-developed and undeveloped countries that deal with currency exchange often face some specific problems that do not exist in well-developed regions of the world. DEVEX helps bank employees to cope with such problems. Since business transactions between developed countries and Third World ones represent an important part of the world's financial affairs, the problems covered in DEVEX go beyond local financial institutions.

A major characteristic regarding developments in the broad held of artificial intelligence (AI) during the 1990s has been an increasing integration of AI with other disciplines. A number of other computer science fields and technologies have been used in developing intelligent systems, starting from traditional information systems and databases, to modern distributed systems and the Internet. This paper surveys the knowledge modeling techniques that have received most attention in recent years among developers of intelligent systems, AI practitioners and researchers. The techniques are described from two perspectives, theoretical and practical. Hence the first part of the paper presents major theoretical and architectural concepts, design approaches, and research issues. The second past deals with several practical systems, applications, and ongoing projects that use and implement the techniques described in the first part.

The purpose of this paper is to introduce a novel similarity measure of intuitionistic fuzzy sets (IFSs). The proposed measure is based on the equivalence relation in the IFS-IBA approach. Due to the logic based background, this measure compares IFS from a different viewpoint than the standard measures, emphasizing comprehension of intuitionism. The IFS-IBA similarity measure has a solid mathematical background and can be combined with various IF aggregation operators. Additionally, we define IFS-IBA distance function as a complement of IFS-IBA similarity. Both IFS-IBA similarity and distance functions may have different realizations that are easy to interpret. Hence, the measures are offering great descriptive power and the ability to model various problems. The benefits of the proposed measure are illustrated on the problem of pattern recognition and classification within k-NN algorithm. Finally, we show that the proposed measure is appropriate for IF hierarchical clustering on the problem of clustering Serbian medium-sized companies according to their financial ratios. Results obtained using the IFS -IBA measure are clear-cut and more meaningful compared to a standard IF distances regardless of the I-fuzzification method used.

This paper presents a novel approach for treating uncertainty in the multi-criteria decision making process by introducing interval rough numbers (IRN). The IRN approach enables decision making using only the internal knowledge incorporated in the data provided by the decision maker. A hybrid multi-criteria model was developed based on IRN, and demonstrated using the example of the bidder selection process in the state administration public procurement procedure. The first segment of the hybrid model deals with the rough interval DEMATEL-ANP (IR'DANP) model, which enables more objective expert evaluation of criteria in a subjective environment than the traditional/crisp approach. In the second segment, the evaluation is enabled by applying the new rough interval MAIRCA method, which introduces mathematical tools and shows high stability concerning changes in the nature and characteristics of the criteria. The results of the hybrid IR'DANP-MAIRCA model were analyzed using 36 scenarios of sensitivity analysis, which showed high stability of the results. The results of the interval rough method were compared with the fuzzy extensions of the TOPSIS, VIKOR, MABAC, TODIM, ELECTRE 1 and DEMATEL-ANP models.

This paper presents a novel approach for treating uncertainty in the multi-criteria decision making process by introducing interval rough numbers (IRN). The IRN approach enables decision making using only the internal knowledge incorporated in the data provided by the decision maker. A hybrid multi-criteria model was developed based on IRN, and demonstrated using the example of the bidder selection process in the state administration public procurement procedure. The first segment of the hybrid model deals with the rough interval DEMATEL-ANP (IR'DANP) model, which enables more objective expert evaluation of criteria in a subjective environment than the traditional/crisp approach. In the second segment, the evaluation is enabled by applying the new rough interval MAIRCA method, which introduces mathematical tools and shows high stability concerning changes in the nature and characteristics of the criteria. The results of the hybrid IR'DANP-MAIRCA model were analyzed using 36 scenarios of sensitivity analysis, which showed high stability of the results. The results of the interval rough method were compared with the fuzzy extensions of the TOPSIS, VIKOR, MABAC, TODIM, ELECTRE 1 and DEMATEL-ANP models.

Considering the ever-increasing speed at which new textual content is generated, an efficient and effective use of large text corpora requires automated natural language processing and text analysis tools. A subset of such tools, namely automated semantic annotation tools, are capable of interlinking syntactical forms of text with their underlying semantic concepts. The optimal performance of automated semantic annotation tools often depends on tuning the values of the tools' adjustable parameters to the specificities of the annotation task, and particularly to the characteristics of the text to be annotated. Such characteristics include the text domain, terseness or verbosity level, text length, structure and style. Since the default configuration of annotation tools is not suitable for the large variety of input texts that different combinations of these attributes can produce, users often need to adjust the annotators' tunable parameters in order to get the best results. However, the configuration of semantic annotators is presently a tedious and time consuming task as it is primarily based on a manual trial-and-error process. In this paper, we propose a Parameter Tuning Architecture (PTA) for automating the task of configuring parameter values of semantic annotation tools. We describe the core fitness functions of PTA that operate on the quality of the annotations produced, and offer a solution, based on a genetic algorithm, for searching the space of possible parameter values. Our experiments demonstrate that PTA enables effective configuration of parameter values of many semantic annotation tools.

In this paper, a novel bio-inspired learning control approach (BILCA) for mobile robots based on Learning from Demonstration (LfD), Firefly Algorithm (FA), and homography between current and target camera view is developed. BILCA consists of two steps: (i) first step in which the actuator commands are learned using FA and demonstrations of desired behavior, and (ii) second step in which the obtained wheel commands are evaluated through the real world experiment. Two different problems are considered in this study: trajectory reproduction, and generation of visual control commands for correction of robot orientation. Developed simulations are used to evaluate BILCA in the domain of learning actuator commands for reproduction of different complex trajectories. Results show that the bigger firefly swarms produce better results in terms of accuracy in the final mobile robot pose, and that the desired trajectory is reproduced with minimal error in final control iteration. Likewise, simulations prove that the FA outperforms other metaheuristic techniques. Experiment conducted on a real mobile robot in indoor environment unifies two considered problems within a single transportation task. Depending of the feature position in the image plane, the homography controller for forward motion or the BILCA based controller for robot orientation correction is employed. Experimental results show the applicability and effectiveness of the developed intelligent approach in real world conditions.

In recent years, the need for greater accountability and improvement in teaching's quality has become a major issue in higher education. The results of student evaluation of teaching are increasingly used to make judgments about teaching quality, career advancement, and the funding of teaching itself. An important component of any such student evaluation is the communication of rating results in a manner that allows for fair and meaningful interpretations and comparisons of results by a wide range of evaluation users. In this paper, an approach for a more objective evaluation of university teachers is proposed, one which is based on previously obtained conjoint analysis data concerning criteria importance from a student's point of view. Therein, an illustrative example is given to demonstrate the efficiency and practical usage of such a proposed approach.

Due to complex and nonlinear dynamics of a braking process and complexity in the tire-road interaction, the control of automotive braking systems performance simultaneously with the wheel slip represents a challenging problem. The non-optimal wheel slip level during braking, causing inability to achieve the desired tire-road friction force strongly influences the braking distance. In addition, steerability and maneuverability of the vehicle could be disturbed. In this paper, an active neuro-fuzzy approach has been developed for improving the wheel slip control in the longitudinal direction of the commercial vehicle. The dynamic neural network has been used for prediction and an adaptive control of the brake actuation pressure, during each braking cycle, according to the identified maximum adhesion coefficient between the wheel and road surface. The brake actuation pressure was dynamically adjusted on the level that provides the optimal level of the longitudinal wheel slip vs. the brake pressure selected by driver, the current vehicle speed, the brake interface temperature, vehicle load conditions, and the current value of longitudinal wheel slip. Thus the dynamic neural network model operates (learn, generalize and predict) on-line during each braking cycle, fuzzy logic has been integrated with the neural model as a support to the neural controller control actions in the case when prediction error of the dynamic neural model reached the predefined value. The hybrid control approach presented here provided intelligent dynamic model - based control of the brake actuation pressure in order to keep the longitudinal wheel slip on the optimum level during a braking cycle.

It is known that most of the key problems in visual servo control of robots are related to the performance analysis of the system considering measurement and modeling errors. In this paper, the development and performance evaluation of a novel intelligent visual servo controller for a robot manipulator using neural network Reinforcement Learning is presented. By implementing machine learning techniques into the vision based control scheme, the robot is enabled to improve its performance online and to adapt to the changing conditions in the environment. Two different temporal difference algorithms (Q-learning and SARSA) coupled with neural networks are developed and tested through different visual control scenarios. A database of representative learning samples is employed so as to speed up the convergence of the neural network and real-time learning of robot behavior. Moreover, the visual servoing task is divided into two steps in order to ensure the visibility of the features: in the first step centering behavior of the robot is conducted using neural network Reinforcement Learning controller, while the second step involves switching control between the traditional Image Based Visual Servoing and the neural network Reinforcement Learning for enabling approaching behavior of the manipulator. The correction in robot motion is achieved with the definition of the areas of interest for the image features independently in both control steps. Various simulations are developed in order to present the robustness of the developed system regarding calibration error, modeling error, and image noise. In addition, a comparison with the traditional Image Based Visual Servoing is presented. Real world experiments on a robot manipulator with the low cost vision system demonstrate the effectiveness of the proposed approach.

In today's highly competitive environment, where market oriented firms aim to maximize profits through customer satisfaction, there is an increasing need to design a product line, rather than a single product. The main goal of designing a profit maximizing product line is to target the 'right product' to the 'right customer'. Although conjoint analysis has turned out to be one of the most widely used techniques for product line design, it falls to explicitly consider retaliatory reactions from competitors. In this paper, we propose a new conjoint-based approach to competitive new product line design, employing the Nash equilibrium concept. The optimal product line design problem for each firm is formulated as a nonlinear integer programming problem. In the absence of a closed-form solution, to compute the Nash equilibrium and to determine the optimal product line, we propose a two-phase procedure: a sequential iterative procedure in the first phase, and backward induction in the second. To solve the optimization problem in each of the iterations of the sequential procedure, we used the branch-and-bound method. The proposed approach is illustrated under several scenarios of competition using previously published conjoint data.

This paper presents the design, implementation, and evaluation of a student model in DEPTHS (Design Pattern Teaching Help System), an intelligent tutoring system for learning software design patterns. There are many approaches and technologies for student modeling, but choosing the right one depends on intended functionalities of an intelligent system that the student model is going to be used in. Those functionalities often determine the kinds of information that the student model should contain. The student model used in DEPTHS is a result of combining two widely known modeling approaches, namely, stereotype and overlay modeling. The model is domain independent and can be easily applied in other learning domains as well. To keep student model update during the learning process, DEPTHS makes use of a knowledge assessment method based on fuzzy rules (i.e.; a combination of production rules and fuzzy logics). The evaluation of DEPTHS performed with the aim of assessing the system's overall effectiveness and the accuracy of its student model, indicated several advantages of the DEPTHS system over the traditional approach to learning design patterns, and encouraged us to move on further with this research.

Intelligent modeling, prediction and control of the braking process are not an easy task if using classical modeling techniques, regarding its complexity. In this paper, the new approach has been proposed for easy and effective monitoring, modeling, prediction, and control of the braking process i.e. the brake performance during a braking cycle. The context based control of the disc brake actuation pressure was used for improving the dynamic control of braking process versus influence of the previous and current values of the disc brake actuation pressure, the vehicle speed, and the brake interface temperature. For these purposes, two different dynamic neural models have been developed and integrated into the microcontroller. Microcontrollers are resource intensive and cost effective platforms that offer possibilities to associate with commonly used artificial intelligence techniques. The neural models, based on recurrent dynamic neural networks, are implemented in 8-bit CMOS microcontroller for control of the disc brake actuation pressure during a braking cycle. The first neural model was used for modeling and prediction of the braking process output (braking torque). Based on such acquired knowledge about the real brake operation, the inverse neural model has been developed which was able to predict the brake actuation pressure needed for achieving previously selected (desired) braking torque value in accordance with the previous and current influence of the pressure, speed, and the brake interface temperature. Both neural models have had inherent abilities for on-line learning and prediction during each braking cycle and an intelligent adaptation to the change of influences of pressure, speed, and temperature on the braking process.

The quality of service of agricultural machinery represents one of the basic factors for successful agricultural production. In this sense, there is a clear need for defining the exact indicator of the quality of these machines, according to which it could be possible to determine which machine is optimal for different working conditions. The concept of effectiveness represents one of synthesis indicators of the quality of service of the technical systems. In this paper the effectiveness is defined using the fuzzy set theory, and reliability, maintainability and functionality are used as influence indicators of the effectiveness. In that sense the model for assessing the effectiveness of tractor as a typical representative of agro machinery has been formed. The model is based on integration of linguistic description of the above mentioned influence indicators using fuzzy set theory and max-min composition. The model was tested on the example of three tractors of the same category, which are exploited in climatic and soil conditions in the wider Belgrade (Serbia) area. Even if the conditions in this experiment were approximately equal, the difference of the achieved effects was attained and very significant, compared to other operation parameters.

This paper presents a new generic approach to real-time monitoring of abrupt changes in cutting process. Proposed method is based on hierarchical fuzzy clustering of patterns obtained from discrete wavelet transform (DWT) of acquired signals correlated with cutting force variation in time. Cutting process is naturally highly dynamical and normally consists of mixture of various dynamic phenomena related to the chip formation process and dynamical responses of machining system, workpiece and tool itself. These phenomena are characterized by different time duration. The class of phenomena related to abrupt changes during short time interval is of special importance since they correspond to the most dramatic changes in cutting process, such as various kinds of tool failure or workpiece damage or even breakage. Due to their short time duration, discovery and recognition of these phenomena is extremely difficult. To solve given problem we have chosen DWT, fuzzy clustering and finite state automata as a formal platform for its analysis. Beside its good time localization properties, DWT is, due to asymmetric and irregular shapes of wavelets, especially suitable for analysis of signals having sharp changes or even discontinuities. Given properties make DWT an efficient means for extraction of representative and reliable information contents, thus making good basis for extraction of discriminative and representative features (as DWT coefficients combinations) for classification that will follow. Robustness of specific pattern recognition and learning may be achieved only by taking into consideration wider context. Therefore, in tool condition pattern recognition we have considered the entire context of changes in cutting process state space that precedes and appears after the phenomenon which should be recognized. The cutting process behavior and its evolution in time are considered rather then momentary state which is represented as a point in adopted feature hyperspace of

The paper describes I-Promise, an expert system for configuration design of complex protective materials. Configuration problem of I-Promise is a part of conceptual design of complex protective materials; hence, the context of conceptual design and configuration problems is briefly described before the detailed description of I-Promise, its problem domain, its design, and its application. Due to the complexity and sequential nature of its task, I-Promise is designed as a series of three specialized expert systems. I-Promise is a real-world system, built on top of the information system of an enterprise that produces and delivers complex protective materials.