International Journal of Information Technology & Decision Making

Manufacturing companies sometimes suffer from unexpected production disruptions/interruptions events (DIEs), affecting the production performance and cost. Since DIEs vary in type and cause, predicting the characteristics of their corresponding production downtimes is a challenging task. Although efforts have been devoted to forecast/prevent specific types of DIEs, such as machine-related events, it is still difficult to deal with the uncertainty caused by a combination of production DIEs of various types. Moreover, the absence of a realistic scenario generator incorporating DIEs has been a challenge in production scheduling under uncertainty. This study investigates the potential use of a hybrid data-driven approach in incorporating the uncertainties of a wide range of DIEs. In this approach, a random forest (RF) method and probability distributions are integrated to forecast the DIEs. The study was carried out based on the recorded DIEs in a Canadian company producing assembly parts for automotive industry. The performance of the proposed methodology for forecasting the production DIEs is evaluated by determining the predicted total downtime (TD) in percent of the expected processing time. The proposed hybrid model yields an overall accuracy of 92.82% in predicting the TD, compared to an overall accuracy of 75.64% when a single RF is used for prediction.

For decreasing service centers’ selection risks in emergency facility location selection, it is crucial to have selected candidate service centers within deeply detailed facility location selection model. To achieve this, a new approach developed in this article involves two stages. In the first stage, the fuzzy multi-attribute group decision making (MAGDM) model for evaluation of the selection of candidate service centers is created. For the aggregation of experts’ assessments of candidate service centers (with respect to attributes) aggregation operators’ approach is used. Experts’ assessments are presented in fuzzy terms with semantic form of triangular fuzzy numbers. For the deeply detailed facility location selection modeling and for the intellectual activity of experts in their evaluations, pairwise interactions between attributes of MAGDM model are considered in the construction of the second-order additive triangular fuzzy valued fuzzy measure (TFVFM). The associated triangular fuzzy probability averaging (As-TFPA) aggregation operators’ family is constructed with respect to TFVFM. Analytical properties of the As-TFPA operators are studied. The new operators are certain extensions of the well-known Choquet integral operator. The extensions, in contrast to the Choquet aggregation, consider all possible pair-wise interactions of the attributes by introducing associated fuzzy probabilities of a TFVFM. At the end of the first stage, a candidate service center’s selection index is defined as As-TFPA operator’s aggregation value on experts’ assessments with respect to attributes.

At the second stage, fuzzy multi-objective facility location set covering problem (MOFLSCP) is created for facility location selection optimal planning with new criteria: (1) maximization of candidate service centers selection index and classical two criteria, (2) minimization of the total cost needed to open service centers and (3) minimization of number of agents needed to operate the opened service centers.

For the constructed two-stage methodology a simulation example of emergency service facility location planning for a city is considered. The example gives the Pareto fronts obtained by As-TFPA operators, the Choquet integral-TFCA operator and well-known TOPSIS approach, for optimal selecting candidate sites for the servicing of demand points. The comparative analysis identifies that the differences in the Pareto solutions, obtained by using As-TFPA operators and TFCA operator or TOPSIS aggregation, are also caused by the fact that TFCA operator or TOPSIS approach considers the pair interaction indexes for only one consonant structure of attributes. While new As-TFPA aggregations provide all pairwise interactions for all consonant structures.

It is desirable to prevent traffic accidents by focusing on elderly people’s brain characteristics. The attention level during driving depends on the amount of information-processing resources. This study first aimed at investigating the effects of the change in attention levels on the electroencephalogram (EEG) waves during the graded working memory tasks for a traffic situation. With the increase in memory loads, reaction times were delayed in the elderly than the young group. The difficult tasks activated the induced $\delta$ and $\theta$ powers in the frontal midline area primarily in the elderly, during the selective task for a target. The elderly could retain the attention level because of the activated slow EEG responses, regardless of the task performance, although the increased $\delta$ wave may reflect drowsiness. Because the assistance system based on drivers’ brain signals can prevent car accidents, this study also aimed at evaluating the analytical method to automatically discriminate the different attentional tasks from the EEG signals. Compared with $k$-nearest neighbors and artificial neural networks, support vector machines more accurately classified attention levels (i.e., task difficulty) during working memory tasks reflecting a change in the induced $\delta$ and $\theta$ waves. This result can be related to a brain-computer interface system to judge the task difficulty during driving and alert a driver to danger. The experimental tasks for this study were limited because they involved simulations only in which participants recognized guided boards and removed irrelevant information. Real-time judgments should be investigated using EEG data to improve systems that can alert drivers to oncoming dangers.

The positioning of roadside units (RSUs) in a vehicle-to-infrastructure (V2I) communication system may have an impact on network performance. Optimal RSU positioning is required to reduce cost and maintain the quality of service. However, RSU positioning is considered a difficult task because numerous criteria, such as the cost of RSUs, the intersection area and communication strength, affect the positioning process and must be considered. Furthermore, the conflict and trade-off amongst these criteria and the significance of each criterion are reflected on the RSU positioning process. Thus, this work proposes a new RSU positioning framework based on multicriteria decision-making (MCDM) in the context of the V2I communication system. Three stages are completed for this purpose. First, a real-time V2I hardware is developed to collect data. The developed hardware consists of multiple mobile nodes (i.e., cars with sending–receiving hardware devices) and physical RSUs. The RSUs and the devices in the cars are connected via the nRF24L01$+$PA/LNA transceiver module with Arduino Uno. Second, seven testing scenarios are identified toward acquiring the required data upon the connection of the V2I devices. Moreover, three evaluation attributes (i.e., number of packet losses [PKL], cost and ratio of intersection area [RIA]) are used to evaluate each scenario. A decision matrix is constructed on the basis of the crossover between ‘RSU positioning scenarios’ and ‘multi-evaluation attributes (i.e., PKL, cost and RIA)’. Third, the RSU positioning scenarios are ranked using MCDM techniques, such as the integrated analytic hierarchy process (AHP), entropy and group Vlsekriterijumska Optimizacija I Kompromisno Resenje (VIKOR). Furthermore, the Borda voting approach is used to aggregate multiple individual rankings into a uniform and final rank. Results indicate the following: (1) integrating AHP, entropy and VIKOR is effective for solving RSU positioning problems; (2) the VIKOR ranking results for individuals vary; (3) the rank of scenarios obtained from the group-VIKOR-based Borda voting context shows that the second scenario, which consists of four RSUs distributed along the street with a maximum distance of 200m between them and 2-m high antennas, is the best in terms of optimally placing the RSUs; and (4) significant differences are observed amongst the scores of the groups, indicating that the ranking results are valid.

In the current world, dealing with some problems with interval data is inevitable. In this case, the methods applied for real data could not be employed. To solve these problems, the modified version of previous methods or new methods should be presented. In this paper, the two-stage ranking method that already has been proposed by the authors is modified to solve the mentioned problems. In each stage, two optimistic and pessimistic attitudes are considered and their corresponding models are presented. Then, an appropriate algorithm for classifying the units based on their obtained interval efficiency is proposed. To demonstrate the applicability of the proposed method, 30 branches of the social security insurance organization in Iran are classified. Also, the validity and consistency of the proposed method are confirmed. The main contributions of this paper are as follows: Decision-making units (DMUs) are ranked with interval inputs and outputs. Inefficiency of the first projection (obtained in the first stage) is applied in the unit rank score. All units are classified in separate classes and all units of each class are ranked. Pareto-efficient projections (practical benchmarks) are obtained for all inefficient units. The proposed model is always feasible and unit invariant.

The analysis of the opinions and experiences of tourists is a key issue in tourist promotion. More precisely, forecasting whether a tourist will or will not recommend a given destination, based on his/her profile, is of utmost importance in order to optimize management actions. According to this idea, this research proposes the application of cutting-edge machine learning techniques in order to predict tourist recommendation of rural destinations. More precisely, classifiers based on supervised learning (namely Support Vector Machine, Decision Trees, and $k$-Nearest Neighbor) are applied to survey data collected in the province of Burgos (Spain). Available data suffer from a common problem in real-life datasets (data unbalance) as there are very few negative recommendations. In order to address such problem, that penalizes learning, data balancing techniques have been also applied. The satisfactory results validate the proposed application, being a useful tool for tourist managers.

To ensure a safe production environment for blue-collar employees, physical ergonomic risk factors (PERFs) such as noise, the vibration must be considered in every workplace. However, PERFs require both theoretical and practical information to figure out their characteristics and negative health effects on employees. This paper proposes an easy-to-follow sustainable ergonomic guide that includes the theoretical and practical information about the calculation details of PERFs, a decision-making (DM) tool to evaluate ergonomic risk factors, and a formula to calculate the total ergonomic risk score of a company. In the DM process, the Pythagorean fuzzy extensions of the Analytical Hierarchy Process were utilized to evaluate PERFs with respect to the social, technical, environmental, and economic criteria. To show the applicability of this guide, a real-world application was conducted in a company located in Ankara. According to the results, technical criteria were determined as the most critical ($w=0.459$), followed by the social, i.e., health-related factor ($w=0.320$). The total risk score of the company was calculated using the measurement of the PERFs and the coefficients obtained from the DM process. A comprehensive sensitivity analysis was conducted to validate the robustness of the guide. The main contribution of this paper is that a formula that can be utilized to calculate the total ergonomic risk score of any type of company was developed, based on the fuzzy DM technique and Threshold Limit Values of risk parameters.

In this study, it is aimed to contribute to the development of companies in order to reach the desired values utilizing the cluster features of the suppliers in a sustainable supplier evaluation process and to guide the supplier in this regard. The clustering features acquired at the end of the study will guide the determination of sustainable supplier criteria and potential new suppliers according to the sector and business line in future studies. The first step of the three-stage model is to apply the best–worst method to determine the weights of the criteria. At the second stage, suppliers are clustered together by the k-means algorithm and silhouette index method. At the third stage, suppliers are divided into categories, and suggestion policies for each supplier category are included. Then, sensitivity analysis has been performed to analyze the effects of variations of criteria weights on the result of the proposed model. Finally, to express the integrated method, a real-life sustainable supplier selection problem is solved and suppliers are categorized according to their similarities. Following evaluation, it was seen that the environmental management system and waste management criteria were the most critical criteria in the categorization of suppliers. When its reasons are investigated, sectoral characteristics come into prominence. The results indicate that the proposed model fits well as a multicriteria decision-making tool for sustainable supplier selection and evaluation decisions. Finally, it is predicted that this model will evaluate new suppliers or new products/services to be procured and selected from the set of features.