Blockchain networks serve as a transparent and secure ledger storage solution, yet they remain vulnerable to attacks. There must be some mechanism to protect the blockchain network from attacks. Among various attacks, the Distributed Denial of Service (DDoS) attack is considered severe, which is challenging to detect accurately and reliably. Machine learning techniques are used to detect the attack, which requires exploring all global attack data in a single system, which is difficult in practice. This article proposes a distributed machine learning mechanism called Federated Machine Learning for detecting the presence of DDoS attacks. But in federated machine learning the model itself can be poisoned by the malicious collaborating node which is another problem that this article solves by storing the model in blockchain and by introducing a new reputation-based miner selection procedure. The proposed framework integrates the federation of machine learning within the blockchain network framework for detecting DDoS attacks. Under the integrated framework, miners are used to train the blocks and they also participate in the machine learning training. A dynamic reputation-based miner selection mechanism that can balance exploration and exploitation is proposed for optimal miner selection, which can ensure the high accuracy of the machine learning model and improve the security of blockchain from attacks like DDoS attacks and 51% attacks. The proposed framework is tested with Random Forest, Multilayer Perceptron, and Logistic Regression machine learning algorithms. The proposed mechanism achieved maximum accuracy of 99.1% using random forest model which is superior to the existing mechanism of detection of DDoS attacks.

The ongoing COVID-19 pandemic has caused a paradigm shift in all aspects of contemporary human life. Everyday activities such as shopping have shifted from traditional methods to the ever-more growing online variants, allowing for an increase in electronic commerce (e-commerce) industry. As more services become available online, consumers often rely on trusted services, which are often reflected on the web and mobile platforms they are presented on. In this paper, we study the factors for successful e-commerce platform design in the Western Balkans region using Fuzzy Analytical Hierarchy Process (FAHP) with triangular fuzzy numbers. After an extensive literature overview, interviews with representatives of top-ranking e-commerce companies in the region, and the analysis of experts’ opinions, we select a number of factors and sub-factors for prioritization, taking into account pre-pandemic factors, as well as the ones of the pandemic itself. We extend the FAHP model, which now consists of five (instead of three) points of view. Finally, we present and discuss the results in the form of tables and graphs, as well as an overall recommendation of what should be taken into account when designing an e-commerce platform. Our results rank service quality and security factors first and criteria such as multilingual support last.

The revolutionary merging of digital twin technology with deep learning methods brings a new age of creativity and efficiency in smart manufacturing. Improving production processes has been much easier with the advent of digital twins (DTs), which simulate real-world procedures in a computer simulation. Affordable and systematic DT implementation has been made possible by the increasing use of cyber-physical systems (CPS), the Internet of Things (IoT), big data analytics, and cloud computing in the manufacturing sector. This can improve manufacturing, track the health of machines and equipment, and offer product support and services. By incorporating deep learning algorithms, digital twins can acquire information, adapt, and make smart predictions using complicated information sources. These digital twins can optimize schedules for manufacturing, anticipate equipment faults, and independently analyze massive datasets with the help of dominant neural networks. With an eye on deep learning’s revolutionary effect on production efficiency, this chapter delves into how insights produced by deep learning help with controlling and monitoring in real-time, guaranteeing flexible and agile production processes, quality assurance, supply chain management, and asset performance. <!----> <!----> <!----> <!---->

In this paper, we discuss the multi-depot vehicle routing problem with regard to mobile beekeeping. Considering that beekeepers usually do not have their own land plots rarely rely on external transporters and often rely on external land plots, we present a fully functional platform to pair beekeepers, transporters, and land plot owners. Two algorithms for calculating optimal mobile beekeeping routes are developed in order to increase honey yield and minimize transport costs. We show that the first algorithm, based on the mathematical model, creates optimal routes in order to transport and distribute beehives by a distributor, while minimizing transport costs. In addition, we consider an option for the case when beekeepers themselves want to handle the transport and use a single external land plot. We show that, for this case, the application of the second algorithm creates a "good enough" solution. This iterative algorithm routes beekeepers to the closest land plot that can handle their beehives. Furthermore, we present a fully functional Web platform that applies these algorithms, enabling beekeepers, transporters and land owners to register and use the platform in Serbia.

In this paper, the performance of intensity-modulated with direct detection (IM/DD) free space optical (FSO) system using the on-off keying is observed. Besides the gamma-gamma atmospheric turbulence degradation, the pointing errors and path loss are included. In order to mitigate degradation effects, the selection combining (SC) receiver is applied. After theoretical derivation of the bit error rate expression, the numerical integration is accomplished and confirmed by Monte Carlo simulations. Also, the BER performance is improved by using the convolutional coding, which is demonstrated by Monte Carlo simulations. Numerical results are presented and effects of normalized jitter, normalized beam width, propagation distance and number of SC apertures are observed.