In this article, a novel method, called attention enhanced reinforcement learning (AERL), is proposed to address issues including complex interaction, limited communication range, and time-varying topology for multi agent cooperation. AERL includes network (CEN), graph spatiotemporal long short-term memory (GST-LSTM), parameters sharing multi-pseudo critic proximal policy optimization (PS-MPC-PPO). Specifically, CEN based on mechanism designed enlarge the agents' range deal with interaction...

Air-to-air close-in combat is based on many basic fighter maneuvers and can be largely modeled as an algorithmic function of inputs. This paper studies autonomous combat, to learn new strategy that adapt different circumstances fight against opponent. Current methods for learning are limited discrete action sets whether in the form rules, actions or sub-polices. In contrast, we consider one-on-one air game with continuous space present a deep reinforcement method proximal policy optimization...

Making efficient motion decisions for a multi-robot system is challenging problem in target encirclement with collision avoidance. Specifically, each robot local communication has to consider cooperative and avoidance simultaneously. In this paper, distributed transferable policy network framework based on deep reinforcement learning proposed solve the of The able process information uncertain number robots obstacles, which desirable property systems. particular, graph attention mechanism...

In this paper, we propose a novel decentralized method based on deep reinforcement learning using robot-level and target-level relational graphs, to solve the problem of multi-target encirclement with collision avoidance (MECA). Specifically, composed three heterogeneous graphs between each robot other robots, targets obstacles, are modeled learned through graph attention networks (GATs) for extracting different spatial representations. Moreover, target within observation robot, is built GAT...

Finding collision-free and efficient paths in an uncertain dynamic environment is a challenge for robot navigation tasks, especially when there are external autonomous agents that also have decision-making abilities the same environment. This paper develops novel method based on DRL with graph attention network (GAT) to solve problem of among (other agents). Specifically, GAT adopted describe other as specific graph, extract spatial structural influence features from graph. Multi-head...

This paper considers a multi-target coverage problem where robot team aims to efficiently cover multi-targets while maintaining connectivity in distributed manner. A novel knowledge-incorporated policy framework is proposed derive distributed, efficient, and guaranteed policy. In particular, knowledge-guided network (KGPnet) designed, which consists of observation attention representation, interaction learning. Giving credit the KGPnet, can be applied different number targets. Moreover,...

Recently, a lot of works have been devoted to researching how agents can learn efficient cooperation in multiagent systems. However, it still remains challenging large-scale systems (MASs) due the complex dynamics between and environment dimension explosion state-action space. In this paper, we propose novel MultiAgent Automatic Curriculum Learning method (MA-ACL) solve learning problems cooperative MASs by beginning from on scenario with few automatically progressively increasing number...

In this paper, we propose a novel distributed method based on attention-based deep reinforcement learning using individual reward shaping, for multiple unmanned aerial vehicles (UAVs) cooperative short-range combat mission. Specifically, two-level attention policy, composed of observation-level and communication-level networks, is designed to enable each UAV selectively focus important environmental features messages, enhancing the effectiveness policy. Moreover, due high complexity...

In this paper we have reviewed our achievements in soft X-ray and extreme ultraviolet (EUV) optics. Up to now, the research system of EUV optics has been established, including light sources, detectors, calibrations, optical testing machining super smooth mirrors, fabrications multilayer film mirrors. Based on achievements, developed two types solar space telescopes for observations. One is an normal incident telescope array 4 different operation wavelength telescopes. The wavelengths are...

Relational databases are widely used today as a mechanism for providing access to structured data. They, however, not suitable typical information finding tasks of end users. There is often semantic gap between the queries users want express and that can be answered by database. In this paper, we propose system bridges using domain knowledge contained in ontologies. Our extends relational with ability answer represented SPARQL, an emerging Semantic Web query language. Users their based on...

Social learning, especially social incentives, is extremely important for humans to achieve a high level of coordination. Inspired by this, we introduce this concept into cooperative multiagent reinforcement learning (MARL), implicitly address the credit assignment problem and promote interagent direct interactions cooperations among agents in games. In article, propose novel intrinsic reward method with peer incentives (IRPI) based on actor–critic policy gradient. This can enable...

Mobile network traffic engineering and management activities require traces where each packet or flow is associated with some ground truth regarding mobile app protocol. This paper presents a system named mobilegt that collects links the to it without rooting devices. It consists of two elements: client server. Mobilegt iteratively probes monitored nodes' kernel obtain socket information on active TCP/UDP sessions. server captures packets generated nodes at aid Virtual Private Network (VPN),...

Since Al Gore created the vision for Digital Earth in 1998, a wide range of research this field has been published journals. However, little attention paid to bibliometric analysis literature on Earth. This study uses methodology publications related Science Citation Index database and Social (via Web online services) during period from 1998 2015. In paper, we developed novel keyword set 'Digital Earth'. Using set, 11,061 scientific articles 23 subject categories were retrieved. Based...

A huge amount of mobile traffic runs on Internet every day. The complex data may degrade the network performance if they are not well managed. As important foundation management, classification aim to map IP packets into a predefined class set. Most work in this area until now has focused schemes. In contrast, little attention been paid towards definition. Various kinds lead difficultly comparing these methods. This paper presents taxonomy based ontology paradigm. Its meta-characteristic is...

In this paper, the impact time control problem has been investigated by applying polynomial shaping method. By modifying kinematics between maneuvering target and interceptor into relative engagement kinematics, paper provides a method to deal with against target. Shaping range remained intercept time, guidance law called (RPITCG) is proposed. The coefficients of terms are determined boundary conditions. Then acceleration given in form proportional navigation (PN) varying gain. Another...

Hamming quasi-cyclic (HQC) cryptosystem, proposed by Aguilar Melchor et al., is a code-based key encapsulation mechanism (KEM) submitted for the NIST standardisation process of post-quantum cryptography (PQC). Under assumption that s -decision syndrome decoding (s -DQCSD) problem hard = 2 and 3, HQC, viewed as public-key encryption scheme, proven to be indistinguishability under chosen plaintext attack (IND-CPA) secure, can transformed into an IND-Adaptive ciphertext secure KEM. However,...

Multi-agent cooperation is one of the most attractive research fields in multi-agent systems. There are many attempts made by researchers this field to promote behavior. However, partially-observable environments, a large number agents and complex interactions among cause huge difficulty for policy learning. Moreover, redundant communication contents caused make effective features hard be extracted, which prevents from converging. To address limitations above, novel method called cognition...