Journal of Applied Research on Industrial Engineering
Volume:10 Issue: 4, Autumn 2023

This study proposes a framework for the automated hyperparameter optimization of a bearing fault detection pipeline for Permanent Magnet Synchronous Motors (PMSMs) without the need for external sensors. An Automated Machine Learning (AutoML) pipeline search is performed through genetic optimization to reduce human-induced bias due to inappropriate parameterizations. A search space is defined, which includes general methods of signal processing and manipulation as well as methods tailored to the respective task and domain. The proposed framework is evaluated on the bearing fault detection use case under real-world conditions. Considerations on the generalization of the deployed fault detection pipelines are also considered. Likewise, attention was paid to experimental studies for evaluations of the robustness of the fault detection pipeline to variations of the motors working condition parameters between the training and test domain.

This paper deals with the modeling of the Food Bank (FB) network in the conditions of uncertainty in the demand of charities and the capacity of donating food. The importance of creating a FB network, along with providing quality food, led to consider the two objective functions of minimizing the costs of the total FB network and maximizing the minimum freshness of the food basket. The simultaneous optimization of the above two objective functions is aimed at making correct routing-inventory and allocation decisions. In this paper, food items in food baskets with high shelf-life and low shelf-life are considered. The results of solving the sample problems by combining the operators of two Genetic Algorithm (GA) and Salp Swarm Algorithm (SSA) showed that with the increase in the freshness of the food baskets, the costs of the FB network have increased. Also, the sensitivity analysis showed that the increase in uncertainty in the network leads to an increase in the cost of FB network and a decrease in the freshness of the food basket. The comparison of the results between the algorithms also showed that the efficiency of HGSSA is much higher than GA and SSA and the problem solving time by these methods is extremely lower. The use of HGSSA has increased the rate of achieving effective solutions by 14.06%.

In this paper, while taking into account the cooperative relationships between units, the problem of revenue allocation is considered as a coalitional game. In order for the allocation to be equitable, by relying on the concept of DEA efficiency, a new characteristic function is presented, and then using the concept of the Shapley value, which is a well-recognized concept in coalitional game theory, a unique solution is obtained for the revenue allocation problem. And finally, to evaluate the equitability of the performed revenue allocation, the Gini coefficient is utilized. A comparison of the Gini coefficient obtained for our method with those of some existing methods showed that our method is more equitable than the previous ones. This demonstrates how impactful the wise and accurate selection of the characteristic function is in the equitability of the results.

Sensitivity analysis in optimization problems is important for managers and decision maker to introduce different strategies. Data Envelopment Analysis (DEA) is a method based on mathematical programming to evaluate the efficiency of a set of Decision-Making Units (DMUs). Due to the importance of sensitivity analysis in an optimization problem, a development of DEA model called inverse model in DEA is presented. The purpose of this model is to analyze the sensitivity of some inputs or outputs to changes in some other inputs or outputs of the unit under evaluation, provided that the amount of efficiency remains constant or improves at the discretion of the manager. In this research, for the first time, we introduce the inverse model in DEA with network structure. In fact, we examine the extent to which the input parameters are likely to change based on the presuppositions of the problem, for the output changes that are applied as the manager desires. One of the key points of this research is that to make the modeling more consistent with reality, the leader-follower method was used in estimating the parameters in the network. In addition, the opinions of the system manager and the decision maker, who have full control over the system under their management, are included in this modeling to estimate the desired values. Another feature of this modeling is the consideration of uncontrollable factors in the inverse model in DEA with network structure. Finally, using a numerical example, the results obtained are analyzed based on the proposed model.

In a lean production environment, reduced setup times can lead to many benefits, including reduced lead times. Previous research has primarily relied on the SMED methodology and mathematical modeling to reduce setup times at machine centers – and both are very useful techniques. We use the Soft Systems Methodology, combined with the Seven Tools of Quality, to provide a structured, illustrative means for diagnosing production and quality issues. A baseline average setup time was established by which future setup times would be compared. The intervention included brainstorming sessions between management and the converting center work crew that disclosed many reasons for increased setup times, some of which were under management’s control. Our findings resulted in a 24% reduction in average setup times and a 62% reduction in the moving range at a bottleneck machine center in the corrugated box industry.

This study aimed to calculate the drift velocity and mobility of holes in organic semiconducting polymers by the Charge Extraction via Linearly Increasing Voltage (CELIV) technique to measure the charge carrier mobility. The charge carrier mobility is defined as carrier drift velocity v in each electric field E. This technique is complementary to Time of Flight (ToF) by providing us with an indication of the material’s properties when other methods are not applicable. Typically, Photo-CELIV is used to measure the charge carrier mobility in Organic Semiconductor (OSCs) due to large bandgap (2 eV) and few thermally generated carriers for extraction in the dark. The effect of the recombination mechanism is investigated on the carrier mobility in the organic layer. The calculation results showed that saturation of extracted charge is linearly proportional to carrier concentration at low concentrations, whereas at high density is saturated due to bimolecular carrier recombination. Langevin recombination mechanisms show that extracted demand saturates at j0, the capacitive displacement current step. Therefore, Δj/j0=1 at high light intensities, the saturation of extracted charge will start to decrease from its maximum value only when tdel is increased to be like tmax. In Langevin recombination, the bimolecular carrier lifetime is much faster than transit time at high carrier concentrations giving the saturation of extracted charge.

In this study, we explain the future potential scenarios of factors affecting stock price fluctuations in the Tehran Stock Exchange concerning the 2026 perspective. This research is applied, cross-sectional, and qualitative, and is implemented as a descriptive survey using the scenario planning approach. The statistical population is a collection of financial experts; We selected 15 of them as a sample using the judgmental/purposive and network (snowball) sampling methods. In the first step, we identified the key uncertainties of the factors affecting stock price fluctuations using the Fuzzy Delphi method, then by identifying the probable modes of each of the key uncertainties, three compatible scenarios were determined by Scenario Wizard software, and finally, the experts suggested strategies for these scenarios.

In the past years, East Azerbaijan province in Iran has always been at the top of the number of incidents in the country in the reports related to the annual analysis of incidents of domestic natural gas subscribers. Despite planning and spending at the expense of previous years, there has been no significant reduction in incident statistics. The purpose of this article is to investigate the root factors affecting the occurrence of incidents in domestic consumers of natural gas in East Azerbaijan province and to provide control and reduction strategies for incidents. To study the statistical analysis of natural gas-related incidents, the big data mining data approach of natural gas incidents in East Azerbaijan province during the years 2014 to 2020 besides Pareto analysis, root analysis, and Delphi have been used. The results of data and information analysis indicate that the most important technical factors affecting the bite are: lack of proper installation of the chimney, use of non-standard chimneys, leakage due to seams between the chimney parts, the presence of cracks, and virtual blockage of the chimney.

With the exponentially increasing volume of online data, searching and finding required information have become an extensive and time-consuming task. Recommender systems as a subclass of information retrieval and decision support systems by providing personalized suggestions helping users access what they need more efficiently. Among the different techniques for building a recommender system, Collaborative Filtering (CF) is the most popular and widespread approach. However, cold start and data sparsity are the fundamental challenges ahead of implementing an effective CF-based recommender. Recent successful developments in enhancing and implementing Deep Learning architectures motivated many studies to propose Deep Learning-based solutions for solving the recommenders' weak points. In this research, unlike the past similar works about using Deep Learning architectures in recommender systems that covered different techniques generally, we specifically provide a comprehensive review of Deep Learning-based CF recommender systems. This in-depth filtering gives a clear overview of the level of popularity, gaps, and ignored areas on leveraging Deep Learning techniques to build CF-based systems as the most influential recommenders.

Company is an organization that provides or produces products/services. Various types of companies and the complexity of the process make the company must be able to continue to grow and compete with competitors. To win the competition, companies must have a strategy to improve performance. TPM is part of the strategy implemented in the company. In Indonesia, not all companies apply TPM, automotive component painting companies apply and measure performance through PQCDSM as a whole. The result of TPM implementation is an increase in production performance which has an impact on reducing quality costs, increasing production, increasing the effectiveness of equipment use because the total of damaged equipment is less. TPM implementation through OEE production performance/engine efficiency increased by 68.7%.