Nowadays, with the development of distributed energy resource (DER) technologies, various DERs would be widely applied to residential houses. By integrating electricity, thermal energy, natural gas, and other forms into sectors, it provides an opportunity for users exploit complementarity multi‐energy resources. However, is also a formidable challenge manage such system (MES) designed The authors propose integrated architecture MES incorporating resources, as renewable energy. Besides,...

The traditional infrastructure in power system is undergoing a transition to the Smart Grid, which communication network and grid will be integrated into cyber-physical (CPPS). Although topological analysis reveals mechanism of cascading failure between two networks, it ignores control redundancy standby lines from grid. robustness CPPS requires more comprehensive model analyze behavior reality. Here, we propose with one-to-multiple interdependency relevant theoretical framework failure. In...

Traffic engineering (TE) is a fundamental task in networking. Conventionally, traffic can take any path connecting the source and destination. Emerging technologies such as segment routing, however, use logical paths that are composed of shortest going through predetermined set middlepoints order to reduce flow table overhead TE implementation. Inspired by this, this paper, we introduce problem node-constrained TE, where must go middlepoints, study its theoretical fundamentals. We show...

To effectively coordinate the distributed energy resources calls for reasonable management methods. Nowadays, there are a great number of studies about home systems (HEMSs) and centralised control. However, former has problem producing new peaks valley when incentives applied, latter only focuses on overall economic benefits ignores individual preferences. Therefore, authors take an emerging decentralised solution, i.e. peer‐to‐peer (P2P) sharing mechanism, into consideration then propose...

Segment routing is an emerging technology to simplify traffic engineering implementation in WANs. It expresses end-to-end logical path as a sequence of segments through set middlepoints. Traffic along each segment routed shortest paths. In this paper, we study practical (TE) with SDN based We consider two common types TE, and show that the TE problem can be solved weakly polynomial time when number middlepoints fixed not part input. However, corresponding linear program large scale...

DNN inference is becoming prevalent for many real-world applications. Current machine learning frameworks usually schedule tasks with the goal of optimizing throughput under predictable workloads and task arrival patterns. Yet, are more dynamic bursty queries generated by various video analytics pipelines which run expensive only on a fraction frames. Thus it imperative to optimize completion time these unpredictable improve customer experience.

Embedding methods are commonly used in recommender systems to represent features about user and item. An impeding practical challenge is that the large number of embedding vectors incurs substantial memory footprint for serving queries, especially as continues grow. We propose an compression system called Saec address this challenge. exploits similarity among within a field they same attribute or item, uses clustering compress embeddings. new fast method relies on empirical heavy-tailed...

This article introduces Electrical Vehicle Wireless Power Transfer (EV-WPT) technology in detail. By comparing with other charging methods, the necessity of EVWPT is highlighted, and differences between electromagnetic induction wireless power supply coupling resonant supply, dynamic system static are discussed On this basis, research status electric vehicles home abroad introduced, results transmission team Chongqing University field introduced Finally, current problems development...

Adding semantics to Web Services can automate the processes of discovery, selection and composition services. Although many annotating models are proposed support this automation, adoption these is significantly hampered by tedious manual annotation process. The unstructured nature descriptions RESTful services making goal even harder achieve than traditional Services. To address problem, we propose ASSARS, namely Automated Structural Semantic Annotation for Services, automatically transform...

In this paper we consider the topology aggregation problem in hierarchical ASON networks. Usually, requirement of scalability and security network is fulfilled by domain division. The relevant information exchanged among domains will be greatly increased for large scale aims at reducing amount link state without losing accuracy routing calculation. We propose to apply a bidirectional shuffle-net model abstraction procedure, which can significantly reduce size information. A genetic algorithm...

Existing opportunistic network routing algorithms usually have two main problems: excessive calculation of key nodes leads to the uneven energy consumption nodes, and limited remaining cache loss important messages. To solve above problems, this paper proposed a new algorithm-EC-CW, which forwards messages according multi-copy mechanism communication willingness between nodes. The simulation results show that EC-CW reduces average latency overhead rate in nodes-sparse scenarios composed...

Owing to the large-scale integration of distributed generation units, traditional distribution networks are gradually turning active networks, which poses impacts on including voltage violations, degradation power losses and failures relay protections. Therefore, higher requirements network in terms planning, control, quality management, operation maintenance have been put forward. The OPF algorithms drawbacks solving non-linear problems with dispersiveness randomness. In order achieve...

The parameter server architecture is prevalently used for distributed deep learning. Each worker machine in a system trains the complete model, which leads to hefty amount of network data transfer between workers and servers. We empirically observe that has non-negligible impact on training time. To tackle problem, we design new called Stanza. Stanza exploits fact many models such as convolution neural networks, most exchange attributed fully connected layers, while computation carried out...

Production recommendation systems rely on embedding methods to represent various features. An impeding challenge in practice is that the large matrix incurs substantial memory footprint serving as number of features grows over time. We propose a similarity-aware compression method called Saec address this challenge. clusters similar within field reduce size. also adopts fast clustering optimization based feature frequency drastically improve implement and evaluate Numerous, production...

The parameter server architecture is prevalently used for distributed deep learning. Each worker machine in a such system trains the complete model, which leads to large amount of network data transfer between workers and servers. We empirically observe that has major impact on training time. present new called Stanza tackle this problem. exploits fact many models as convolution neural networks, most exchange attributed fully connected layers, while computation carried out convolutional...