1. Comparing algorithms for predictive data analytics : magistrsko deloGoran Kirov, 2024, master's thesis Abstract: The master’s degree thesis is composed of theoretical and practical parts. The theoretical part describes the basics of predictive data analytics and machine learning algorithms for classification such as Logistic Regression, Decision Tree, Random Forest, SVM, and KNN. We also describe different evaluation metrics such as Recall, Precision, Accuracy, F1 Score, Cohen’s Kappa, Hamming Loss, and Jaccard Index that are used to measure the performance of these algorithms. Additionally, we record the time taken for the training and prediction processes to provide insights into algorithm scalability.
The key part master’s thesis is the practical part that compares these algorithms with a self-implemented tool that shows results for different evaluation metrics on seven datasets. First, we describe the implementation of an application for testing where we measure evaluation metrics scores. We tested these algorithms on all seven datasets using Python libraries such as scikit-learn. Finally, w
Keywords: data analytics, machine learning, classification, evaluation metrics
Published in DKUM: 15.01.2025; Views: 0; Downloads: 39
 Full text (2,68 MB) |
2. Ergonomic evaluation of human–robot collaborative order picking : a combined laboratory and simulation studyMinqi Zhang, Jakob Marolt, Primož Bencak, Eric Grosse, Tone Lerher, 2024, published scientific conference contribution Abstract: Thanks to rapid technological developments in robotics, various automation technologies are being applied in warehouses today. Order picking, as a key process in warehouse operations, has drawn attention in academia and practice for decades. In addition to many studies dedicated to manual and fully automated order picking, efforts have also been made to study semi-automated warehouses in which humans and robots collaborate. However, these studies mostly focused on system efficiency and ignored ergonomic aspects. Order picking was confirmed as a labor-intensive process in an environment in which workers are at a high risk of developing health problems. Therefore, this study addresses the investigation of physical human working conditions in both manual and robot-assisted order picking systems via real-life laboratory experiments and simulation modeling. We used a motion capture system to assess human working postures when working with and without robot assistance. In addition, we estimated the daily workload by applying the energy expenditure concept. Using simulation experiments, we were able to extend the results to various practical scenarios with different design variables, for example warehouse layouts, order sizes, and human-robot team configuration. Our preliminary results reveal that human-robot collaboration can reduce human workload. Posture evaluation also shows a slight improvement.
Keywords: order picking, autonomous mobile robot, human factors, ergonomics, assisted order picking, hybrid order picking, simulation, performance evaluation
Published in DKUM: 21.11.2024; Views: 0; Downloads: 2
Full text (713,52 KB)
This document has many files! More... |
3. Analytical model and swapping policy assessment of a vertical lift module – buffer integrated storage systemJakob Marolt, Fabio Sgarbossa, Jesus A. Jimenez, Abhimanyu Sharotry, Tone Lerher, 2024, original scientific article Abstract: To meet the rising demands of global trade and e-commerce, efficient warehousing relies on inte-grated and cooperative material handling systems. This paper investigates the extension of VerticalLift Module (VLM) storage capability with a Buffer System and assesses the impact of this integra-tion on performance. We developed an analytical model to calculate the expected dual commandcycles, forming the basis for evaluating the VLM – Buffer integrated storage system’s performance.Our research emphasises minimising unnecessary swaps between the VLM and the Buffer Systemto enhance throughput performance. We introduce the Look Ahead Strategy (LAS) to minimiseinter-system swaps and develop a Binary Integer Program (BIP) to benchmark its performance. Theresults indicate that LAS performs on par with BIP, due to its ability to consider product popular-ity during the final selection of the outbound swapping tote. Through a comprehensive analysisof the analytical model with an empirical correction, utilising Pareto-based order sequences, theresults show deviations of less than 1% on average, affirming the analytical model’s accuracy. Ourresearch provides insights on using the VLM-Buffer integrated storage system, emphasising efficienttote swapping policies like LAS for enhanced warehouse operations, and allows managers to assesssystem performance through scenario-based analyses.
Keywords: automated warehouses, vertical lift module, toteswapping, analytical and numerical modelling, performance evaluation
Published in DKUM: 11.10.2024; Views: 0; Downloads: 1
Full text (3,56 MB)
This document has many files! More... |
4. Green sourcing: effects on supplier performance metrics in fast food restaurants in frontier marketsPaul Mukucha, Bongani E. Mushanyuri, Divaries Cosmas Jaravaza, 2021, original scientific article Abstract: Following the banning of expanded polystyrene packaging material, the fast food restaurant industry was serendipitously plunged into green sourcing initiatives for its packaging materials leading to heightened efforts on evaluating the suitability of various suppliers of green packaging material. This study therefore sought to assess the performance of selected suppliers on supplier performance metrics of total cost of ownership, quality of sourced material and delivery performance, in Zimbabwe’s restaurant industry. Methodology: Data was collected from senior procurement officers of 30 officially registered fast food restaurants in Zimbabwe. Each restaurant randomly selected 5 suppliers from their supplier lists. Each unit was asked to rate their 5 major suppliers of packaging material leading to an effective sample size of 150 suppliers. Structural Equation Modeling (SEM) was used to analyse data in AMOS. The five steps in SEM, that is: model specification, model identification, parameter estimation, model evaluation and modification were done. Findings: The findings in this study revealed that green sourcing leads to improved procurement performance in terms of total cost of ownership, quality of the materials sourced, and delivery performance. Implications for theory and practice: It was therefore recommended that practically the restaurant industry should explore various options of green packaging material ranging from recyclable and reusable plastic, paper, aluminium and steel material. Originality and value: Although the study was done in a single market, there is dearth in literature on green sourcing in frontier markets, of which the study provides empirical evidence from the restaurant industry in SubSaharan Africa.
Keywords: green sourcing, supplier evaluation, fast food restaurant, frontier markets, supplier performance metrics, green packaging material, total cost of ownership, quality of sourced material, delivery performance
Published in DKUM: 23.08.2024; Views: 80; Downloads: 7
Full text (1,33 MB)
This document has many files! More... |
5. Developing an integrated framework for supplier evaluation based on relevant attributes and performance measuresMohsen Cheshmberah, 2020, original scientific article Abstract: Supplier evaluation and selection is essential to any organization, and planning an effective and comprehensive approach to that end seems inevitable. Meanwhile, determining the requisite criteria for evaluating and selecting suppliers is probably one of the most important steps to be taken towards developing an evaluation and selection model in the organization. In this article, first a review of the literature on the criteria and the field of supplier evaluation and selection are provided. These criteria are then placed into proper categories. In order to formulate a supplier evaluation and selection framework for the manufacturing organization under study, the implemented categorization is applied where a list of fifteen attributes and performance criteria is created; where upon it is secured with the help of a designated panel (project team). These features are then screened using Lawshe's method the "social attribute" is removed from the list of fifteen. The remaining 14 other criteria are configured within the SEAP (Suppliers Evaluation based on Attributes and Performances) framework. The framework follows the objective of continually evaluating suppliers, both potential and actual ones through incorporating their performances into their qualification ratings. Based on the proposed framework, suppliers are evaluated on the basis of two types of criteria, - feature (attribute) and performance.
Keywords: framework, supplier evaluation, supplier selection, criteria, performance measurement, attributes, Lawshe's method
Published in DKUM: 22.08.2024; Views: 73; Downloads: 8
Full text (617,06 KB)
This document has many files! More... |
6. Multicriteria evaluation of intermodal (rail/road) freight transport corridorsMilan Janić, 2020, original scientific article Abstract: This paper deals with the multi-criteria evaluation of the intermodal (rail/road) freight corridors as competing transport alternatives. For such a purpose, the methodology has been developed consisting of two main components; i) the analytical models for estimating the indicators and measures of the corridors’ physical/spatial or infrastructural, technical/technological, operational, economic, social, and environmental performance; and ii) the MCDM (Multi-Criteria Decision Making) method using the above-mentioned indicators and measures of performance as the evaluation attributes/criteria in ranking and identifying the preferred among the several mutually competing freight transport alternative corridors. The proposed methodology has been applied to two Trans-European intermodal rail/road freight transport corridors. As such, it has shown to be of use, in addition to the researchers, also to the other potential DMs (Decision Maker(s)). These could be, for example, the freight shippers/receivers as the users of the already existing intermodal (rail/road) transport services, the transport and intermodal terminal operators and infrastructure providers, and the business and policy makers facing with the problems of allocating the usually limited investments in the social-economic feasible way to the corresponding infrastructure at the local, regional, national, and international scale.
Keywords: intermodal (rail/road) freight transport corridors, indicators and measures of performance, multicriteria evaluation
Published in DKUM: 22.08.2024; Views: 66; Downloads: 9
Full text (999,74 KB)
This document has many files! More... |
7. Implementation of the modern immersive learning model CPLMMatej Veber, Igor Pesek, Boris Aberšek, 2022, original scientific article Abstract: The digitalization of industrial processes is being driven forward worldwide. In parallel, the education system must also be transformed. Currently, education does not follow the opportunities and development of technologies. We can ask ourselves how we can integrate technologies into a traditional learning process or how we can adapt the learning process to these technologies. We focused on robotics education in secondary vocational education. The paper contains research results from a modern learning model that addresses student problem-solving using cyber-physical systems. We proposed a reference model for industrial robotics education in the 21st century based on an innovative cyber-physical didactic model (CPLM). We conducted procedure time measurements, questionnaire evaluations, and EEG evaluations. We could use VR to influence the improvement of spatial and visual memory. The more intense representation of the given information influences multiple centers in the brain and, thus, the formation of multiple neural connections. We can influence knowledge, learning more effectively with short-term training in the virtual world than with classical learning methods. From the studied resources, we can conclude that the newer approach to teaching robotics is not yet available in this form. The emerging modern technologies and the possibility of developing training in this area should be investigated further.
Keywords: VR technologies, educational robotics, education, innovative learning method development, evaluation
Published in DKUM: 08.07.2024; Views: 105; Downloads: 11
Full text (2,43 MB)
This document has many files! More... |
8. |
9. Microstructure, mechanical properties and fatigue behaviour of a new high-strength aluminium alloy AA 6086Franc Zupanič, Jernej Klemenc, Matej Steinacher, Srečko Glodež, 2023, original scientific article Abstract: This study presents the comprehensive experimental investigation of the microstructure, mechanical and fatigue properties of a new high-strength aluminium alloy AA 6086, which was developed from a commercial aluminium alloy AA 6082. The new alloy possesses a higher content of Si, and, it also contains Cu and Zr. The alloy was characterised in the as-cast condition after homogenisation, extrusion, and T6 heat treatment. Light microscopy, scanning and transmission electron microscopy with energy dispersive spectrometry were used to analyse the microstructure and the fractography of broken specimens. The quasi-static and fatigue tests were performed on the MTS Landmark 100 kN servo-hydraulic test machine, controlled with a mechanical extensometer with a 25 mm gauge length. The quasi-static strength of the analysed aluminium alloy AA 6086 was found to be significantly higher if compared to some other AA 6xxx alloys, while the ductility was kept almost the same. The experimental results of the comprehensive fatigue tests in a Low Cycle Fatigue (LCF) and High Cycle Fatigue (HCF) regime showed a good fatigue resistance, and represent a good basis for engineering design applications of the newly developed aluminium alloy AA 6086.
Keywords: aluminijeve zlitine, karakterizacija materiala, utrujanje, eksperimentalno testiranje, statistično ovrednotenje, Aluminium Alloy AA 6086, material characterisation, fatigue behaviour, experimental testing, statistical evaluation
Published in DKUM: 02.04.2024; Views: 281; Downloads: 23
 Link to full text
This document has many files! More... |
10. Robotic bin-picking : benchmarking robotics grippers with modified YCB object and model setTone Lerher, Primož Bencak, Luka Bizjak, Darko Hercog, Boris Jerman, 2023, published scientific conference contribution Abstract: Robotic bin-picking is increasingly important in the order-picking process in intralogistics. However, many aspects of the robotic bin-picking process (object detection, grasping, manipulation) still require the research community's attention. Established methods are used to test robotic grippers, enabling comparability of the research community's results. This study presents a modified YCB Robotic Gripper Assessment Protocol that was used to evaluate the performance of four robotic grippers (twofingered, vacuum, gecko, and soft gripper). During the testing, 45 objects from the modified YCB Object and Model Set from the packaging categories, tools, small objects, spherical objects, and deformable objects were grasped and manipulated. The results of the robotic gripper evaluation show that while some robotic grippers performed substantially well, there is an expressive grasp success variation over diverse objects. The results indicate that selecting the object grasp point next to selecting the most suitable robotic gripper is critical in successful object grasping. Therefore, we propose grasp point determination using mechanical software simulation with a model of a two-fingered gripper in an ADAMS/MATLAB cosimulation. Performing software simulations for this task can save time and give comparable results to real-world experiments.
Keywords: intralogistics, robotic bin-picking, YCB protocol, robotic gripper evaluation, mechanical software simulations, performance analysis
Published in DKUM: 21.03.2024; Views: 298; Downloads: 11
Full text (773,38 KB)
This document has many files! More... |
