Graph-based Retrieval-Augmented Generation (RAG) typically operates on binary Knowledge Graphs (KGs). However, decomposing complex facts into binary triples often leads to semantic fragmentation and longer reasoning paths, increasing the risk of retrieval drift and computational overhead. In contrast, n-ary hypergraphs preserve high-order relational integrity, enabling shallower and more semantically cohesive inference. To exploit this topology, we propose HyperRAG, a framework tailored for n-ary hypergraphs featuring two complementary retrieval paradigms: (i) HyperRetriever learns structural-semantic reasoning over n-ary facts to construct query-conditioned relational chains. It enables accurate factual tracking, adaptive high-order traversal, and interpretable multi-hop reasoning under context constraints. (ii) HyperMemory leverages the LLM's parametric memory to guide beam search, dynamically scoring n-ary facts and entities for query-aware path expansion. Extensive evaluations on WikiTopics (11 closed-domain datasets) and three open-domain QA benchmarks (HotpotQA, MuSiQue, and 2WikiMultiHopQA) validate HyperRAG's effectiveness. HyperRetriever achieves the highest answer accuracy overall, with average gains of 2.95% in MRR and 1.23% in Hits@10 over the strongest baseline. Qualitative analysis further shows that HyperRetriever bridges reasoning gaps through adaptive and interpretable n-ary chain construction, benefiting both open and closed-domain QA. Our codes are publicly available at https://github.com/Vincent-Lien/HyperRAG.git.

The new generations of mobile devices have high processing power and storage, but they lag behind in terms of software systems for big data storage and processing. Hadoop is a scalable platform that provides distributed storage and computational capabilities on clusters of commodity hardware. Building Hadoop on a mobile network enables the devices to run data intensive computing applications without direct knowledge of underlying distributed systems complexities. However, these applications have severe energy and reliability constraints (e.g., caused by unexpected device failures or topology changes in a dynamic network). As mobile devices are more susceptible to unauthorized access, when compared to traditional servers, security is also a concern for sensitive data. Hence, it is paramount to consider reliability, energy efficiency and security for such applications. The MDFS (Mobile Distributed File System) [1] addresses these issues for big data processing in mobile clouds. We have developed the Hadoop MapReduce framework over MDFS and have studied its performance by varying input workloads in a real heterogeneous mobile cluster. Our evaluation shows that the implementation addresses all constraints in processing large amounts of data in mobile clouds. Thus, our system is a viable solution to meet the growing demands of data processing in a mobile environment.

Today's integration of mobile technologies and traditional cloud computing exploits the abundant computation and storage resources in the cloud, to enhance the capabilities of end-user mobile devices. The designs that rely on remote cloud services, however, sometimes overlook the abundant resources (e.g., storage, communication, and computation) on mobile devices. In particular, when the remote cloud services are unavailable (due to service downtime or network issues), these smart devices can no longer function. We propose a Heterogeneous Mobile Cloud (HMC) computing design that efficiently utilizes the communication and computation resources to support data storage and data processing services in a group of mobile devices. Each mobile device may have different energy, communication and computation capabilities, but our Mobile Storage & Processing System (MSPS) ensures that: i) the communication and computation tasks are executed in an energy-efficient manner, ii) task allocation considers device heterogeneity and achieves system-wide load balancing, and iii) the stored data are fault-tolerant. Through extensive simulations and real hardware implementations on Android devices, we demonstrate the performance and feasibility of deploying MSPS in a real heterogeneous mobile environment.

Despite the advances in hardware for hand-held mobile devices, resource-intensive applications (e.g., video and image storage and processing or map-reduce type) still remain off bounds since they require large computation and storage capabilities. Recent research has attempted to address these issues by employing remote servers, such as clouds and peer mobile devices. For mobile devices deployed in dynamic networks (i.e., with frequent topology changes because of node failure/unavailability and mobility as in a mobile cloud), however, challenges of reliability and energy efficiency remain largely unaddressed. To the best of our knowledge, we are the first to address these challenges in an integrated manner for both data storage and processing in mobile cloud, an approach we call k-out-of-n computing. In our solution, mobile devices successfully retrieve or process data, in the most energy-efficient way, as long as k out of n remote servers are accessible. Through a real system implementation we prove the feasibility of our approach. Extensive simulations demonstrate the fault tolerance and energy efficiency performance of our framework in larger scale networks.

Resource Allocation has been widely used for improving various performance metrics in wireless networks. Applying resource allocation to a Mobile Ad Hoc Network (MANET), however, is a challenging problem because of dynamic network topology. In this paper, we develop a novel resource allocation scheme designed for MANETs that minimizes the communication cost for accessing distributed resources while improving the reliability by adopting the k-out-of-n system, a widely used technique for reliability control. Specifically, we propose a scheme that allocates resources to n nodes, called service centers, such that the expected energy consumption for nodes to access k service centers out of the n service centers (k⩽n) is minimized. Our scheme accounts for dynamic network topology by estimating the failure probabilities of nodes and monitoring the network for significant topology changes. In addition, an Importance Sampling technique is used to reduce the computation-overhead. To evaluate the performance, we build a mobile distributed file system based on our resource allocation scheme. Through both extensive simulations and real hardware implementation on Smartphones, we show that our resource allocation scheme effectively reduces energy consumption by up to 45% and increases the successful data retrieval rate by up to 50% in comparison with a greedy algorithm.

With the advance of mobile devices, cloud computing has enabled people to access data and computing resources without spatiotemporal constraints. A common assumption is that mobile devices are well connected to remote data centers and the data centers securely store and process data. However, for systems like mobile cloud deployed in infrastructureless dynamic networks (i.e., with frequent topology changes because of node failure/unavailability and mobility), reliability and energy efficiency remain largely unaddressed challenges. To address these issues, we develop the first 'k-out-of-n computing' framework that ensures nodes retrieve or process data stored in mobile cloud with minimum energy consumption as long as k out of n storage/processing nodes are accessible. We demonstrate the feasibility and performance of our framework through both hardware implementation and extensive simulations.

This paper implements an optimal speed control system of synchronous reluctance motor using sliding mode with genetic algorithm. The mathematical model of synchronous reluctance motor includes core loss to improve the performance in the current control. The sliding mode speed control is used to avoid the load perturbations and parameter variation. The genetic algorithm is used to search the uncertain parameters effort under the occurrence of uncertainties in the control procedure. The control system can be stable and optimize the torque and speed when the uncertain parameters are unknown/bounded. Finally, the simulation results is demonstrated the effectiveness of the proposed scheme.

Message Sequence Charts (MSCs) are a graphical language for description of scenarios in terms of message exchanges between communicating components in a distributed environment. The language is popular in capturing system requirements in the design of reactive systems and communication protocols. In this paper, we review a design anomaly, called race conditions, in an MSC specification and present a few drawbacks with the current solution. We propose a new approach to correcting race conditions, and the limitation of this approach is also discussed.