There is a large potential for automation and optimization in radio access networks (RANs) using a data-driven approach to efficiently handle the increase in complexity due to the steep growth in traffic and new technologies introduced with the development of 5G. Reinforcement learning (RL) has natural applications in RAN control loops such as link adaptation, interference management and power control at different timescales commonly occurring in the RAN context. Elevating the status of data-driven solutions in RAN and building a new, scalable, distributed and data-friendly RAN architecture will be needed to competitively tackle the challenges of coming 5G networks. In this work, we propose a systematic, efficient and robust methodology for applying RL on different control problems. Firstly, the proposed methodology is evaluated using a well-known control problem. Then, it is adapted to a real-world RAN scenario. Extensive simulation results are provided to show the effectiveness and potential of the proposed approach. The methodology was successfully created but results on a RAN-simulator were not mature / Det finns en stor potential automatisering och optimering inom radionätverk (RAN, radio access network) genom att använda datadrivna lösningar för att på ett effektivt sätt hantera den ökade komplexiteten på grund av trafikökningar and nya teknologier som introducerats i samband med 5G. Förstärkningsinlärning (RL, reinforcement learning) har naturliga kopplingar till reglerproblem i olika tidsskalor, såsom länkanpassning, interferenshantering och kraftkontroll, vilket är vanligt förekommande i radionätverk. Att förhöja statusen på datadrivna lösningar i radionätverk kommer att vara nödvändigt för att hantera utmaningarna som uppkommer med framtida 5G nätverk. I detta arbete föreslås vi en syetematisk metodologi för att applicera RL på ett reglerproblem. I första hand används den föreslagna metodologin på ett välkänt reglerporblem. Senare anpassas metodologin till ett äkta RAN-scenario. Arbetet inkluderar utförliga resultat från simuleringar för att visa effektiviteten och potentialen hos den föreslagna metoden. En lyckad metodologi skapades men resultaten på RAN-simulatorn saknade mognad.

In today’s industry, the competitiveness between companies is continuously increasing and thus it is important to continue utilizing new technologies to make the operation more efficient and it is vital to strive towards continuous improvement. At this moment it is getting more and more vital to utilize the 4th industrial revolution in the production industry but to utilize this new wave of technology it is crucial to understand what in-efficiencies exist in a production plant and how to work with Lean production as well as implementing the new technologies such as Artificial intelligence and machine learning. These technologies are relatively new to the industrial industry and naturally, with new technologies, there are challenges, benefits but also risks involving new technology implementation. This must be done for companies to stay competitive within the sector even though it is not easy to implement, hence this thesis focuses on realizing these challenges and risks as well as understanding the benefits that can be gained when implementing artificial intelligence in production. When working with this kind of implementation it is important to consider all aspects that are affected by the change, this includes humans as well as the production itself in addition to that also the environment. It is important to understand that when working with this type of implementation it is important to realise what the lean wastes are and by understanding this it will be easier to know what AI can do to minimize or eliminate these wastes and thus making the operation more efficient. This study focuses on the challenges that this type of implementation might have as well as what benefits and risks that AI aided scheduling will have when its implemented. In this study, the findings are connected to a case study made at a focal company in the wood industry as well as an extensive literature study within the field. This thesis provided information in the analysis chapter about how these subjects are linked to the industry and how it's linked to the four main fields of this study. By combining the literature search with the findings from the case study at the focal company, vital information could be gathered and analysed. The areas of this study that was analysed and later discussed when answering the three research questions. The result of this thesis was later used as a base for suggestions for possible future implementations within the field. In addition to that, this study also acts as a framework for how AI implementations can benefit a company’s operation within the industrial sector

Business Intelligence (BI) handlar huvudsakligen om att samla in, analysera och konvertera data till värdefull information som sedan används av beslutstagare för att vidareutveckla och optimera verksamheten (Negash, 2004). Utvinning av positiva effekter till följd av BI-implementering är dock inte något som organisationer kan eller bör ta för givet. Tidigare forskning påvisar att ett stort antal organisationer upplever svårigheter att utvinna nytta ur BI-initiativ och att satsningar på BI därmed ofta betraktas som ett misslyckande (Chenoweth, Corral & Demirkan, 2006). Baserat på tidigare forskning bedöms frekvensen av misslyckade BI-projekt vidare ligga någonstans kring 50-80% av alla BI-satsningar (Meehaan, 2011; Legodi & Barry, 2010). I dagsläget finns begränsad forskning om faktorer som påverkar BI-implementeringars framgång på grund av att utvecklingen av BI främst drivits av IT- industrin och dess leverantörer (Yeoh & Koronios, 2010). Detta pekar enligt oss på att det finns en diskrepans mellan forskning och praktik. Vår förhoppning är således att vår studie kan fylla denna kunskapslucka genom att presentera nya empiriska insikter kopplat till framgångsfaktorer för implementering av BI hos svenska organisationer, med hjälp av en vetenskaplig förankring. / Business Intelligence (BI) mainly concerns collecting, analyzing and converting data into valuable information that is then used by decision makers to further develop and optimize the business (Negash, 2004). However, the extraction of positive effects as a result of BI implementation is not something organizations can or should take for granted. Previous research shows that a large number of organizations find it difficult to derive benefits from BI initiatives and that investments in BI are often regarded as failures (Chenoweth, Corral & Demirkan, 2006). Furthermore, the frequency of failed BI projects is assessed to be somewhere around 50-80% of all BI initiatives (Meehaan, 2011; Legodi & Barry, 2010). There is currently limited research on factors that affect the success of BI implementations due to the fact that the development of BI has been mainly driven by the IT industry and its suppliers (Yeoh & Koronios, 2010). This leads us to believe that there is a discrepancy between research and practice. Our hope is therefore that we through our study will be able to fill this knowledge gap by presenting new empirical insights linked to success factors for BI implementation in Swedish organizations, using a scientific foundation.

Detecting failures in tricone drill bits is crucial in the mining industry due to their potential consequences, including operational losses, safety hazards, and delays in drilling operations. Timely identification of failures allows for proactive maintenance and necessary measures to ensure smooth drilling processes and minimize associated risks. Accurate failure detection helps mining operations avoid financial losses by preventing unplanned breakdowns, costly repairs, and extended downtime. Moreover, it optimizes operational efficiency by enabling timely maintenance interventions, extending the lifespan of drill bits, and minimizing disruptions. Failure detection also plays a critical role in ensuring the safety of personnel and equipment involved in drilling operations. Traditionally, failure detection in tricone drill bits relies on manual inspection, which can be time-consuming and labor-intensive. Incorporating artificial intelligence-based approaches can significantly enhance efficiency and accuracy. This thesis uses machine learning methods for failure inference in tricone drill bits. A classic Convolutional Neural Network (CNN) classification method was initially explored, but its performance was insufficient due to the small dataset size and imbalanced data. The problem was reformulated as an object detection task to overcome these limitations, and a post-processing operation was incorporated. Data augmentation techniques enhanced the training and evaluation datasets, improving failure detection accuracy. Experimental results highlighted the need for revising the initial CNN classification method, given the limitations of the small and imbalanced dataset. However, You Only Look Once (YOLO) algorithms such as YOLOv5 and YOLOv8 models exhibited improved performance. The post-processing operation further refined the results obtained from the YOLO algorithm, specifically YOLOv5 and YOLOv8 models. While YOLO provides bounding box coordinates and class labels, the post-processing step enhanced drill bit failure detection through various techniques such as confidence thresholding, etc. By effectively leveraging the YOLO-based models and incorporating post-processing, this research advances failure detection in tricone drill bits. These intelligent methods enable more precise and efficient detection, preventing operational losses and optimizing maintenance processes. The findings underscore the potential of machine learning techniques in the mining industry, particularly in mechanical drilling, driving progress and enhancing overall operational efficiency