Deep reinforcement learning (DRL) integrates the feature representation ability of deep with decision-making so that it can achieve powerful end-to-end control capabilities. In past decade, DRL has made substantial advances in many tasks require perceiving high-dimensional input and making optimal or near-optimal decisions. However, there are still challenging problems theory applications DRL, especially limited samples, sparse rewards, multiple agents. Researchers have proposed various...

Up to now, immunization of disease propagation has attracted great attention in both theoretical and experimental researches. However, vast majority existing achievements are limited the simple assumption single layer networked population, which seems obviously inconsistent with recent development complex network theory: each node could possess multiple roles different topology connections. Inspired by this fact, we here propose strategies on multiplex networks, including node-based random...

Efficient security patch distribution is of essential importance for updating anti-virus software to ensure effective and timely virus detection cleanup. In this paper, we propose a mixed strategy combine the advantages traditional centralized decentralized strategy. A novel network model that contains central node multiplex composed dissemination layer propagation presented, competing spreading dynamical process on top simulates interplay between developed. Such new framework helps in...

How to find the effective approach of immunizing a population is one open question in research complex systems. Up now, there have been great number works focusing on efficiency various immunization strategies. However, majority these existing achievements are limited isolated networks, how affects disease spreading multiplex networks seems need further exploration. In this letter, we explore impact acquaintance where two kinds strategies, node-based and layer immunization, proposed. With...

Freespace detection is an essential component of autonomous driving technology and plays important role in trajectory planning. In the last decade, deep learning based freespace methods have been proved feasible. However, these efforts were focused on urban road environments few specifically designed for off-road due to lack dataset benchmark. this paper, we present ORFD dataset, which, our knowledge, first dataset. The was collected different scenes (woodland, farmland, grassland...

Epidemiological models based on traditional networks have made important contributions to the analysis and control of malware, disease, rumor propagation. However, higher-order are becoming a more effective means for modeling epidemic spread characterizing topology group interactions. In this article, we propose composite degree Markov chain approach (CEDMA) describe discrete-time dynamics networks. approach, nodes classified according number neighbors hyperedges in different states...

Models for simulation are vital the authenticity of simulation. Modeling doubly fed induction generation (DFIG) power system stability analysis has attracted a great deal research interests in past few years due to wide applications DFIG wind generation. The models developed by generator manufacturers not only complicated, but also proprietary and very specific their own products. Many simplified generic have been proposed. Among them, current-source based model is most widely used analysis....

Abstract From transportation networks to complex infrastructures and social economic networks, a large variety of systems can be described in terms multiplex formed by set nodes interacting through different network layers. Network robustness, as one the most successful application areas has attracted great interest myriad research realms. In this regard, how respond potential attack is still an open issue. Here we study robustness under layer node-based random or targeted attack, which...

The way of information diffusion among individuals can be quite complicated, and it is not only limited to one type communication, but also impacted by multiple channels. Meanwhile, easier for an agent accept idea once the proportion their friends who take goes beyond a specific threshold. Furthermore, in social networks, some higher-order structures, such as simplicial complexes hypergraph, describe more abundant realistic phenomena. Therefore, based on classical multiplex network model...

Over the past decade, coupled spread of information and epidemic on multiplex networks has become an active interesting topic. Recently, it been shown that stationary pairwise interactions have limitations in describing inter-individual , thus, introduction higher-order representation is significant. To this end, we present a new two-layer activity-driven network model, which considers partial mapping relationship among nodes across two layers simultaneously introduces simplicial complexes...

With the continuous development and popularization of Internet, there has been an increasing number network security problems appearing. Among them, rapid growth in malware emergence variants have seriously affected Internet. Traditional detection methods require heavy feature engineering, which affects efficiency detection. Existing deep-learning-based such as poor generalization ability long training time. Therefore, we propose a classification method based on transfer learning for...

The data collected by sensors is streaming in the Internet of Things (IoT). Although existing deep-learning-based anomaly detection methods generally perform well on static data, they struggle to respond timely after distribution changes. However, suffers from conceptual drift due highly dynamic nature IoT. In network security, concept drift-oriented a crucial task, because it can adjust model adapt latest and detect attacks time. Existing are confronted with some challenges, including...