Search

Recently, using unmanned platforms to perform target search tasks has received extensive attention and research. The complexity of the search scenario and the unpredictability of the target movement pose significant challenges for unmanned platforms to perform the search task. Developing efficient search strategies is crucial for their success. In this study, we model the problem as a Partially Observable Markov Decision Process (POMDP) and propose a Hierarchical Deep Q Network with Multi-criteria Negative Feedback method named MNF-HDQN to solve the problem efficiently. The MNF-HDQN incorporates point and area evaluations instead of the original sparsity calculation to enhance the cooperation competence of unmanned platforms in searching for a target in various search tasks. Additionally, the integration of convex polygon theory into reward shaping and the design of a new two-stage search strategy encouragement function further improve the performance of the proposed method. We conduct detailed experiments to verify the superiority of the MNF-HDQN. And experimental results show that our method provides a high successful search rate in a shorter timestep compared with state-of-the-art baselines. This advantage is more evident when the search scenario is more complex or the target movement is more unpredictable. • We propose a new reinforcement learning method for the multi-agent target search strategy optimization problem. • We design a novel two-stage reward function based on convex polygon theory to handle reward sparsity. • Our method develops a hindsight experience replay with multi-criteria negative feedback, the new feedback evaluation that combines two special indicators. • Experimental results demonstrate the superiority and the generalization ability of MNF-HDQN. [ABSTRACT FROM AUTHOR]

With the development of cloud storage system and its application in complex environment, its data security has been more and more attention. On the one hand, node crashes or external invasion are likely to lead to incomplete data; on the other hand, when the data is incomplete, because the cloud service provider deliberately concealed or other factors, the user cannot be promptly informed of the change. In view of the above problems, this paper makes a deep research, and puts forward a secure storage system based on how to ensure the data availability when data integrity and data are not complete. In this paper, we begin with the availability of data; the research focuses on the confidentiality of data, the loss of data recovery and data recovery. In this paper, we propose a data secure storage scheme based on Tornado codes (DSBT) by combining the technique of symmetric encryption and erasure codes. Program uses boot password to solve the traditional data encryption in the problem of key preservation and management; system design by correcting Tornado data redundancy code delete code in order to solve problems and recover lost data; through a hash keyed to Tornado code with error correction function so as to solve the problem of data tampering. On this basis, the paper continues to carry out research on data retrieval (POR). Based on the classic POR algorithm based on BLS short signature, the trusted log is introduced, and the trusted log is used to provide the user with the test results. Finally, combined with the DSBT scheme, the computational efficiency of the POR algorithm is optimized, which has nothing to do with the file size, which can achieve the calculation complexity of the constant level. According to the above scheme, this paper implements a secure cloud storage prototype system based on Cassandra. The test shows that the system can provide strong data loss recovery ability, effectively resist the Byzantine fault, in the back of the desirable detection ability is also prominent, but also has very high computation efficiency, especially in the face of large files. This paper studies the modeling and analysis methods of some key problems of data security in cloud storage, such as encryption storage, integrity verification, access control, and verification and so on. Through the data segmentation and refinement rules algorithm to optimize the access control strategy, using the data label verification cloud data integrity, using replica strategy to ensure the data availability, the height of authentication to strengthen security, attribute encryption method using signcryption technology to improve the algorithm efficiency, the use of time encryption and DHT network to ensure that the cipher text and key to delete the data, so as to establish a security scheme for cloud storage has the characteristics of privacy protection. [ABSTRACT FROM AUTHOR]

This paper studies the algebraic formulation and optimization control for a class of networked evolutionary games with switched topologies via the semi-tensor product method. First, an algebraic expression is formulated for the given networked evolutionary games, based on which, the behavior of corresponding evolutionary games is analyzed. Then, under some certain assumptions, the existence of fixed points for the given systems is proved and a free-type control sequence is designed to guarantee the best strategy profile reachable globally. Finally, an illustrative example is studied to support our new results. [ABSTRACT FROM AUTHOR]

Building projects consume large amount of energy and resources, and emit solid waste and CO 2 harmful to the eco-environment. In order to study the eco-environmental impact of green building development policies, a “Green Building Eco-environment (GBE)” model is constructed with the method of System Dynamics and implemented using the Vensim software. The model is used to simulate and evaluate the current state and future trend of variation of the eco-environmental impact of green building development in Wuhan during the years 2008–2050 under current green building development policies. Some policy factors are then adjusted in the simulation to determine the optimal green building development policies, under which the quality of the regional ecological environment in Wuhan would be improved most economically. [ABSTRACT FROM AUTHOR]

During the past 70 years, prestressed-concrete cylinder pipe (PCCP) has been used to construct approximately 19,000 mi of US water transmission mains. Numerous unanticipated performance shortfalls, caused largely by the actions of one manufacturer during the 1970s, resulted in concerns among utilities related to the use of PCCP that continue to this day. An emerging PCCP risk management strategy is to use the combination of electromagnetic inspection to gauge the current extent of deterioration, finite-element-based structural analysis to gauge the significance of deterioration relative to risk of failure, and acoustic monitoring to gauge the rate of deterioration. This article examines the state of the art as it relates to PCCP risk assessment and provides constructive input related to the limitations associated with the various technologies currently being used. Additionally, suggestions for improvements and ways to get the most out of a risk assessment are given. [ABSTRACT FROM AUTHOR]

• Focus on double auction-based microgrid, a multi-agent reinforcement learning approach is adopted to achieve distributed energy scheduling and strategy-making. • A multi-agent deep Q-network (MADQN) algorithm which is a model-free reinforcement learning for agents to learn independent market strategy is first proposed. • An optimal equilibrium selection mechanism is applied to improve the performance of MADQN from benefit fairness, execution efficiency, and privacy protection. • A deep neural network is used in MADQN to approximate the Q-value update in the learning process. With the increasing demands for autonomous decision, market initiative, and privacy protection, a distributed energy management and strategy-making framework is more appropriate for the future microgrid. Besides, the integration of various distributed generations (DGs) and load facilities with high randomness poses challenges to traditional model-free microgrid market management approaches. Therefore, this paper concentrates on the distributed energy management and strategy optimization for a regional microgrid from the following aspects. First, a multi-agent reinforcement learning (MARL) framework is applied to generate independent real-time market decisions and to keep the benefits balance during agents' interactive learning. Moreover, to solve the defects of traditional RL in the microgrid case of continuous state inputs, a multi-agent deep Q-network (MADQN) is adopted as the approximation of value function. Finally, an optimal equilibrium selection (OES) mechanism is proposed to calculate the collective update objective for promoting learning efficiency and to ensure benefits equilibrium. Combining theoretical analysis and simulation results, the proposed MARL framework can improve the microgrid operation performance from economy, independence, and benefit balance among microgrid participants. [ABSTRACT FROM AUTHOR]