检索结果-内蒙古大学图书馆

dynamic routing via Reinforcement Learning for Network Traffic Optimization

学校读者我要写书评

暂无评论

Informatica (Slovenia)

Informatica (Slovenia) 2025年第8期49卷 1-14页

作者： Ma, Jian Zhu, Chaoyong Fu, Yuntao Zhang, Haichao Xiong, Wenjing State Grid Yingda CO. LTD. Beijing100005 China State Grid Yingda International Holdings CO. LTD. Beijing100005 China Information Technology CO. LTD. Beijing100077 China

With the rapid development of the Internet, network traffic has shown explosive growth, which puts forward higher requirements for the network routing system. Traditional static routing methods are no longer able to meet the needs of today's complex and ever-changing network environment, as they cannot be flexibly adjusted according to real-time network conditions. In order to address this challenge, this paper proposes an innovative dynamic routing method. This method is based on reinforcement learning, especially Q-learning algorithm, which realizes the dynamic adjustment of routing decisions through continuous learning and adaptation to changes in the network environment. Our goal is to minimize root mean square error (RMSE) to improve routing accuracy, while at the same time improving load balancing efficiency to ensure that network resources are fully utilized. In order to verify the effectiveness of this method, we conducted detailed simulation experiments. Experimental results show that compared with the baseline method, our dynamic routing method significantly improves the throughput of the network, which increases by 30%, effectively reduces the delay, and reduces 25%. These positive results not only prove the effectiveness of our method in network traffic optimization, but also provide new ideas for the development of network routing system in the future. © 2025 Slovene Society Informatika. All rights reserved.

关键词： dynamic routing algorithms

EADR: Entropy Adjusted dynamic routing Capsule Networks 59

学校读者我要写书评

暂无评论

EADR: Entropy Adjusted Dynamic Routing Capsule Networks

59th Annual Conference on Information Sciences and Systems, CISS 2025

作者： Babaei, Reza Pham, Huong Cheng, Samuel Zhao, Shangqing The University of Oklahoma School of Electrical and Computer Engineering Norman United States The University of Oklahoma School of Computer Science Norman United States

ISBN: (纸本)9798331513269

Capsule networks offer a useful approach for modeling part-whole hierarchies in visual data, yet they remain limited in effectively handling uncertainty in part-object relationships. This work introduces an entropy-adjusted dynamic routing (EADR) algorithm that leverages information-theoretic principles to enhance both the performance and interpretability of capsule networks. By incorporating an entropy-based regularization term into the final iteration of the dynamic routing process, our approach refines routing decisions, reducing reliance on uncertain capsule connections. Experimental evaluations on the CIFAR-10 dataset demonstrate that our method achieves a mean accuracy of 85.73%, surpassing the baseline accuracy of 84.22% achieved by standard dynamic routing. Comparative analysis with recent entropy-based routing methods highlights our approach's balance of computational efficiency and routing flexibility, achieved without much additional model complexity. These results suggest that EADR routing can be a useful tool for enhancing both the interpretability and efficacy of capsule networks in uncertain environments. © 2025 IEEE.

关键词： dynamic routing algorithms

DRAMA: A dynamic Packet routing Algorithm using Multi-Agent Reinforcement Learning with Emergent Communication

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhang, Wang Liu, Chenguang Pi, Yue Zhang, Yong Huang, Hairong Rao, Baoquan Ding, Yulong Yang, Shuanghua Jiang, Jie Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China Huawei Technologies Co. Ltd. China School of Engineering University of Warwick CoventryCV4 7AL United Kingdom Department of Computer Science University of Reading United Kingdom Beijing China

The continuous expansion of network data presents a pressing challenge for conventional routing algorithms. As the demand escalates, these algorithms are struggling to cope. In this context, reinforcement learning (RL) and multi-agent reinforcement learning (MARL) algorithms emerge as promising solutions. However, the urgency and importance of the problem are clear, as existing RL/MARL-based routing approaches lack effective communication in run time among routers, making it challenging for individual routers to adapt to complex and dynamic changing networks. More importantly, they lack the ability to deal with dynamically changing network topology, especially the addition of the router, due to the non-scalability of their neural networks. This paper proposes a novel dynamic routing algorithm, DRAMA, incorporating emergent communication in multi-agent reinforcement learning. Through emergent communication, routers could learn how to communicate effectively to maximize the optimization objectives. Meanwhile, a new Q-network and graph-based emergent communication are introduced to dynamically adapt to the changing network topology without retraining while ensuring robust performance. Experimental results showcase DRAMA’s superior performance over the traditional routing algorithm and other RL/MARL-based algorithms, achieving a higher delivery rate and lower latency in diverse network scenarios, including dynamic network load and topology. Moreover, an ablation experiment validates the prospect of emergent communication in facilitating packet routing. © 2025, CC BY-NC-SA.

关键词： dynamic routing algorithms

The Convergence of dynamic routing between Capsules

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Ye, Daoyuan Li, Juntao Shen, Yiting School of Informatics Beijing Wuzi University Beijing China

Capsule networks(CapsNet) are recently proposed neural network models with new processing layers, specifically for entity representation and discovery of images. It is well known that CapsNet have some advantages over traditional neural networks, especially in generalization capability. At the same time, some studies report negative experimental results. The causes of this contradiction have not been thoroughly analyzed. The preliminary experimental results show that the behavior of routing algorithms does not always produce good results as expected, and in most cases, different routing algorithms do not change the classification results, but simply polarize the link strength, especially when they continue to repeat without stopping. To realize the true potential of the CapsNet, deep mathematical analysis of the routing algorithms is crucial. In CapsNet, each capsule uses the length of the pose vector to represent the probability of the presence of the entity represented by it. An irregular nonlinear function is required in the dynamic routing algorithm to keep the vector length less than 1. Based on this, Hinton et al. didn’t believe the existence of any sensible objective function that is minimized by the dynamic routing algorithm. However, in this paper, we will give the objective function that is minimized by the dynamic routing algorithm, which is a concave function. The dynamic routing algorithm can be regarded as nonlinear gradient method to solving an optimization algorithm under linear constraints, and its convergence can be strictly proved mathematically. Furthermore, the mathematically rigorous proof of the convergence is given for this class of iterative routing procedures. We analyze the relation between the objective function and the constraints solved by the dynamic routing algorithm in detail, and perform the corresponding routing experiment to analyze the effect of our convergence proof. © 2025, CC BY-SA.

关键词： dynamic routing algorithms

LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned dynamic routing

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Li, Yang

The rapid advancement in large language models (LLMs) has brought forth a diverse range of models with varying capabilities that excel in different tasks and domains. However, selecting the optimal LLM for user queries often involves a challenging trade-off between accuracy and cost, a problem exacerbated by the diverse demands of individual queries. In this work, we present a novel framework that formulates the LLM selection process as a multi-armed bandit problem, enabling dynamic and intelligent routing of queries to the most appropriate model. Our approach incorporates a preference-conditioned dynamic routing mechanism, allowing users to specify their preferences at inference time, thereby offering a customizable balance between performance and cost. Additionally, our selection policy is designed to generalize to unseen LLMs, ensuring adaptability to new models as they emerge. Experimental results demonstrate that our method achieves significant improvements in both accuracy and cost-effectiveness across various LLM platforms, showcasing the potential of our framework to adaptively optimize LLM selection in real-world scenarios. © 2025, CC BY-NC-ND.

关键词： dynamic routing algorithms

High-Performance Computing for Distributed Route Computation in Traffic Flow Models 2

学校读者我要写书评

暂无评论

High-Performance Computing for Distributed Route Computation...

2nd EuroHPC user day, EuroHPC 2024

作者： Silva, Paulo Smolková, Pavlína Michailidu, Sofia Beránek, Jakub Macháček, Roman Slaninová, Kateřina Martinovič, Jan Cmar, Radim IT4Innovations VSB - Technical University of Ostrava Ostrava Czech Republic Sygic A.s. Twin City C Mlynské Nivy 16 Bratislava Slovakia

In the dynamic landscape of smart cities and traffic management, it is necessary to further explore the synergistic potential of realtime traffic data and high-performance computing to optimise traffic flow through dynamic re-routing strategies. High-performance computing plays an essential role in achieving effective traffic flow optimisation. Our research builds upon existing studies highlighting the positive correlation between the integration of live traffic updates and route optimisation. The methodology involves simulations with our Ruth traffic simulator, where vehicles dynamically adjust routes based on up to date traffic information available to them at different levels. Scalability tests are conducted with varying numbers of CPUs and nodes to assess the simulator's capacity to scale. The results showcase the impact of live traffic data on both driving time and average speed, emphasising the adaptability of our approach for broader applications. In conclusion, our work not only advances the understanding of real-time traffic optimisation but also underscores the critical role of high-performance computing in achieving scalable solutions. The findings present practical implications for the implementation of dynamic re-routing strategies in transportation systems, paving the way for future research and real-world applications on smart cities. © 2025 The Authors. Published by Elsevier B.V.

关键词： dynamic routing algorithms

DialectMoE: An End-to-End Multi-dialect Speech Recognition Model with Mixture-of-Experts 23rd

学校读者我要写书评

暂无评论

DialectMoE: An End-to-End Multi-dialect Speech Recognition M...

23rd China National Conference on Computational Linguistics, CCL 2024

作者： Zhou, Jie Gao, Shengxiang Yu, Zhengtao Dong, Ling Wang, Wenjun Faculty of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming650500 China Yunnan Key Laboratory of Artificial Intelligence Kunming University of Science and Technology Yunnan Kunming650500 China

ISBN: (纸本)9789819783663

Dialect speech recognition has always been one of the challenges in Automatic Speech Recognition (ASR) systems. While lots of ASR systems perform well in Mandarin, their performance significantly drops when handling dialect speech. This is mainly due to the obvious differences between dialects and Mandarin in pronunciation and the data scarcity of dialect speech. In this paper, we propose DialectMoE, a Chinese multi-dialects speech recognition model based on Mixture-of-Experts (MoE) in a low-resource conditions. Specifically, DialectMoE assigns input sequences to a set of experts using a dynamic routing algorithm, with each expert potentially trained for a specific dialect. Subsequently, the outputs of these experts are combined to derive the final output. Due to the similarities among dialects, distinct experts may offer assistance in recognizing other dialects as well. Experimental results on the Datatang dialect public dataset show that, compared with the baseline model, DialectMoE reduces Character Error Rate (CER) for Sichuan, Yunnan, Hubei and Henan dialects by 23.6%, 32.6%, 39.2% and 35.09% respectively. The proposed DialectMoE model demonstrates outstanding performance in multi-dialects speech recognition. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： dynamic routing algorithms

dynamic Directional routing of Freight in the Physical Internet

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Shaikh, Sahrish Jaleel Muthukrishnan, Praveen Lai, Yijun Montreuil, Benoit H. Milton Stewart School of Industrial & Systems Engineering Georgia Institute of Technology Atlanta United States

The Physical Internet (PI) envisions an interconnected, modular, and dynamically managed logistics system inspired by the Digital Internet. It enables open-access networks where shipments traverse a hyperconnected system of hubs, adjusting routes based on real-time conditions. A key challenge in scalable and adaptive freight movement is routing—determining how shipments navigate the network to balance service levels, consolidation, and adaptability. This paper introduces directional routing, a dynamic approach that flexibly adjusts shipment paths, optimizing efficiency and consolidation using real-time logistics data. Unlike shortest-path routing, which follows fixed routes, directional routing dynamically selects feasible next-hop hubs based on network conditions, consolidation opportunities, and service level constraints. It consists of two phases: (1) area discovery, which identifies candidate hubs, and (2) node selection, which determines the next hub based on real-time parameters. This paper advances the area discovery phase by introducing a Reduced Search Space Breadth-First Search (RSS-BFS) method to systematically identify feasible routing areas while balancing service levels and consolidation. The proposed approach enhances network fluidity, scalability, and adaptability in PI-based logistics, advancing autonomous and sustainable freight movement. © 2025, CC BY.

关键词： dynamic routing algorithms

CTWR: A congestion, temperature and wear-aware routing algorithm for partially-connected 3D network-on-chip

学校读者我要写书评

暂无评论

Computers and Electrical Engineering 2025年 124卷

作者： Asadboland, Mojtaba Mehranzadeh, Amin Mosleh, Mohammad Department of Computer Engineering Dezful Branch Islamic Azad University Dezful Iran

Three-dimensional Network-on-Chip (3D-NoC) is an efficient solution to overcome communication limitations in complex System-on-Chip (SoC) architectures. However, challenges such as increased temperature, traffic congestion, and link wear-out significantly impact network performance and lifespan. In this study, we propose an adaptive routing algorithm named CTWR (Congestion, Temperature and Wear-aware routing), which simultaneously considers temperature, congestion, and wear-out while utilizing both intra-layer and inter-layer routing approaches to enhance network performance. The algorithm employs a dynamic approach to assess the real-time status of vertical links to control and reduce wear-out, selecting paths that mitigate thermal hotspots, balance traffic distribution, and extend the lifespan of interconnects. Extensive assessments and simulations performed under diverse traffic scenarios and multiple vertical link or elevator layout configurations indicate that the CTWR algorithm outperforms ETW, EF, HE, and Nezarat routing methods in reducing average packet delay by 92.71 %, 67.84 %, 56.33 %, and 26.91 %, respectively. Furthermore, our proposed approach enhances average network throughput by 9.88 %, 4.38 %, 2.64 %, and 1.66 % compared to these methods. Thermal analysis of the chip surface also reveals a lower overall temperature and a more balanced heat distribution than competing techniques. © 2025 Elsevier Ltd

关键词： dynamic routing algorithms