Federated learning (FL) has considerably emerged as a promising solution to enhance user privacy and data security by enabling collaboratively multi-party model learning without exchanging confidential data. Neverthel...
详细信息
Multi-Sequence Alignment (MSA) is considered an NP problem in bioinformatics. Compared to traditional techniques, nature-inspired techniques produce accurate results. In this article, Improved Chemical Reaction Optimi...
详细信息
The precise control of an exoskeleton, especially for pediatric subjects, is a challenging task due to the presence of parametric perturbations and external disturbances (PPED). The sliding mode control is a well-know...
详细信息
Modern deep Reinforcement Learning (RL) techniques are quite good at choosing the best possible rules to maximise rewards. By using rich visual information for policy selection, this method combined with Deep Learning...
详细信息
ISBN:
(数字)9798350394474
ISBN:
(纸本)9798350394481
Modern deep Reinforcement Learning (RL) techniques are quite good at choosing the best possible rules to maximise rewards. By using rich visual information for policy selection, this method combined with Deep Learning methodologies shows promise for difficult tasks. In this work, we have used Double DQN and also three variants of Duelling DQN to implement the ice hockey environment. The Duelling DQN outperformed the Double DQN and gave better results.
In our swiftly digitizing world, the importance of low-earth-orbit satellite (LEO SAT) communication is escalating. Advancements in 5G, IoT, and AI underscore the need for a robust communication network. Research reve...
In our swiftly digitizing world, the importance of low-earth-orbit satellite (LEO SAT) communication is escalating. Advancements in 5G, IoT, and AI underscore the need for a robust communication network. Research reveals that nearly 40% of the global population faces Internet service challenges due to limited high-performance access by 2022. This technology not only addresses coverage limitations but also mitigates latency issues. Multinational corporations like SpaceX and Amazon are already exploring its potential. As we approach the 6G era, users’ computational demands surge, prompting satellite edge computing with onboard processing. Our paper proposes a multi-hop code offloading model using LEO SAT, reducing energy consumption, and introduces a dynamic optimization algorithm for LEO SAT - mobile edge computing. The research aims to enhance satellite-based communication networks, with a focus on optimizing satellite edge computing efficiency.
Edge computing processes and stores data on user devices located at the edge of the Internet to support latency-sensitive services such as autonomous vehicles, drones, and so on. This adopts container technology to qu...
详细信息
ISBN:
(纸本)9781665423830
Edge computing processes and stores data on user devices located at the edge of the Internet to support latency-sensitive services such as autonomous vehicles, drones, and so on. This adopts container technology to quickly initialize and deploy applications. Containers interact with each other through Plugins implemented based on Container Network Interface (CNI). In particular, the network latency is different depending on the plugin. In this paper, we study plugins to provide low-latency edge services and evaluate the effect on data processing performance when multiple plugins are used at the same time. The SR-IOV plugins showed twice throughput than the kernel-based virtual network interface. Also, it had more than 50% bandwidth regardless of the addition of other virtual network interfaces on a single connection.
With the IoT connecting the world, finding parking lots is much easier with smart parking solutions. Some of the parking areas in the world use at least some kind of smart parking solution these days. There are existi...
详细信息
Artificial intelligence (AI) is becoming more active than ever in everyday life and steadily being incorporated to healthcare. AI, with its seemingly limitless power, affirms a promising future to a revolutionized hea...
详细信息
An enhanced variant of the Grey Wolf Optimisation (GWO) technique is presented in this paper, addressing a number of issues, including reduced population diversity, slow convergence and vulnerability to local optima. ...
详细信息
ISBN:
(数字)9798331522100
ISBN:
(纸本)9798331522117
An enhanced variant of the Grey Wolf Optimisation (GWO) technique is presented in this paper, addressing a number of issues, including reduced population diversity, slow convergence and vulnerability to local optima. The proposed variant incorporates innovative techniques, including Opposition Learning, Refraction Learning (RL), Differential Evolution Mutation and a non-linear time-varying hyperparameter tuning mechanism. Combining these strategies greatly enhances the standard GWO’s performance, leading to a more streamlined and effective search procedure. This enhancement leads to better performance in multi-modal landscapes, facilitates escaping local optima and allows the search space to be explored more extensively. Extensive testing across 23 benchmark functions demonstrates that the new variant outperforms the original GWO as well as other contemporary optimization methods. The results indicate substantial improvements in convergence speed, global search capability and solution accuracy, demonstrating how effective this method can be as an optimisation tool for solving challenging real-world issues.
暂无评论