Remote sensing (RS) images change detection (CD) is essential for the surveillance and prevention of geohazards. Nevertheless, the current deep learning (DL)-based CD methods still face challenges such as pseudo changes, missed detections, and edge noise due to the inadequate research of temporal differences and inconsistent viewing angles between the dual-temporal images. In order to improve the perception of spatiotemporal variations and effectively manage complex motion in spatiotemporal data, this paper proposes a spatiotemporal multi-scale fusion and optical flow warping network (SMOW-Net). Initially, the internal fusion property of 3D convolution enables the simultaneous extraction and fusion of feature information in dual-temporal images. The spatiotemporal multi-scale feature encoder (SMFE) module is proposed to mitigate the semantic gap between low-level and high-level features. This module is designed to aggregate complementary feature information between each level through temporal and spatial independent processing and flexible temporal transposed convolutional layers. Furthermore, the optical flow warper (OFW) module is intended to improve the spatiotemporal dynamic modeling capability in order to manage complex motion data effectively, where a two-channel spatial deformation field is autonomously learned by the network to guide feature alignment. The performance advantage of our network over eleven state-of-the-art methods (SOTA) on the GVLM-CD, LEVIR-CD, WHU-CD, S2Looking, and LEVIR-CD+ datasets is validated by experimental results. Finally, we also introduce SMOW-Net-LW, a lightweight variant with significantly reduced model complexity, suitable for resource-constrained settings, while still achieving excellent performance. The code for this work is available at https://github.com/ChundeLiao/SMOW-Net .

Federated learning (FL) is well-suited for multimodal tasks due to its ability to protect privacy and support local training. However, the complexity of real-world sensor environments causes modality heterogeneity across clients. Some modalities may be missing altogether, making it difficult to construct a generalized global model. Existing multimodal federated learning methods often address modality-missing scenarios under simplified assumptions of modality heterogeneity, typically focusing on unimodal clients and modality-complete multimodal clients. Moreover, to mitigate performance degradation caused by missing modalities, some approaches assume the availability of auxiliary information at the server, which may be impractical in real-world scenarios. Therefore, we propose a novel heterogeneous multimodal Federated Learning with Mask-Restoration and Self-Guidance (FL-MRSG). The Mask-Restoration employs a masking strategy to simulate missing data during feature extraction, enabling the network to learn semantic features of missing modality. Furthermore, we introduce an innovative self-guidance mechanism that leverages the restored data as guidance information, enabling the network to distinguish between complete and missing data representations. In addition, we propose a personalized decoupled aggregation strategy to facilitate the collaborative training of a global model across heterogeneous modality clients. We extend the multimodal test set to arbitrary modality combinations to evaluate the robustness of the global model. Extensive experiments on MOSI and SIMS datasets demonstrate the effectiveness of the proposed FL-MRSG for arbitrary missing modalities.

Ureteral stent tubes are important medical devices used to repair ureteral obstruction or injury. However, relevant experiments of ureteral stent tubes are usually time-consuming and expensive. This research introduces a mechanical model that can simulate the force and deformation of ureteral stents. In addition, a novel optimization algorithm called improved exploration-enhanced gray wolf optimizer (IEE-GWO) is proposed to optimize parameters of the model. In order to balance exploration and exploitation of gray wolf optimizer (GWO), a dimension learning-based hunting (DLH) search strategy and a nonlinear control parameter strategy are integrated into the IEE-GWO. The experimental results show that the proposed IEE-GWO has better performance, such as fast convergence speed and high solution quality. Furthermore, the novel approach can improve the accuracy of the mechanical modal.

The search for multiple escaping targets is a significant issue of cooperative control in multi-agent systems since targets consciously seek to avoid being captured. Moreover, the assumption of continuous observations in existing works is not always suitable due to the limit of measuring equipment and uncertain movement of targets. Therefore, the problem with searching for escaping targets, which can be more aptly labeled "multiple escaping-targets search with random observation conditions" (MESROC), is difficult to address by conventional methods. Inspired by machine learning and the immune response mechanism of human bodies, a self-learning immune co-evolutionary network (SLICEN) is proposed. The SLICEN consists mainly of an immune cellular network (ICN) and an immune learning algorithm (ILA). The ICN provides feasible solutions to MESROC. Different kinds of network models are introduced to work as an ICN, such as convolutional neural networks, extreme learning machines, and support vector machines. The ILA evaluates the performance of feasible solutions and selects the optimal ones to further strengthen ICN reversely. Solutions are repeatedly improved through the co-evolution of ICN and ILA. An essential distinction to conventional machine learning approaches is that SLICEN works well without training samples. Simulations and comparisons demonstrate that patterns of advanced cooperative behavior among searchers function properly. SLICEN is an efficient method for solving MESROC.

In this paper, a computing speed improvement for the clonal selection algorithm (CSA) is proposed based on a degeneration recognizing (DR) method. The degeneration recognizing clonal selection algorithm (DR-CSA) is designed for solving complex engineering multimodal optimization problems. On each iteration of CSA, there is a large amount of eliminated solutions which are usually neglected. But these solutions do contain the knowledge of the nonoptimal area. By storing and utilizing these data, the DR-CSA is aimed to identify part of the new population as degenerated and eliminate them before the evaluation operation, so that a number of evaluation times can be avoided. This pre-elimination operation is able to save computing time because the evaluation is the main reason for the time cost in the complex engineering optimization problem. Experiments on both test function and a real-world engineering optimization problem (wet spinning coagulating process) are conducted. The results show that the proposed DR-CSA is as accurate as regular CSA and is effective in reducing a considerable amount of computing time.

Most of the multivariable industrial production processes have latent failures on account of its complexity and uncertainty. This brings a great hidden threat to the control system's safety. Actuator fault is one of the commonest failures in control systems. This paper addresses a reconfigurable control system whose controller can be reconstructed when any actuator is faulty. Regard the control system as a biological immune system, an immune reconfigurable control system (IRCS) is designed. The IRCS includes an actuator monitor to imitate the immune surveillance mechanism, a decision module to imitate the immune response mechanism, a control strategy repository to imitate the immune memory mechanism and an adaptive learning module to imitate the immune learning mechanism. The IRCS is applied to the polyacrylonitrile carbon fiber coagulation bath, and the simulation results show that when there is any actuator fault, the IRCS can guarantee better performance compared with the conventional control schemes.

This paper studies on the predicted regression model of support vector machines (SVM). Aiming at the shortage that with the amount of samples grows, training time increases rapidly as well, we propose an optimization algorithm to optimize it for large scale dataset. The optimization algorithm is based on ensemble learning and dynamic clustering. Firstly, we use dynamic cluster method to generate different types of sub training set based on fuzzy granular. Then we construct SVM sub-learners. Afterwards we synthesize outputs of each sub-learner by using the strategy of mean squared error. Simulation experimental results demonstrate that the optimization algorithm can increase training speed obviously, and keep the original accuracy compared to traditional SVM.

The stretching process is one of the key sections in fiber production, which is decisive to the quality of the final fiber products. Such a process raises high requirements on the control of the rollers with proper stretching ratios, and the large number of rollers with their special characteristics and the demand for synchronous running usually make the design of a good control scheme difficult. In this paper, a novel bioinspired multilayered intelligent cooperative controller (BMLICC) is proposed to provide a control plan for the interlinked rollers by organizing them into unified stretching units. Based on the multilayer regulation networks of neuroendocrine system in the human body, a networked controller structure is established. It consists of several components like rollers, distributed controllers, communication paths, and conversion units. The rollers in the same unit can exchange the working information rapidly to implement simultaneous response and cooperation. The stretching ratio can be kept stable and has strong resistance against the external disturbances on the stretching system. Both computer-simulation- and device-based experimental results demonstrate that the stretching unit with the proposed BMLICC can maintain its stretching ratio and effectively resist the external disturbances. This is beneficial to improve the performance of the stretched precursors and, furthermore, produce fibers with high quality. The proposed BMLICC can be easily extended to productions with multiple stretching units or industrial processes with similar mechanical structures for better control quality.

Next-generation grid systems where the emphasis shifts to distributed global collaboration, a service-oriented approach, and information layer issues exhibit a strong sense of automation. Requirements for these systems resemble the self-organizing and the self-healing properties of natural ecosystems. Inspired by this resemblance, we introduce some key ecological concepts and mechanisms into the design for the third-generation grid systems. In this article, a novel Ecological Network-based Grid Middleware (ENGM), which is based on Ecological Network Computing Environment (ENCE), is proposed. First, we discuss how to design the ENCE by agent-oriented approaches based on the key concepts and principles of ecosystems. ENCE provides a new computing and problem-solving paradigm by combining natural ecosystem mechanisms with agent technologies. Then, we design the ENGM with built-in mechanisms to support desirable requirements of new grid systems, namely scalability, adaptability, self-organization, simplicity, and survivability. Based on Jeffery's conceptual model, we also present a corresponding grid-computing prototype that embeds ENGM layers from the implementation point of view. The ENGM will be useful to address the challenges of the third-generation grid systems. Finally, as a demonstration, we built an ENGM platform-based commercial grid service environment and developed a prototype of enterprise supply chain management system. The experimental results demonstrate that the proposed ENGM satisfies the requirements of the next-generation grid and is suitable for new generation grid applications. © 2004 Wiley Periodicals, Inc. Int J Int Syst 19: 859–884, 2004.

A new bio-network model for network emergence computation based on mutual-coupled immune network hypothesis in biological immune systems is proposed. The model is useful for service emergence, which is implemented by the super-entities generated by the interaction among the bio-entities created in the bio-network architecture. The bio-entities, the super-entities, and the interaction among the super-entities are designed and implements in our bio-network simulation platform. The emergent services generated from the mutual-coupled immune network computation model can satisfy various users' requirements. Finally, The feasibility of the model is demonstrated for the supply chain formation and reformation through computer simulation experiments on the platform. The simulation results demonstrate that the immune network model developed in our platform can be employed to complicated enterprise applications.

Next generation grid systems where the emphasis shifts to distributed global collaboration, a service-oriented approach and information layer issues exhibit a strong sense of automation. Requirements for these systems resemble the self-organizing and the healing properties of natural ecological networks. Inspired by this resemblance, we introduce some key ecological concepts and mechanisms into the design for the next generation grid systems. In this paper, a novel Ecological Network-based Grid Middleware (ENGM), which is based on Ecological Network Computing Environment (ENCE), is proposed. This middleware system can address the challenges of the next generation grid systems. That is, it achieves built-in mechanisms to support desirable requirements of new grid systems such as self-organization, survivability, self-evolution, scalability, adaptability, and dynamic configuration and reconfiguration.

In this article, we propose a new approach to the virus DNA–based evolutionary algorithm (VDNA-EA) to implement self-learning of a class of Takagi-Sugeno (T-S) fuzzy controllers. The fuzzy controllers use T-S fuzzy rules with linear consequent, the generalized input fuzzy sets, Zadeh fuzzy logic and operators, and the generalized defuzzifier. The fuzzy controllers are proved to be nonlinear proportional-integral (PI) controllers with variable gains. The fuzzy rules are discovered automatically and the design parameters in the input fuzzy sets and the linear rule consequent are optimized simultaneously by the VDNA-EA. The VDNA-EA uses the VDNA encoding method that stemmed from the structure of the VDNA to encode the design parameters of the fuzzy controllers. We use the frameshift decoding method of the VDNA to decode the DNA chromosome into the design parameters of the fuzzy controllers. In addition, the gene transfer operation and bacterial mutation operation inspired by a microbial evolution phenomenon are introduced into the VDNA-EA. Moreover, frameshift mutation operations based on the DNA genetic operations are used in the VDNA-EA to add and delete adaptively fuzzy rules. Our encoding method can significantly shorten the code length of the DNA chromosomes and improve the encoding efficiency. The length of the chromosome is variable and it is easy to insert and delete parts of the chromosome. It is suitable for complex knowledge representation and is easy for the genetic operations at gene level to be introduced into the VDNA-EA. We show how to implement the new method to self-learn a T-S fuzzy controller in the control of a nonlinear system. The fuzzy controller can be constructed automatically by the VDNA-EA. Computer simulation results indicate that the new method is effective and the designed fuzzy controller is satisfactory. © 2003 Wiley Periodicals, Inc.

A new DNA-based evolutionary algorithm (DNA-EA) for automatic design of a class of Takagi-Sugeno (TS) fuzzy controllers is discussed in this paper. The fuzzy rules are automatically discovered, and the design parameters in the input fuzzy sets and the linear rule consequent are optimized simultaneously by the DNA-EA. The DNA-EA uses the DNA encoding method stemmed from the structure of the biological DNA to encode the design parameters of the fuzzy controllers. The gene transfer operation and bacterial mutation operation inspired by a microbial evolution phenomenon are introduced into the DNA-EA. Moreover, frameshift mutation operations based on the DNA genetic operations are used in the DNA-EA. Our encoding method is suitable for complex knowledge representation, and is easy for the genetic operations at gene level to be introduced into the DNA-EA. The length of the chromosome is variable and it is easy to insert and delete parts of the chromosome. As a demonstration, we show how to implement the new method to design automatically a TS fuzzy controller in the control of a nonlinear system. The fuzzy controller can be automatically constructed by the DNA-EA. Computer simulation results indicate that the new method is effective and the designed fuzzy controller is satisfactory.

A new DNA-based immune genetic algorithm (DNA-IGA) is proposed to design a class of Takagi-Sugeno fuzzy control systems. The DNA-IGA is developed from the structures of biological immune system and DNA. The genetic operators of the method are based on the DNA genetic operations and evaluation of immune system. It uses DNA encoding method stemmed from the biological DNA to encode the design parameters of the fuzzy controllers. The DNA encoding method can significantly shorten the code length of antibodies and is suitable for complex knowledge representation. The proposed method has the local search ability and prevents the premature convergence. As a demonstration, we show how to implement the new method to optimize the design parameters of the TS fuzzy controller in control of a nonlinear system. Computer simulation indicates that the designed fuzzy controller is satisfactory.

In this paper, we focus on design of the efficient algorithms and tools for search and retrieval of documents on Internet. We propose an Internet information retrieval model based on mobile agent technique, fuzzy index algorithm, and genetic search methods. The whole work frame and the structure of the model are discussed. Mobile agents can be used to optimize network traffic for distributed applications. Efficient search of documents can be done with genetic algorithm that exploits the principles of temporal and spatial locality. Fuzzy index algorithm is used to intelligently evaluate the importance of the information searched on Internet. We have also implemented a software platform for experimenting with intelligent information retrieval on Internet using Java language. The experimental results demonstrate the effectiveness of our model.

A new universal controller structure suitable for various low or high order plants is presented. The controller includes a basic fuzzy self-tuning P-type immune controller inspired from the immune feedback mechanism of biological immune systems and an incremental block with the purpose to treat with the order of the plants. The P-type immune feedback law is automatically tuned by a fuzzy logic controller. The effectiveness of this control approach is examined by designing a fuzzy self-tuning PID-type immune controller to control tissue temperature during hyperthermia. Computer simulation results demonstrate that the control performance of the immune controller is superior to that of the corresponding linear controller.

A new DNA genetic algorithm (DNA-GA) based on the mechanism of biological DNA and genetic information is proposed. The genetic operators of the DNA-GA are discussed. The DNA encoding method is suitable for the representation of complex knowledge. The DNA-GA is employed to design effective generalized fuzzy systems (GFS) for the modeling and control applications. GFS employ arbitrary fuzzy rules, e(-|/spl alpha//sub 1/x/sub 1/+/spl alpha//sub 2/|)-type input fuzzy sets containing almost arbitrary continuous input fuzzy sets, arbitrary singleton output fuzzy sets, arbitrary fuzzy logic AND, and the generalized defuzzifier containing the widely-used centroid defuzzifier as a special case. The DNA-GA is used to select input variables and to tune the design parameters and membership functions of GFS. As such, the fuzzy rule sets of GFS can be obtained. The work in this paper provides a useful way for the design of fuzzy controllers and fuzzy models.

We propose a new DNA-based genetic algorithm (DNA-GA) to optimize the design parameters of a generalized membership-type Takagi-Sugeno fuzzy controller (GTSFC). The GTSFC employs TS fuzzy rules with linear consequent, e/sup -|ax+b|(c)/-type input fuzzy sets containing almost arbitrary continuous input fuzzy sets, Zadeh fuzzy logic AND operation, and the widely-used centroid defuzzier. The GTSFC is proved to be a nonlinear PI controller with variable gains. The optimized design parameters are the input fuzzy sets and the linear consequent of the rules. The DNA-GA uses a DNA encoding method stemmed from the structure of the biological DNA to encode the design parameters of the GTSFC. The genetic operators of the method are based on the DNA genetic operations. The encoding method can significantly shorten the code length of DNA chromosomes and is suitable for complex knowledge representation. As a demonstration, we show how to implement the new method to optimize the design parameters of the GTSFC to control a nonlinear system. Computer simulation results indicate that the performance of the designed fuzzy controller is satisfactory.