Solving nonlinear equation systems (NESs) is important in engineering and science. In general, NESs have more than one root, and it is a great challenge to locate multiple roots in a single run for numerical computati...
详细信息
Solving nonlinear equation systems (NESs) is important in engineering and science. In general, NESs have more than one root, and it is a great challenge to locate multiple roots in a single run for numerical computation. In order to tackle this challenging task, we propose a neighborhood-based particle swarm optimization with discrete crossover (MNPSO). MNPSO has three main characteristics: (1) it adopts a two-stage framework to balance exploitation and exploration well;(2) a distance-based adaptive neighborhood technique is proposed, which can form an appropriate neighborhood for each particle;(3) a discrete crossover operator improves the performance of PSO in the high dimension problems. We choose thirty NESs with different features as the test suite to evaluate the performance of MNPSO. The experimental results reveal that MNPSO can locate multiple roots in a single run, and the root ratio and the success ratio of MNPSO are higher than other methods.
作者:
Zhang, LupingXu, FeiHuazhong Univ Sci & Technol
Sch Artificial Intelligence & Automation Key Lab Image Informat Proc & Intelligent Control Minist China Wuhan 430074 Hubei Peoples R China
Homogenous spiking neural P systems (HSNP systems) are a class of neuron-inspired computing models, where each neuron contains an identical set of rules. It remains open how to design universal asynchronous HSNP syste...
详细信息
Homogenous spiking neural P systems (HSNP systems) are a class of neuron-inspired computing models, where each neuron contains an identical set of rules. It remains open how to design universal asynchronous HSNP systems. In this work, we introduce local rule synchronization into asynchronous HSNP systems, and such systems are abbreviated as AHSNPR systems. Specifically, a family of the rule sets is specified;a rule in a specified rule set is applied synchronously with all the other rules in the same set, and a rule not in any specified rule set is applied asynchronously. We investigate the number generating power of AHSNPR systems. It is proved that general and unbounded AHSNPR systems are universal, and bounded AHSNPR systems are only able to characterize the semilinear sets of numbers achieving the corresponding properties of decidability and closure. The results show that the local rule synchronization is useful in constructing universal asynchronous HSNP systems. (C) 2022 Elsevier B.V. All rights reserved.
The complex networks exhibit significant heterogeneity in node connections, resulting in a few nodes playing critical roles in various scenarios, including decision-making, disease control, and population immunity. Th...
The complex networks exhibit significant heterogeneity in node connections, resulting in a few nodes playing critical roles in various scenarios, including decision-making, disease control, and population immunity. Therefore, accurately identifying these influential nodes that play crucial roles in networks is very important. Many methods have been proposed in different fields to solve this issue. This paper focuses on the different types of disassortativity existing in networks and innovatively introduces the concept of disassortativity of the node, namely, the inconsistency between the degree of a node and the degrees of its neighboring nodes, and proposes a measure of disassortativity of the node (DoN) by a step function. Furthermore, the paper analyzes and indicates that in many real-world network applications, such as online social networks, the influence of nodes within the network is often associated with disassortativity of the node and the community boundary structure of the network. Thus, the influential metric of node based on disassortativity and community structure (mDC) is proposed. Extensive experiments are conducted in synthetic and real networks, and the performance of the DoN and mDC is validated through network robustness experiments and immune experiment of disease infection. Experimental and analytical results demonstrate that compared to other state-of-the-art centrality measures, the proposed methods (DoN and mDC) exhibits superior identification performance and efficiency, particularly in non-disassortative networks and networks with clear community structures. Furthermore, we find that the DoN and mDC exhibit high stability to network noise and inaccuracies of the network data.
Performance degradation in solid oxide fuel cells (SOFCs) leads to shorter service life and unexpected downtime. To reduce economic losses and accelerate commercialization, accurately predicting the degradation is con...
详细信息
Performance degradation in solid oxide fuel cells (SOFCs) leads to shorter service life and unexpected downtime. To reduce economic losses and accelerate commercialization, accurately predicting the degradation is conducted in this study. First, a comprehensive analysis of performance degradation through experiments on a real SOFC system is investigated. Then, three dada-driven robust models, that is, vector autoregressive moving average (VARMA), radial basis function neural network (RBFNN), and neural basis expansion analysis for time series (N-BEATS) models are proposed to predict the SOFC's performance degradation. Herein, the top 60-90% of the experimental datasets are used for training and the bottom 40-10% for testing. After training, the prediction performance testing of these 3 models is compared with that of the bi-long short-term memory networks (bi-LSTM) and bi-gated recurrent units (bi-GRU) models. Simulation results show that both the VARMA and N-BEATS models are superior to the bi-LSTM and bi-GRU models in predicting the performance degradation of the SOFC. While the test performance of the RBFNN model is worst, especially under the top 60% training datasets condition. These indicate it is feasible to respectively establish the VARMA model and the N-BEATS model for predicting the SOFC's performance degradation.
Deep learning-based methods have shown their wide application prospects in the field of solid oxide fuel cell(SOFC) prediction. However, the irrationality of the prediction object and the lack of prediction accuracy h...
详细信息
Industrial manufacturing processes present unique challenges in implementing demand response due to their complex equipment interactions, diverse operation modes, and safety constraints. Conventional model-based optim...
详细信息
Industrial manufacturing processes present unique challenges in implementing demand response due to their complex equipment interactions, diverse operation modes, and safety constraints. Conventional model-based optimization methods often struggle in this context, requiring complete mathematical models and accurate uncertainty distributions. In this article, we propose a model-free safe deep reinforcement learning approach for real-time scheduling of industrial manufacturing systems, ensuring constraint adherence and accommodating diverse equipment operation modes. Specifically, the approach formulates the industrial demand response problem as a constrained Markov decision process with hybrid action space, capturing the intricate interplay between variable-speed equipment and discrete actuators, while considering safety constraints. To enhance exploration and robustness, the proposed method combines Lagrange multipliers based on soft actor-critic to satisfy the constraints. The cross-attention mechanism is utilized to find the association rules between hybrid actions to solve the challenges posed by the spatial gradient of hybrid actions. The proposed approach is trained and tested on a real-world dataset, demonstrating its superior performance in achieving significant cost reductions for manufacturing while satisfying operational constraints. Furthermore, sensitivity analysis underpins robustness against the variable real-time prices, showcasing its industrial applicability.
In this paper, the p-th moment synchronization problem for a class of stochastic multi-layer neural networks with intra-layer and inter-layer connections is investigated. Due to the multiple connections with delays an...
详细信息
In this paper, the p-th moment synchronization problem for a class of stochastic multi-layer neural networks with intra-layer and inter-layer connections is investigated. Due to the multiple connections with delays and stochastic noise, the typical methodologies that build a canonical linear or expanded matrix model to analyze its stability by constraining eigenvalues in the left-half plane, such as the Kronecker product method, linear matrix inequality and M-matrix approach are tough to tackle the problem. Consequently, a graph-theory-based Lyapunov functional is constructed by combining multiplicative principles and a graph-theoretic approach to help examine the effect of inter- and intra-layer connectivity on a unified framework. With the proposed adaptive fixed-time controller, sufficient conditions for the p-th moment synchronization in a fixed time are derived in terms of algebraic inequality. A corollary, together with a constant-gain fixed-time controller, is presented in case there is no delay. Finally, a confirmatory and two comparative simulations show the effectiveness and convenient implementation of the proposed control strategy.
Asynchronous spiking neural P systems with rules on synapses (ARSSN P systems) are a class of computation models, where spiking rules are placed on synapses. In this work, we investigate the computation power of ARSSN...
详细信息
Asynchronous spiking neural P systems with rules on synapses (ARSSN P systems) are a class of computation models, where spiking rules are placed on synapses. In this work, we investigate the computation power of ARSSN P systems working in the rule synchronization mode, where a family of rule sets are specified, and all the rules in a such set should be synchronously used or not. We prove that ARSSN P systems working in the rule synchronization mode are universal as number generating. Additionally, two universal ARSSN P systems working in the rule synchronization mode are constructed to accept numbers and compute functions, respectively. The results indicate that rule synchronization set is a powerful ingredient for resource-saving, as the constructed universal ARSSN P systems with rule synchronization sets use less neurons than the counterpart universal systems without such sets.
Most current infrared small target detection methods attempt to fuse local and global information by using single-scale inputs and creating a multiscale feature pyramid during network feeding forward. Further to this,...
详细信息
Most current infrared small target detection methods attempt to fuse local and global information by using single-scale inputs and creating a multiscale feature pyramid during network feeding forward. Further to this, our research finds that using high-resolution inputs can improve recall, while low-resolution inputs improve precision. Nevertheless, solely focusing on global or local information can result in missing targets and false alarms. To address these issues, we propose the BPR-Net to balance precision and recall via a novel multiscale attention mechanism, which combines semantic and shallow features of multiscale inputs (MS). We first scale the input image into multiple images with varying resolutions and feed them into the network. In the encoder, the scale fusion module (SFM) fuses features from corresponding images of different resolutions. In the decoder, a channel fusion module (CFM) fuses useful information from multiple channels. Furthermore, a wavelet transform cross-layer skip layer (WTL) is employed to enhance the interaction between decoder layers for more effective multiscale feature fusion. Experimental results demonstrate that our approach achieves a balance between recall and precision and yields state-of-the-art performance on challenging benchmarks including Sirst, miss detection versus false alarm (MDvsFA), and small infrared aerial target detection (SIATD). Notably, our approach achieves an F1 score of 0.9409 on the challenging benchmark SIATD, surpassing the state-of-the-art method by 16.7%.
Advanced cyber-physical power systems (CPPS) has been put forward by the strong integration of energy networks and communication networks. While CPPS brings a promising solution with high efficiency, strong flexibilit...
详细信息
Advanced cyber-physical power systems (CPPS) has been put forward by the strong integration of energy networks and communication networks. While CPPS brings a promising solution with high efficiency, strong flexibility, great scalability, and improved reliability, it inevitably poses some security challenges. In order to address these challenges, it is essential to accurately describe the attack behavior and system security situation. In this article, a dynamic risk propagation evaluation approach is proposed for accurately predicting attacks and quantitatively analyzing system risk. It is equipped with a partitioned cellular automata model to deal with spatial heterogeneity in the partitioned system. The intentions of targeted attack are also considered for predicting attacks. Then, the cyber-to-physical risk is quantitatively identified from multiple dimensions. Finally, the verification of attack intention is designed to dynamically update and adjust the predicted result. The presented approach is demonstrated through a case study on a CPPS.
暂无评论