Search

The second-hand vessel market has quite different dynamics than the market of the new-building vessel, and highly complicated and conflicting criteria and many uncertainties affect the ship selection process. Therefore, it is required to use a robust mathematical model to solve these kinds of decision-making problems. For this purpose, this paper presents an extended version of WASPAS (Weighted Aggregated Sum Product ASsessment) techniques with the help of T2NN based on the Bonferroni function (T2NN WASPAS'B). The three main focal points of the proposed approach are (i) setting the influential criteria to select the appropriate Ro-Ro vessel in the second-hand vessel market; and (ii) presenting a flexible group decision-making approach, which is proper to real decision-making problems. (iii) detecting the interrelations among criteria and eliminating the negative impacts of undesirable and excessive values in input variables on the results. Practical use of the proposed approach is demonstrated to select the appropriate Ro-Ro vessel in the second-hand market. The analysis results show that the most effective and determinative factor is Trailer Lane length, and the most effective alternative is GREIFSWALD. Besides, the consistency and validity of the obtained results have been verified with the help of a stability and robustness check. The results prove that the proposed novel T2NN WASPAS'B model is robust, powerful, and reliable for making rational and realistic decisions.

Traffic congestion is, nowadays, one of the most important highway problems. Highway tolls with booth operators are one of the causes of traffic congestion on highways, especially in rush hour periods, or during seasonal holiday travels. The value of driver waiting time (needed to stop and pay the toll) and the cost of the toll booth operators can reach up to about one-third of the revenue. In this paper we propose a novel methodology for continuous-time optimal control of highway tolls by predicting the optimal number of active modules (booths) in toll stations. The proposed methodology is based on a combination of recurrent neural networks, queuing theory, and metaheuristics. We utilized several recurrent neural network architectures for predicting the average intensity of vehicle arrivals. Moreover, the prediction error of the first recurrent neural network was modelled by another one in order to provide confidence estimates, additional regularization, and robustness. The predicted intensity of vehicle arrival rates was used as an input of the queuing model, whereas differential evolution was applied to minimize the total cost (waiting and service costs) by determining the optimal number of active modules on a highway toll in continuous time. The developed methodology was experimentally tested on real data from highway E70 in the Republic of Serbia. The obtained results showed significantly better performance compared to the currently used toll station opening pattern. The solutions obtained by solving a system of differential equations of the queuing model were also validated by a simulation procedure.

The current study is mainly devoted to explore and extend the measurement and ranking of alternatives according to the compromise solution under the background of 2-tuple linguistic q-rung picture fuzzy sets. The capability of q-rung picture fuzzy sets to accommodate a broad range of information, and the salient features of 2-tuple linguistic term sets to manage qualitative data motivate us to present the notion of 2-tuple linguistic q-rung picture fuzzy set. The linguistic terms are very useful to deal with the situations where the data providing the assessment of attributes are not easily quantified. This paper proposes and utilizes the Dombi operations to develop certain aggregation operators: 2-tuple linguistic q-rung picture fuzzy Dombi weighted averaging operator and 2-tuple linguistic q-rung picture fuzzy Dombi weighted geometric operator. The presented methodology utilizes the aforementioned operators to handle and aggregate data in decision-making problems. Furthermore, an example to select the best option among the collection of considered projects is handled through the proposed method to substantiate the extended approach. Ultimately, the validation of the developed technique is conducted through a comparative analysis.

This study investigates the adoption of blockchain technology in small and medium-sized enterprises’ (SMEs) supply chains. Blockchain has been deemed a “Game Changer” for its potential benefits but also has several limitations due to its properties and SMEs’ limited resources. Small and medium-sized companies (SMEs) have more problems employing Industry 4.0 Technologies in practice than larger companies. So, the employment of BC in SMEs’ supply chains, as the economy's backbone, deserves a full investigation which has been addressed in this paper. We use the Technology–Organization–Environment framework to examine both positive and negative factors affecting blockchain adoption in SMEs. A Delphi method was used to refine a list of 42 factors down to the 22 most essential ones. These factors were then prioritized using the Fuzzy Logarithmic Additive Weights Methodology based on expert preferences. The findings indicate that while financial factors play a crucial role in adoption, transparency, and traceability are significant benefits that motivate organizations to adopt blockchain. The organizational dimension also has a substantial impact on decision-making. In this study, we conclude that SMEs must carefully consider their circumstances before adopting blockchain technology, given its infancy state. The barriers to adoption are just as significant as incentives, and SMEs must weigh the financial and organizational factors before making the decision. This study extends the literature by comprehensively looking at all possible factors influencing BC adoption by SMEs. Furthermore, we have highlighted valuable insights for policymakers and SME owners to make the best decision.

Traffic congestion is, nowadays, one of the most important highway problems. Highway tolls with booth operators are one of the causes of traffic congestion on highways, especially in rush hour periods, or during seasonal holiday travels. The value of driver waiting time (needed to stop and pay the toll) and the cost of the toll booth operators can reach up to about one-third of the revenue. In this paper we propose a novel methodology for continuous-time optimal control of highway tolls by predicting the optimal number of active modules (booths) in toll stations. The proposed methodology is based on a combination of recurrent neural networks, queuing theory, and metaheuristics. We utilized several recurrent neural network architectures for predicting the average intensity of vehicle arrivals. Moreover, the prediction error of the first recurrent neural network was modelled by another one in order to provide confidence estimates, additional regularization, and robustness. The predicted intensity of vehicle arrival rates was used as an input of the queuing model, whereas differential evolution was applied to minimize the total cost (waiting and service costs) by determining the optimal number of active modules on a highway toll in continuous time. The developed methodology was experimentally tested on real data from highway E70 in the Republic of Serbia. The obtained results showed significantly better performance compared to the currently used toll station opening pattern. The solutions obtained by solving a system of differential equations of the queuing model were also validated by a simulation procedure.

Because of the high competition among IT sectors, companies are planning to migrate to the cloud for effective growth and development. Driven by the importance of the cloud, new cloud vendors emerge each day to satisfy the demand of IT sectors. As a result, selection of an apt cloud vendor is critical. To this end, researchers have proposed different decision models, but these models do not effectively capture uncertainty during the decision -making process. To handle this issue, probabilistic linguistic information (PLI) is adopted in this paper, which associates occurrence probability to each term. Furthermore, weights of criteria are systematically determined using a deviation method, and cloud vendors are prioritized using a mathematical model under the PLI context. These methods are integrated to form the decision model, validated for its applicability using real case data from Cloud Armor. Finally, the advantages and weaknesses of the model are analyzed by using sensitivity analysis and comparison with extant models.

Artificial intelligence is steadily increasing its impact on everyday life. Therefore, the societal issues of artificial intelligence have become an important concern in the AI research. The presence of data that reflects human biases towards historically discriminated groups defined by sensitive features such as race and gender, results in machine learning models which discriminate against these groups. In order to tackle the impact of bias in data, researchers developed a variety of specialized machine learning algorithms which are able to satisfy different fairness constraints imposed on the model. Group fairness constraints do not fit standard machine learning formulations easily due to their non-differentiable nature. In this paper we developed a technique for learning a fair classifier by Monte Carlo policy gradient method which naturally deals with such non-differentiable constraints. Our methodology focuses on direct optimization of both group fairness metric and predictive performance of the model. In addition, we propose two different variance reduction techniques of gradient estimation. We compare our models to seven other related and state-of-the-art models and demonstrate that they are able to achieve better trade-off between accuracy and unfairness. To the best of our knowledge, this is the first fair classification algorithm which solves the issue of non-differentiable constraints by reinforcement learning techniques.

Early fault detection of engineering systems allows early warnings of anomalies and provides time to initiate proactive mitigation actions before the anomaly has developed to a problem that either requires extensive maintenance or affects the productivity of the system. In this paper, a new fault detection method using signal reconstruction based on Auto-Associative Extreme Learning Machines (AAELM) is proposed. AAELM are applied for fault detection on an artificially generated dataset to test the performance of the algorithm under controlled conditions and a real case study based on condition monitoring data from a combined-cycle power plant compressor. The performance of AAELM is compared to that of two other commonly used signal reconstruction methods: Auto-Associative Kernel Regression (AAKR) and Principal Component Analysis (PCA). The results from the two case studies demonstrate that AAELM achieve a smaller reconstruction error, shorter detection delay, lower spillover and a higher distinguishability compared to AAKR and PCA on the evaluated datasets. The obtained results are generalized to elaborate guidelines for industrial users for selecting suitable signal reconstruction algorithms based on their specific requirements and boundary conditions.

Early fault detection of engineering systems allows early warnings of anomalies and provides time to initiate proactive mitigation actions before the anomaly has developed to a problem that either requires extensive maintenance or affects the productivity of the system. In this paper, a new fault detection method using signal reconstruction based on Auto-Associative Extreme Learning Machines (AAELM) is proposed. AAELM are applied for fault detection on an artificially generated dataset to test the performance of the algorithm under controlled conditions and a real case study based on condition monitoring data from a combined-cycle power plant compressor. The performance of AAELM is compared to that of two other commonly used signal reconstruction methods: Auto-Associative Kernel Regression (AAKR) and Principal Component Analysis (PCA). The results from the two case studies demonstrate that AAELM achieve a smaller reconstruction error, shorter detection delay, lower spillover and a higher distinguishability compared to AAKR and PCA on the evaluated datasets. The obtained results are generalized to elaborate guidelines for industrial users for selecting suitable signal reconstruction algorithms based on their specific requirements and boundary conditions.

Possible ways of speeding-up the inference cycle of inference engines capable of reasoning with external data are studied in this paper. The design and implementation of such inference engines on personal computers and PC-based workstations are illustrated by a description of FLEXIE, a PC-based inference engine. FLEXIE has been developed as an integral part of FLEX, a prototype real-time expert system shell for personal computers. The FLEX shell is briefly reviewed in the paper, and FLEXIE is examined in detail. A fast pattern-matcher, based on the RETE algorithm, is built into FLEXIE. During the pattern-matching phase, FLEXIE communicates with FLEXEN, the FLEX interface to the environment, in order to include external information into the reasoning process. The design of the pattern-matching algorithm makes it suitable for hardware implementation, which can make the pattern-matching process run considerably faster. An architecture for such a hardware pattern-matcher is proposed, and initial simulation results are presented. Two possibilities for concurrent implementation of the inference cycle are used in the design of the hardware pattern-matcher. The contribution of parallelism to the pattern-matching speed-up is also discussed. Simulation shows the overall speed-up to be more than an order of magnitude.

In this paper, we present new Learning from Demonstration-based algorithm that generalizes and extracts relevant features of desired motion trajectories for differential drive mobile robots. The algorithm is tested through series of simulations and real world experiments in which desired task is demonstrated by the human teacher while teleoperating the mobile robot in the working environment. In the first step of the developed method, Gaussian Mixture Model (GMM) is built for incremental motions of the mobile robot between two consecutive poses. After this, the hidden Markov model is used to capture transitions between states (temporal variations of the data between clusters) which are missing from static GMM representation. Generalization of the motion is achieved by using the concept of keyframes, defined as points in which significant changes between GMM/HMM states occur. In the second step, the resulting GMM/HMM representation is used to generate optimal state sequences for each demonstration and to temporally align them, using 1D dynamic time warping, with respect to the one most consistent with the GMM/HMM model. This phase implies extraction of keyframes along all state sequences and projecting them into control space, in which controls are aligned in time as well. Finally, the generalized controls are obtained by averaging over all controls at the keyframes; simple piecewise cubic spline method is used for interpolation between generated control values. The main advantage of the developed algorithm is its ability to learn and generalize from all demonstrated examples which results in high quality reproductions of the motion. The proposed approach is verified both in simulated environment and using real mobile robot.

Second-generation expert systems emerged recently in order to overcome the shortcomings and limitations of first-generation expert systems. This paper presents a systematic and extended overview of second-generation expert-system features. In order to emphasize the improvements introduced by the second generation, a detailed discussion of the shortcomings and limitations of first-generation expert systems is also presented. Several examples of second-generation expert-system research, development and application are discussed. The examples provide a precise insight into the state of the art in the field.

REX, an expert systems development shell for robotics applications, is presented in the paper. The shell is intended for building expert systems in various domains of robotics, including design, planning, control, fault-detection, and navigation. REX's knowledge and data representation techniques include several standard techniques, like rules, frames, parameters, and variables, as well as certain techniques which are not commonly used by other expert systems building tools. These specific techniques include sensor data, status data, models and algorithms. The REX inference engine provides a means for reasoning with all of these knowledge and data types. The inference engine features both off-line and on-line modes of operation. REX has powerful interfaces for communicating both with the user and with other software modules in a complex application. Certain learning techniques are built into the shell. An example of applying REX for building an expert system in the domain of robot control is also presented. Finally, a discussion is provided on the relation between REX and second-generation expert systems.