Process simulation is a critical cornerstone of chemical engineering design. Current automated chemical design methodologies focus mainly on various representations of process flow diagrams. However, transforming these diagrams into executable simulation flowsheets remains a time-consuming and labor-intensive endeavor, requiring extensive manual parameter configuration within simulation software. In this work, we propose a novel multi-agent workflow that leverages the semantic understanding capabilities of large language models(LLMs) and enables iterative interactions with chemical process simulation software, achieving end-to-end automated simulation from textual process specifications to computationally validated software configurations for design enhancement. Our approach integrates four specialized agents responsible for task understanding, topology generation, parameter configuration, and evaluation analysis, respectively, coupled with Enhanced Monte Carlo Tree Search to accurately interpret semantics and robustly generate configurations. Evaluated on Simona, a large-scale process description dataset, our method achieves a 31. 1% improvement in the simulation convergence rate compared to state-of-the-art baselines and reduces the design time by 89. 0% compared to the expert manual design. This work demonstrates the potential of AI-assisted chemical process design, which bridges the gap between conceptual design and practical implementation. Our workflow is applicable to diverse process-oriented industries, including pharmaceuticals, petrochemicals, food processing, and manufacturing, offering a generalizable solution for automated process design.

Large language models (LLMs) have demonstrated remarkable capabilities in complex reasoning tasks, yet they often suffer from hallucinations and lack reliable factual grounding. Meanwhile, knowledge graphs (KGs) provide structured factual knowledge, but lack the flexible reasoning abilities of LLMs. In this paper, we present Reason-Align-Respond (RAR), a novel framework that systematically integrates LLM reasoning with knowledge graphs for knowledge graph question answering (KGQA). Our approach consists of three key components: a Reasoner that generates human-like natural language reasoning chains, an Aligner that maps these chains to valid KG paths, and a Responser that synthesizes the final answer. We formulate this process as a latent variable mixture model and optimize it using the Expectation-Maximization algorithm, which iteratively refines the reasoning chains and knowledge paths. Extensive experiments on multiple benchmarks demonstrate the effectiveness of RAR, achieving state-of-the-art performance with Hit scores of 93.3% and 91.0% on WebQSP and CWQ respectively. Human evaluation confirms that RAR generates high-quality, interpretable reasoning chains well-aligned with KG paths while maintaining computational efficiency during inference.

Few-shot class-incremental learning (FSCIL) presents a greater challenge compared with few-shot task-incremental learning (FSTIL) due to the need to classify all previous classes without prior knowledge of the session identifier (session-ID). To address this, we propose Bamboo, a novel framework for FSCIL that introduces a cascading inference mechanism to explicitly infer the session-ID for each sample. This mechanism is enabled by a novel, session-specific equiangular tight frame prototype (ETF-P) classifier. By adaptively fusing session-agnostic and session-specific semantics, the ETF-P classifier reliably determines if a sample belongs to its associated session, which is the core decision required at each step of the cascade. Considering the incremental nature of the learning process, which resembles the continuous growth of bamboo, we treat the base session classifier as the foundational bamboo node and progressively add new session classifiers as additional nodes on top. During the testing phase, each sample flows sequentially through the bamboo nodes, from top to bottom, to determine its session-ID and to be classified accordingly. Overall, the Bamboo framework is capable of perceiving session-ID without prior knowledge and classifying each sample within the correct session, leading to state-of-the-art performance on multiple benchmark datasets.

This article investigates the distributed aggregation optimization problem in multiagent systems (MASs), with a particular focus on addressing the aggregation effect commonly encountered in modern engineering and technological applications. In such scenarios, the local objective function of an agent depends not only on its own decision variables but also interacts with the decision variables of other agents, resulting in complex coupling relationships. To solve these challenges while ensuring that the optimization variables satisfy the coupled inequality constraints, this article introduces a novel framework called distributed aggregative parameter projection (DAPP). Specifically, the proposed distributed protocol is based on an improved parameter projection, including two direction updates, which minimizes the cost function and keeps the search direction obeying the inequality at each iteration. In addition, the linear convergence performance of the proposed scheme over the undirected and connected graph is ensured by rigorous theoretical proof with mild assumptions. Finally, simulation results demonstrate the superior performance of DAPP with smaller global function and faster convergence speed in comparison to the existing method.

Neural Architecture Search (NAS) has emerged as a promising approach to automating the discovery of optimal neural network architectures. However, the computational cost of evaluating candidate architectures through full training presents a significant barrier to efficient search. While surrogate models can accelerate this process by predicting network performance, they struggle to accurately model the vast architecture search space when working with limited training data. To mitigate this challenge, we propose a Multi-Task, Multi-Surrogate Assisted Evolutionary NAS framework (MT-MSAENAS) that combines multiple surrogate models to enhance search efficiency. To fully exploit the limited training data, MT-MSAENAS constructs both strong and weak surrogate predictors, a global model (strong) that captures overall search space patterns and a local model (weak) that specializes in promising regions. Based on the two surrogate model, MT-MSAENAS employs evolutionary multi-tasking optimization by treating the strong and weak model-assisted search as two related optimization tasks to facilitate knowledge transfer between these models and improve the search efficiency. Experiments on the NAS-Bench101 and NAS-Bench201 search spaces show that the proposed algorithm outperforms state-of-the-art methods in architecture search.

Due to the interference of background noise, existing video anomaly detection methods are prone to detect some normal events in complex scenes as anomalies. Meanwhile, we note that the diversity of normal patterns has not been adequately considered, i.e., the normal events that are worthy of reference in the test data have not been properly utilized, which raises the risk of missed and false detections. In this work, we combine the tasks of next-frame prediction and predicted-frame reconstruction to propose a noise-resistant video anomaly detection method. For the prediction task, we develop an RGB Error-Guided Multiscale Predictive Coding (EG-MPC) framework to overcome the interference of background noise on the learning of appearance and motion features of objects at various scales, thus achieving high-quality frame prediction. For the reconstruction task, we introduce the Dynamic Memory Modules (DMMs) into the reconstruction network, and design sparse aggregation and selective update strategies for the memory items in the DMMs to effectively represent diverse normal patterns while increasing the difficulty of reconstructing anomalies, thus making it easier to distinguish between normal and abnormal frames. Extensive experiments on four benchmark datasets demonstrate that our proposed method outperforms state-of-the-art approaches, especially in complex scenarios.

In video anomaly detection task, to mitigate the interference of background noise on learning the appearance and motion features of foreground objects, existing object-centric methods often directly disregard scene information, making it challenging for them to detect scene-dependent anomalies. Moreover, we observe that most methods focus on reconstructing or predicting complete frame-level or object-level RGB information, which limits their inference speed. In this work, we propose a novel inter-frame RGB difference reconstruction network for efficient video anomaly detection. Specifically, we construct separate scene-dependent memory banks (SDMBs) for different scenes to store exclusive normal patterns, thus enabling sensitive detection of scene-dependent anomalies. Meanwhile, we design sparse aggregation and similarity-driven updating mechanisms for the memory items in the SDMBs, which effectively increase the reconstruction error of anomalies by adequately learning the diverse normal patterns in both the training and testing data, thus making the distinction between normal and abnormal frames easier. Extensive experiments on three public datasets demonstrate that our method outperforms state-of-the-art approaches in terms of detection accuracy, false negative rate, and inference speed, particularly in situations with more complex scene and event types.

This paper presents a robust$H_{\infty}$performance analysis method for an uncertain linear parameter-varying (LPV) system to achieve less conservative disturbance rejections. Specifically, by adopting Lyapunov function, a feedback controller is designed based on linear matrix inequalities (LMIs) framework to eliminate the influences of unknown uncertainties for continuous-time and discrete-time LPV systems. Meanwhile, a convex optimization technique is explored for solving optimization problems of bilinear matrix inequalities (BMIs), in which several sufficient conditions guaranteeing system stability are relaxed. Furthermore, compared with some existing approaches, the superior performances of the presented scheme are demonstrated by rigorous theoretical analysis and numerical simulations on a practical truck-trailer system.Note to Practitioners—This paper is motivated by the problem of minimal control domain range (also called conservativeness) in robust$H_{\infty}$control approaches for LPV systems. Generally, LMI conditions are accompanied by a certain level of conservativeness for the robust stability of LPV systems. However, excessive conservatism makes the proposed control strategy impossible to implement in practical engineering. Considering this gap, we formulate a lemma to reduce the conservativeness in robust stability conditions while dealing with the effects of system uncertainty. Strict theoretical proofs and comparative simulation experiments are also conducted to verify the effectiveness of the proposed robust control scheme in continuous and discrete LPV systems.

In this paper, a novel sliding mode control (SMC) method is designed to investigate the finite-time consensus tracking (FTCT) problem of the second-order leader-following multi-agent systems (MASs) with imprecise communication topology of each agent. First, a T-S fuzzy model is introduced to characterize the inexact communication topology of the leader-following MASs. Moreover, a fuzzy SMC law is designed to ensure the reachability of the constructed sliding mode surface (SMS). Meanwhile, in light of the partitioning strategy, sufficient conditions for FTCT of the leader-following MASs are established. It is worth mentioning that the distributed SMC method proposed in this paper is based on the unknown sliding gain, which can greatly reduce the conservatism. Finally, the simulation study on a group of single-link robots and a numerical simulation demonstrate the feasibility of proposed control method. Note to Practitioners—This paper aims to ensure the transient performance and good robustness of second-order multi-agent systems. Generally, the classic nonlinear control method, sliding mode control (SMC), is accompanied by a certain level of conservativeness due to its fixed sliding gain matrix. Moreover, excessive conservatism may prevent the proposed control strategy from being implemented in practical engineering. To address this issue, we formulate an improved finite-time SMC framework to reduce the conservatiness. We also designed an improved algorithm based on genetic algorithm (GA) and linear matrix inequalities (LMIs) to solve the difficulties brought by the new control algorithm, thereby facilitating its practical implementation and enhancing the overall performance of the second-order multi-agent systems.

Considering the complexity of centralized plant-wide optimization and the presence of communication delay in production units, a distributed asynchronous optimization framework is developed for energy consumption optimization problem during ethylene production. First, the energy consumption optimization problem of ethylene process is formulated into a distributed asynchronous optimization problem. Then, considering multiple production units and the material transfer time involved, each unit is treated as a node, and each node is decomposed into calculation nodes, constraint nodes, and delay nodes. Specifically, each production unit can be optimized asynchronously without the necessity for synchronization. To overcome the impact of communication delay on asynchronous process, state vectors are incorporated into the distributed parameter projection algorithm. Additionally, to ensure that the inseparable coupled constraints between nodes during asynchronous operations are met, the parameter projection algorithm is employed. Numerical experiments and industrial simulations demonstrate the proposed algorithm exhibits faster convergence speeds compared to distributed synchronous algorithms, and the energy consumption is lower than the results obtained by the centralized algorithm. Note to Practitioners—As the scale of ethylene production expands, traditional centralized optimization methods struggle to meet the demands of plant-wide optimization due to their high model complexity and slow convergence rates. In this study, we develop a distributed optimization framework, in which the large-scale complex model is decomposed into small parts, and the global optimization task is accomplished cooperatively through local information exchange. Furthermore, considering the time delay due to material residence time in processing units, a distributed asynchronous parameter projection algorithm is proposed to solve the energy consumption issue in ethylene plant. Experimental results demonstrate that the proposed method is capable of converging to the feasible solution in the presence of time delay. Compared with the existing centralized methods, the proposed asynchronous distributed algorithm exhibits better performance, i.e. lower energy consumption. Besides, the proposed asynchronous distributed optimization framework holds potential for application in other process industries, such as the metallurgical industry and steel production process.

Evolutionary algorithms (EAs) have significantly contributed to addressing large-scale crude oil scheduling problems (COSPs) that exceed the capabilities of mathematical programming. However, current research using EAs to resolve COSPs is limited to discrete-time models that assume uniform operation durations. While this simplification transforms COSPs into fixed-length optimization problems suitable for conventional EAs, it fails to adequately tackle the challenge posed by variable operation durations in COSPs. This limitation results in a compromised solution quality for EAs. To bridge the gap between large-scale COSPs and real-world refineries, this paper introduces the formulation of a continuous-time model for a large-scale COSP, accounting for varying operating durations and practical requirements. In an effort to enhance the effectiveness of EAs in optimizing COSPs represented in continuous time, we propose a variable-length constrained evolutionary algorithm (VLCEA). The variable-length encoding in VLCEA, tailored to the specific problem, facilitates the generation of high-quality solutions with variable lengths. This feature addresses the challenge of determining solution length in advance based on a priori knowledge using conventional methods. Additionally, VLCEA incorporates a sorting strategy and a niching-based selection to expedite the search for promising regions, aiming to achieve better solutions in less time. Experimental results on practical cases not only verify that the continuous-time model produces higher quality solutions compared to the discrete-time model but also demonstrate that VLCEA significantly outperforms five state-of-the-art large-scale EAs with fixed length when applied to the continuous-time model.Note to Practitioners—Crude oil scheduling is one of the most challenging problems in the refinery scheduling process. With the increasing size of scheduling tasks, traditional methods have become inefficient and inaccurate. To address this, evolutionary algorithms (EAs) have emerged as powerful tools for solving large-scale crude oil scheduling problems (COSPs). However, existing EA-based approaches often rely on a fixed number of decisions, resulting in suboptimal solutions when confronted with insufficient prior knowledge. To overcome these obstacles, we have developed a comprehensive continuous-time model specifically designed to address COSPs in real-world refineries. Additionally, we propose a variable-length constrained evolutionary algorithm (VLCEA) as an effective approach for resolving these complex scheduling problems. Computational results demonstrate that the VLCEA achieves superior performance compared to three state-of-the-art EAs in practical large-scale cases. Consequently, the proposed VLCEA exhibits significant potential for solving real-world refinery scheduling problems and can be extended to tackle other scheduling problems with an unfixed number of decisions.

This article investigates the problem of distributed Nash equilibrium (NE) seeking in noncooperative games within a directed communication network. For promoting the efficiency of communication among players, a gradient-based dynamic event-triggered mechanism is proposed, where Zeno behavior is excluded. Moreover, based on the Lyapunov stability theory, we derive sufficient conditions for exponential convergence and demonstrate that the seeking strategy proposed facilitates the convergence of players' actions toward the NE. To illustrate the effectiveness of the proposed strategies, simulation results are presented in a system consisting of five agents.

Near-infrared (NIR) technology is an essential online measurement and monitoring technology in modern industrial processes. Although various methods have been developed for spectral multivariate analysis, they often ignore the potential correlation and complementary relationship between wavelengths. To this end, this work proposes the spectral correlation information embedded (SCIE) modeling method to improve NIR model performance by incorporating the underlying correlation structure among spectral wavelengths, in which the structural model is developed to integrate the spectral correlation information matrix into NIR regression analysis. Furthermore, the theory-guided initialization and structural correlation constraint are introduced to assist model learning and avoid overfitting. The effectiveness of the proposed SCIE method is demonstrated on the benchmark dataset and real-world NIR datasets from the refining process. Comparative results indicate that the proposed method effectively selects wavelengths associated with functional groups and achieves accurate prediction of petrochemical product properties.

Few-shot class-incremental learning (FSCIL) requires learning new knowledge without forgetting old knowledge. Forward compatibility can reserve space for novel classes while maintaining base class knowledge in incremental learning. Better forward compatibility is crucial for effectively mastering all knowledge, especially when dealing with a few unknown new classes. In this article, we propose the better forward compatibility pretraining (BFCP) to further enhance forward compatibility in FSCIL. We adopt a two-stage training for the backbone network in the base session. First, we train the backbone network at the image-level to enhance its feature extraction capability, enabling the model to extract valuable information from unknown class images. Second, we fine-tune the backbone network at the feature-level with fake prototypes and instances to achieve clustering base classes and reserve space for unknown new classes. For all incremental new sessions, we freeze the backbone network and employ prototype rectification without further training to refine the prototypes of the novel classes. We conduct extensive experiments with different input scales, including federated cross-domain pretraining and cross-domain class-incremental experiments. BFCP efficiently handles both novel and base classes of each incremental session and significantly outperforms state-of-the-art methods, achieving an average accuracy of 63.47% on the CIFAR100 dataset.

Near-infrared (NIR) technology has gained wide acceptance in practical processes and is now the measurement of choice in many sectors. However, with increasing spectral dimensionality, it is challenging to establish a prediction model with satisfactory stability and generalization. Stochastic configuration networks (SCNs) based on supervisory learning mechanism have demonstrated significant advantages in developing nonlinear learners. However, existing incremental learning strategies make it difficult to achieve fast convergence while obtaining a suitable-scale network in high-dimensional spectra modeling. In addition, the linear or regularization weight estimation methods are vulnerable to outliers and noise in NIR analysis. To accelerate model construction and improve model performance in high-dimensional spectra analysis, the adaptive robust SCN (AR-SCN) algorithm is proposed in this work, which can perform adaptive incremental learning according to the prediction residual and robustly estimate the output weights by the global-local shrinkage strategy. Comparison results on three benchmark NIR datasets and real-world gasoline blending process verify the effectiveness of the proposed method. Compared with the state-of-the-art SCNs, the AR-SCN method can simultaneously improve the construction efficiency and robustness of SCNs.

Many industrial parks, which are connected to the main grid, have integrated renewable energy to reduce carbon emission for achieving the goal of Industry 5.0. However, the optimal scheduling is challenging due to fluctuations in renewable energy generation. Hydrogen, which plays an important role in the future development of the power grid in Industry 5.0, offers an attractive option to coordinate with the batteries. This work focuses on the day-ahead scheduling of a hydrogen-based microgrid for an industrial park. A day-ahead scheduling model is established by taking into consideration the detailed nonlinear energy conversion behavior of the electrolyzer and fuel cell, as well as the two-timescale property of a battery energy storage system (BESS) and the hydrogen system, including an electrolyzer, a hydrogen energy storage system (HESS), and a fuel cell. Note that the optimization problem is a mixed integer nonlinear programming, which is challenging to be solved. A novel multilearning rate reinforcement learning algorithm is proposed and its convergence is also proved based on two-timescale stochastic approximation theory. Simulation results, based on real-world traces in Belgium at a 15-min resolution, are presented, which shows that the proposed method has a higher reward, lower-operating costs and less computing time. It is also found that the shorter scheduling period for the BESS can lead to reduced operating costs by decreasing the required purchasing power and the renewable energy curtailment power.

In recent years, there has been a growing research interest in decision-focused learning, which embeds optimization problems as a layer in learning pipelines and demonstrates a superior performance than the prediction-focused approach. However, for distributionally robust optimization (DRO), a popular paradigm for decision-making under uncertainty, it is still unknown how to embed it as a layer, i.e., how to differentiate decisions with respect to an ambiguity set. In this paper, we develop such differentiable DRO layers for generic mixed-integer DRO problems with parameterized second-order conic ambiguity sets and discuss its extension to Wasserstein ambiguity sets. To differentiate the mixed-integer decisions, we propose a novel dual-view methodology by handling continuous and discrete parts of decisions via different principles. Specifically, we construct a differentiable energy-based surrogate to implement the dual-view methodology and use importance sampling to estimate its gradient. We further prove that such a surrogate enjoys the asymptotic convergency under regularization. As an application of the proposed differentiable DRO layers, we develop a novel decision-focused learning pipeline for contextual distributionally robust decision-making tasks and compare it with the prediction-focused approach in experiments

Enabled by the widely deployed distributed flexible alternating current transmission system (D-FACTS) devices in practical systems, moving target defense (MTD) has been considered as an effective way to detect stealthy false data injection (FDI) attacks by actively changing branch parameters. However, existing MTD methods heavily depend on the assumption that opponents can not timely obtain the newly changed branch parameters. In this paper, a parameter-estimate-first FDI (PEF-FDI) attack is proposed to reveal vulnerabilities of MTD methods in AC state estimation, which can bypass bad data detectors in the existence of MTD. Specifically, a PEF-FDI attack model is proposed to timely construct attack vector and stealthily misguide the results of alternating current (AC) state estimation in the presence of MTD. Requirements of constructing PEF-FDI attacks on eavesdropped measurements are deduced to reveal the limitation on capability of attackers. Simulations in the IEEE 118-bus system verify the performance of the proposed PEF-FDI attacks.

Low-rank matrix recovery is a major challenge in machine learning and computer vision, particularly for large-scale data matrices, as popular methods involving nuclear norm and singular value decomposition (SVD) are associated with high computational costs and biased estimators. To overcome this challenge, we propose a novel approach to learning low-rank matrices based on the matrix volume and a nonconvex logarithmic function. The matrix volume is the product of all the nonzero singular values of a matrix and has unique geometric properties and connections with other convex and nonconvex functions. We establish a generalized nonconvex regularization problem using the penalty function strategy and introduce an accelerated proximal alternating linearized minimization (AccPALM) algorithm with double acceleration, which combines Nesterov’s acceleration and power strategy. The algorithm reduces computational costs and has provable convergence results under the Kurdyka-Łojasiewicz (KŁ) inequality with mild conditions. Our approach shows superior accuracy, efficiency, and convergence behavior compared to other low-rank matrix learning methods on robust matrix completion (RMC) and low-rank representation (LRR) tasks. We analyze the impact of algorithm parameters on convergence and performance and present visually appealing results to further demonstrate the effectiveness of our approach. The proposed methodology represents a promising advance in the field of low-rank matrix recovery, and its effectiveness has been validated via extensive numerical experiments. The source code for the proposed algorithms is accessible at https://github.com/ZhangHengMin/AccPALMcodes.

The near-infrared (NIR) method has shown great potential in estimating key parameters in various industrial processes. Selecting characteristic wavelengths from high-dimensional spectra is crucial in building a prediction model with the satisfactory performance. However, existing wavelength selection methods based on the wavelength or waveband importance estimation are time-consuming and unstable. To improve the stability and generalization of the established model, a novel adaptive robust method is proposed for NIR modeling in this work, in which the pattern-coupled sparse model has been developed to estimate the spectral peak wavebands adaptively. Furthermore, a robust method based on global-local shrinkage is developed for characteristic wavelength selection. Compared with the state-of-the-art techniques, the proposed method is more accurate and robust for NIR modeling on both benchmark and real datasets.

The hydrogen-based multimicrogrid (HMMG) has emerged as a game changer for energy transition. However, it encounters new challenges in tackling uncertainty data streams stemming from intermittent renewable energy and load. This article presents a multiple-time-scale HMMG energy management framework. In the day-ahead stage, the optimal scheduling is determined. The deviations of day-ahead predictions are redressed by intraday rescheduling. To accommodate the uncertainty data streams, a novel data-stream-driven distributionally robust model predictive control is proposed for the HMMG real-time operation. Specifically, a Dirichlet process mixture model is leveraged to construct an online-updated ambiguity set, which adequately characterizes the multimodality and local moment information of uncertainties. Based upon this ambiguity set, the data-stream-driven distributionally robust model predictive control enhances the real-time tracking performance of energy storage references. Its salient feature is the capability of greatly reducing conservatism while ensuring probabilistic operational constraints even under time-varying uncertainty distributions. Since this real-time operation is an intractable infinite-dimensional optimization problem, a novel constraint-tightening technique is proposed to address the computational challenge. Case studies demonstrate that the proposed approach offers advantages over state-of-the-art methods in out-of-sample performance.

Enabled by the widely existed distributed flexible alternating current transmission system devices in power systems, moving target defense (MTD) has been considered as an effective way to detect stealthy false data injection (FDI) attacks. However, due to the limitation of power system topology, not all stealthy FDI attacks can be detected in power system with MTD. In this article, the authors propose a joint meter coding (MC) and moving target defense (MC-MTD) method to cost-effectively improve the detection of stealthy FDI attacks through integrating MC with MTD. Detection conditions and requirements on MC-MTD are theoretically analyzed, which reveal the close coupling between MC and MTD in collaboratively detecting stealthy FDI attacks. The design of the coding matrix and the selection of encoded measurements are theoretically analyzed to integrate MC with MTD in a special case that the coding matrix is diagonal. An optimization of MC-MTD is formulated and approximately solved to improve detection effectiveness with a small defending cost. Finally, simulations are carried out on both direct and alternating current state estimations to validate the performance of MC-MTD.

This article focuses on solving a distributed convex optimization problem of multiagent systems with multiple inequality constraints. Considering communications between agents are prone to failures and not synchronized among themselves in some cases, a novel asynchronous adaptive step sizes-DIGing algorithm is proposed through integrating the projection operators and logic-andframework. In specific, a bilaterally adjustable adaptive step size mechanism is introduced to automatically abandon the irrational evolutionary route which relieves the limitations of traditional DIGing algorithm. By adopting the operator theory, the almost sure convergence of the proposed algorithm under asynchronous communication is proved. Finally, the theoretical and simulation results for the plantwide optimization problem in the ethylene production process illustrate the effectiveness of the proposed algorithm.

This article presents a distributed optimization framework in order to solve the plant-wide energy-saving problem of an ethylene plant. First, the ethylene production process is abstracted into a distributed network, and then, a new distributed consensus algorithm is proposed, which is called adaptive step-size-based distributed proximal consensus algorithm (ASS-DPCA). This algorithm can dynamically adjust the step size and automatically abandon the irrational evolutionary route while eliminating the dependence of optimization algorithms on model gradient information. Moreover, the designed algorithm is able to converge to an optimal solution for any convex cost functions and approach to a convex constraint set of agents over an undirected connected graph. Finally, the results of numerical simulation and industrial experiments show that the algorithm can reduce the total energy consumption of an ethylene plant with less computing time and assured consensus.

The strengthening and the weakening of synaptic strength in existing Bienenstock-Cooper-Munro (BCM) learning rule are determined by a long-term potentiation (LTP) sliding modification threshold and the afferent synaptic activities. However, synaptic long-term depression (LTD) even affects low-active synapses during the induction of synaptic plasticity, which may lead to information loss. Biological experiments have found another LTD threshold that can induce either potentiation or depression or no change, even at the activated synapses. In addition, existing BCM learning rules can only select a set of fixed rule parameters, which is biologically implausible and practically inflexible to learn the structural information of input signals. In this article, an evolved dual-threshold BCM learning rule is proposed to regulate the reservoir internal connection weights of the echo-state-network (ESN), which can contribute to alleviating information loss and enhancing learning performance by introducing different optimal LTD thresholds for different postsynaptic neurons. Our experimental results show that the evolved dual-threshold BCM learning rule can result in the synergistic learning of different plasticity rules, effectively improving the learning performance of an ESN in comparison with existing neural plasticity learning rules and some state-of-the-art ESN variants on three widely used benchmark tasks and the prediction of an esterification process.

This article proposes a distributed optimization approach upon an undirected topology, through only local computation and communication, with the goal of optimizing global function which consists of a host of local functions under complex constraints. In particular, the accelerated distributed Nesterov gradient descent subject to complex constraints (Acc-DNGD-CCs) algorithm is developed for smooth and strongly convex functions. By adopting an estimation mechanism of gradient and only using the history information, the fast optimization of the presented algorithm is ensured. Subsequently, the parameter projection scheme is employed for handling constraints of uncertain parameters introduced by the coupling relationship between the nodes. Meanwhile, the rigorous theoretical proofs along with stability analysis are given to prove the linear convergence of the Acc-DNGD-CC algorithm. Furthermore, compared with some existing algorithms, the superior performances of Acc-DNGD-CC are verified by numerical simulation on a plant-wide ethylene separation optimization process in terms of energy saving.

Distributed optimization achieves a minimized objective function through collaboration among distributed agents. Considering limited communication capabilities and privacy concerns, this article proposes a dynamic event-triggered differentially private gradient-tracking algorithm for distributed optimization. The communication requirement is reduced by event triggering, while the$\epsilon$-differential privacy is guaranteed by perturbations on states and the tracking of the average gradient. The convergence point is uniquely determined by the noise injected to the tracking. Sufficient conditions for stepsizes are established theoretically to guarantee the convergence in mean and almost surely. Moreover, the theoretical privacy level is rigorously obtained and the positive effect of the event-triggered communication on the privacy is also discussed. Simulations are conducted for the classification of the dataset on the stability of a 4-node star power system to verify the theoretical findings.

The success of multitask learning (MTL) can largely be attributed to the shared representation of related tasks, allowing the models to better generalise. In deep learning, this is usually achieved by sharing a common neural network architecture and jointly training the weights. However, the joint training of weighting parameters on multiple related tasks may lead to performance degradation, known as negative transfer. To address this issue, this work proposes an evolutionary multitasking neural architecture search (EMT-NAS) algorithm to accelerate the search process by transferring architectural knowledge across multiple related tasks. In EMT-NAS, unlike the traditional MTL, the model for each task has a personalised network architecture and its own weights, thus offering the capability of effectively alleviating negative transfer. A fitness re-evaluation method is suggested to alleviate fluctuations in performance evaluations resulting from parameter sharing and the mini-batch gradient descent training method, thereby avoiding losing promising solutions during the search process. To rigorously verify the performance of EMT-NAS, the classification tasks used in the empirical assessments are derived from different datasets, including the CIFAR-10 and CIFAR-100, and four MedMNIST datasets. Extensive comparative experiments on different numbers of tasks demonstrate that EMT-NAS takes 8% and up to 40% on CIFAR and MedMNIST, respectively, less time to find competitive neural architectures than its single-task counterparts.

Federated Continual Learning (FCL) approaches exist two major problems of the probability bias and the imbalance in parameter variations. These two problems lead to catastrophic forgetting of the network in the FCL process . Therefore, this paper proposes a novel FCL framework, Federated Probability Memory Recall (FedPMR), to mitigate the probability bias problem and the imbalance in parameter variations. Firstly, for the probability bias problem, this paper designs the Probability Distribution Alignment (PDA) module, which consolidates the memory of old probability experience. Specifically, PDA maintains a replay buffer and uses the probability memory stored in the buffer to correct the offset probabilities of the previous tasks during the two-stage training. Secondly, to alleviate the imbalance in parameter variations, this paper designs the Parameter Consistency Constraint (PCC) module, which constrains the magnitude of neural weight changes for previous tasks. Concretely, PCC applies a set of adaptive weights to subsets of the regularization term that constrains parameter changes, forcing the current model to be sufficiently close to the past model in parameter space distance. Experiments with various levels of task similitude across clients demonstrate that our technique establishes the new state-of-the-art performance when compared to previous FCL approaches.

In multiobjective decision making, most knee identification algorithms implicitly assume that the given solutions are well distributed and can provide sufficient information for identifying knee solutions. However, this assumption may fail to hold when the number of objectives is large or when the shape of the Pareto front is complex. To address the above issues, we propose a knee-oriented solution augmentation (KSA) framework that converts the Pareto front into a multimodal auxiliary function whose basins correspond to the knee regions of the Pareto front. The auxiliary function is then approximated using a surrogate and its basins are identified by a peak detection method. Additional solutions are then generated in the detected basins in the objective space and mapped to the decision space with the help of an inverse model. These solutions are evaluated by the original objective functions and added to the given solution set. To assess the quality of the augmented solution set, a measurement is proposed for the verification of knee solutions when the true Pareto front is unknown. The effectiveness of KSA is verified on widely used benchmark problems and successfully applied to a hybrid electric vehicle controller design problem.

In infrared scenes, humans can easily observe objects in the scene with their eyes, even dim ones. To make the robot have the same visual ability, this paper proposes a pyramid-feature fusion target detection network, called RLPGB-Net, which combines reinforcement learning with aerial targets in the infrared scene. It makes use of the powerful decision-making ability of reinforcement learning to give corresponding weights to the extracted features and highlight the significant features of infrared dim small targets. In reinforcement learning, we use priori strategy guidance and long-term training methods to train weight-regulating agents. To eliminate the local influence on the detection results, such as bright interference points similar to the target, and to solve the problem of dim target detection effectively, the global context boundary attention module is introduced to eliminate the disadvantage of local comparison by using the global characteristics of different dimensions. At the same time, it can prevent the edge information of the refined target from being submerged in the background. Experimental results on SAITD and SIRST data sets show the effectiveness of the proposed method.

Due to the close relevance to the reliability and efficiency of power systems, network parameters such as branch admittance have been the target of various cyberattacks. However, existing attack models are generally based on the impractical assumption that attackers can directly modify the data of network parameter stored in well-secured control centers. This article proposes a practical attack model and designs an optimal strategy to detect malicious modification of critical network parameters. Specifically, the vulnerability of network parameter error processing is discovered and exploited to indirectly modify the data of network parameter without accessing to the well-secured control center. A model of false-data-injection-enabled network parameter modification is proposed, which significantly reduces the requirements on attackers’ capability and system information. An optimal detection strategy is designed based on the analysis of the minimal protection set at a single branch, which can significantly reduce the number of protected measurements in detecting malicious modification of critical network parameters. Finally, numerical simulations are carried out on the PJM 5-bus and the IEEE 118-bus test systems to validate the theoretical results.

Plant-wide optimization plays a vital role in improving the overall performance of large-scale industrial processes. Considering the modeling complexity and convergence difficulty of centralized plant-wide optimization, in this article, we propose a distributed framework by decomposing the global optimization problem into a set of subproblems, where multiple local units interact with each other between nodes. According to the proposed framework, plant-wide optimization problem can be effectively solved by distributed optimization. To eliminate the limitations of existing distributed algorithms, we introduce constraint node to describe the inseparable coupled constraints between nodes. By combining Lagrange duality and parameter projection, the proposed algorithm can solve optimization problems with multiple constraints. Taking ethylene production process as an example, the global energy consumption optimization is guaranteed without the whole-process mechanism model. Numerical simulation and industrial experimental results demonstrate that the proposed algorithm can reduce the energy consumption of the entire ethylene process with fewer computation time.

In the face of high-dimensional and complex data, effective subspace can preserve specific statistical properties and provide an appropriate representation of data, which generally facilitates the underlying tasks such as clustering or classification. Meanwhile, multiple kernel learning is a technique to combine multiple kernels from different feature spaces effectively. Thus, by incorporating multiple kernels into the process of subspace learning, different feature spaces can be projected into a unified subspace. This paper proposes the Multiple Kernel Subspace Learning (MKSL) for embedding the original space into a unified subspace. Multiple kernels of different feature spaces are combined by MKSL in the process of learning, which can extend the suitability for various applications. Moreover, to generate the optimal combination kernel of subspace learning, we propose a two-step iteration strategy to learn the appropriate kernel weights and transformation matrix of projecting simultaneously. Furthermore, our proposed formulation of MKSL can introduce different prior knowledge such as class information and neighborhood relationships. Thus it is competent to the unsupervised learning, semi-supervised learning, and supervised learning. Extensive experiments are conducted on diverse datasets, and the performances are comprehensively evaluated on different tasks. The experimental results indicate that the proposed algorithm is outstanding in unsupervised clustering task and effective in supervised and semi-supervised classification tasks.

Accompanied with the increasing popularity of linear regression approaches, most of the existing minimization problems are related with several convex measurements, e.g.,$\ell_1$/$\ell_2$/$\ell_{2,1}$-norm of a vector and$L_1$/$L_{2,1}$/Frobenius/nuclear norm of a matrix, where the regularized function and the loss function are usually studied for two objective terms case by case, respectively. To address this issue, this work combines these linear regression problems into a unified expression framework by employing an adaptive and flexible function, in which we need to choose different variable elements and adjust an inner parameter, properly. Besides this, they are equipped with some corresponding relationships and their interesting properties. Intuitively speaking, the proposed framework can generalize several traditional linear regression formulations and even more complex ones into an extended representation. For further optimizations, an iteratively re-weighted penalty solution (IRwPS) is devised without any inner loops, making the iteration programming easy to perform. Meanwhile, the theoretical results are provided for guaranteeing that the mathematical convergence analysis is solid and meaningful. Finally, by performing real-world applications in supervised, unsupervised, and semi-supervised tasks, numerical experiments are conducted to validate the theoretical properties and the superiority over some of the state-of-the-art.

In multilabel images, the changeable size, posture, and position of objects in the image will increase the difficulty of classification. Moreover, a large amount of irrelevant information interferes with the recognition of objects. Therefore, how to remove irrelevant information from the image to improve the performance of label recognition is an important problem. In this article, we propose a convolutional network based on feature denoising and details supplement (FDDS) to address this issue. In FDDS, we first design a cascade convolution module (CCM) to collect spatial details of upper features, in order to enhance the information expression of features. Second, the feature denoising module (FDM) is further put forward to reallocate the weight of the feature semantic area, in order to enrich the effective semantic information of the current feature and perform denoising operations on object-irrelevant information. Experimental results show that the proposed FDDS outperforms the existing state-of-the-art models on several benchmark datasets, especially for complex scenes.

Recently, electroencephalogram (EEG) emotion recognition has gradually attracted a lot of attention. This brief designs a novel frame-level teacher-student framework with data privacy (FLTSDP) for EEG emotion recognition. The framework first proposes a teacher-student network without prior professional information for automated filtering of useful frame-level features by a gated mechanism and extracting high-level features by using knowledge distillation to capture the results of EEG emotion recognition from a teacher network and student networks. Then, the results from subnetworks are integrated by using the novel decision module, which, motivated by the voting mechanism, adjusts the composition of feature vectors and improves the weight of accurate prediction to optimize the integration effect. During training, an innovative data privacy protection mechanism is applied for avoiding data sharing, where each student network only inherits weights from all trained networks and does not inherit the training dataset. Here, the framework can be repeatedly optimized and improved by only training the next student subnetwork on new EEG signals. Experimental results show that our framework improves the accuracy of EEG emotion recognition by more than 5% and gets state-of-the-art performance for EEG emotion recognition in the subject-independent mode.

Autonomous systems possess the features of inferring their own state, understanding their surroundings, and performing autonomous navigation. With the applications of learning systems, like deep learning and reinforcement learning, the visual-based self-state estimation, environment perception, and navigation capabilities of autonomous systems have been efficiently addressed, and many new learning-based algorithms have surfaced with respect to autonomous visual perception and navigation. In this review, we focus on the applications of learning-based monocular approaches in ego-motion perception, environment perception, and navigation in autonomous systems, which is different from previous reviews that discussed traditional methods. First, we delineate the shortcomings of existing classical visual simultaneous localization and mapping (vSLAM) solutions, which demonstrate the necessity to integrate deep learning techniques. Second, we review the visual-based environmental perception and understanding methods based on deep learning, including deep learning-based monocular depth estimation, monocular ego-motion prediction, image enhancement, object detection, semantic segmentation, and their combinations with traditional vSLAM frameworks. Then, we focus on the visual navigation based on learning systems, mainly including reinforcement learning and deep reinforcement learning. Finally, we examine several challenges and promising directions discussed and concluded in related research of learning systems in the era of computer science and robotics.

Due to the wide range of time scales involved in the ordinary differential equations (ODEs) describing chemical reaction kinetics, multidimensional numerical simulation of chemical reactive flows using detailed combustion mechanisms is computationally expensive. To confront this issue, this article presents an economic data-driven tabulation algorithm for fast combustion chemistry integration. It uses the recurrent neural networks (RNNs) to construct the tabulation from a series of current and past states to the next state, which takes full advantage of RNN in handling long-term dependencies of time series data. The training data are first generated from direct numerical integrations to form an initial state space, which is divided into several subregions by the K-means algorithm. The centroid of each cluster is also determined at the same time. Next, an Elman RNN is constructed in each of these subregions to approximate the expensive direct integration, in which the integration routine obtained from the centroid is regarded as the basis for a storing and retrieving solution to ODEs. Finally, the alpha-shape metrics with principal component analysis (PCA) are used to generate a set of reduced-order geometric constraints that characterize the applicable range of these RNN approximations. For online implementation, geometric constraints are frequently verified to determine which RNN network to be used to approximate the integration routine. The advantage of the proposed algorithm is to use a set of RNNs to replace the expensive direct integration, which allows to reduce both the memory consumption and computational cost. Numerical simulations of a H2/CO-air combustion process are performed to demonstrate the effectiveness of the proposed algorithm compared to the existing ODE solver.

Decomposing data matrix into low-rank plus additive matrices is a commonly used strategy in pattern recognition and machine learning. This article mainly studies the alternating direction method of multiplier (ADMM) with two dual variables, which is used to optimize the generalized nonconvex nonsmooth low-rank matrix recovery problems. Furthermore, the minimization framework with a feasible optimization procedure is designed along with the theoretical analysis, where the variable sequences generated by the proposed ADMM can be proved to be bounded. Most importantly, it can be concluded from the Bolzano-Weierstrass theorem that there must exist a subsequence converging to a critical point, which satisfies the Karush-Kuhn-Tucher (KKT) conditions. Meanwhile, we further ensure the local and global convergence properties of the generated sequence relying on constructing the potential objective function. Particularly, the detailed convergence analysis would be regarded as one of the core contributions besides the algorithm designs and the model generality. Finally, the numerical simulations and the real-world applications are both provided to verify the consistence of the theoretical results, and we also validate the superiority in performance over several mostly related solvers to the tasks of image inpainting and subspace clustering.

In large refineries, the resource-oriented plan of crude oil is commonly required to be tractable and decomposable for practical operation scheduling, especially for large-scale scheduling. To this end, a framework of hybrid optimization based on plan decomposition (FHO/PD) is proposed, which mainly depends on evolutionary algorithms to realize the flexible decomposition from large-scale planning to scheduling and takes advantage of mathematical programming to improve the solving efficiency synchronously. Finally, the experimental results on a practical case suggest that the proposed method has shown great flexibility and applicability in crude oil scheduling.

Semisupervised multiview learning gains extensive research attention due to its strong capability to utilize the heterogeneous features and the label information of a few labeled samples. However, the supervision information is not well utilized in the process of exploring the consensus structure of the multiview data. In this paper, we propose a novel unified pseudolabel-guided multiview consensus (PMvC) learning framework for the semisupervised classification problem, which learns the consensus structure of multiview data by fully exploiting the supervised information of labeled samples. Specifically, PMvC first assigns multiple pseudolabels to the unlabeled samples by selecting the nearest labeled sample in each view separately, and then labels the part of unlabeled samples by selecting the pseudolabel that agrees across all views. By doing so, the high-confident pseudolabeled samples can be selected to enlarge the labeled sample pool and the supervision information can be exploited further in the learning process. In addition, to capture the consensus structure of the multiview data, PMvC learns a consensus graph from the view-specific self-representation graph guided by enhanced supervision information, which better preserves the manifold structure of samples. Meanwhile, the label information is also propagated from the labeled samples to the unlabeled samples by the learned consensus graph simultaneously. Accordingly, an effective optimization algorithm is derived to find the optimal solution for PMvC. Extensive experiment results on several real-world data sets demonstrate the feasibility and superiority of PMvC. The source code of PMvC is available at https://github.com/justcallmewilliam/PMvC.

Multiple empirical kernel learning (MEKL) is a scalable and efficient supervised algorithm based on labeled samples. However, there is still a huge amount of unlabeled samples in the real-world application, which are not applicable for the supervised algorithm. To fully utilize the spatial distribution information of the unlabeled samples, this paper proposes a novel semi-supervised multiple empirical kernel learning (SSMEKL). SSMEKL enables multiple empirical kernel learning to achieve better classification performance with a small number of labeled samples and a large number of unlabeled samples. First, SSMEKL uses the collaborative information of multiple kernels to provide a pseudo labels to some unlabeled samples in the optimization process of the model, and SSMEKL designs pseudo-empirical loss to transform learning process of the unlabeled samples into supervised learning. Second, SSMEKL designs the similarity regularization for unlabeled samples to make full use of the spatial information of unlabeled samples. It is required that the output of unlabeled samples should be similar to the neighboring labeled samples to improve the classification performance of the model. The proposed SSMEKL can improve the performance of the classifier by using a small number of labeled samples and numerous unlabeled samples to improve the classification performance of MEKL. In the experiment, the results on four real-world data sets and two multiview data sets validate the effectiveness and superiority of the proposed SSMEKL.

In the Few-Shot Learning (FSL), much of the related efforts only rely on the few available labeled samples (support set) building approach. However, the challenge is that the support set is easy-to-be-biased, so that they cannot be competent prototypes and are hard to represent the class distribution, leading to performance bottlenecks. In this paper, we propose to solve this obstacle by capturing the distribution of the unlabeled samples (query set). We propose two sampling methods: DeepSearch ($\cal DS$) and WideSearch ($\cal WS$). Both approaches are simple to implement and have no trainable parameters. They search the query samples near to the support set in different manners. Afterward, the statistic information is calculated, and we generate the latent samples according to it. The generated latent set is promising. First, it brings the query set distribution information to the classifier, which significantly improves the performance of the cross-entropy-based classifier. Second, it helps the support set become the better prototypes, which boosts the performance of the prototype-based classifier. Third, we find few latent samples are enough to boost the performance. Abundant experiments prove the proposed method achieves state-of-the-art performance on the few-shot tasks. Finally, rich ablation studies explain the compelling details of our approach.

The multi-label image classification problem is one of the most important problems in the field of computer vision, which needs to predict and output all the labels in an image. Multiple labels to be classified in an image increases the difficulty of image classification, and multi-label image classification usually requires additional attention to the positions of the object with different scales and poses. Hence, how to use the dependency relationship between labels to improve the recognition accuracy is an important problem when the object is difficult to directly identify. In this paper, we propose a designed network called the Semantic Supplementary Network with Prior Information (SSNP) to address this problem. The proposed SSNP first generates prior information by using a prior information network with different convolutional layers. Then the semantic supplementary module generates semantic information of the potential labels that is highly relevant to the current information based on the prior information, thereby effectively using the dependency relationship between the labels to improve the classification accuracy. Different from existing methods which pay more attention to the image feature extraction process, we focus on the impact of high-level semantic information generated after feature extraction on the results and tap the potential of high-level semantic information through a semantic supplementary module to strengthen the potential dependence between labels. Experimental results on public benchmark datasets demonstrate that the proposed architecture achieves the state-of-the-art performance, especially when predicting some semantically dependent labels.

In recent years, most of the studies have shown that the generalized iterated shrinkage thresholdings (GISTs) have become the commonly used first-order optimization algorithms in sparse learning problems. The nonconvex relaxations of the$\ell _{0}$-norm usually achieve better performance than the convex case (e.g.,$\ell _{1}$-norm) since the former can achieve a nearly unbiased solver. To increase the calculation efficiency, this work further provides an accelerated GIST version, that is, AGIST, through the extrapolation-based acceleration technique, which can contribute to reduce the number of iterations when solving a family of nonconvex sparse learning problems. Besides, we present the algorithmic analysis, including both local and global convergence guarantees, as well as other intermediate results for the GIST and AGIST, denoted as (A)GIST, by virtue of the Kurdyka-Łojasiewica (KŁ) property and some milder assumptions. Numerical experiments on both synthetic data and real-world databases can demonstrate that the convergence results of objective function accord to the theoretical properties and nonconvex sparse learning methods can achieve superior performance over some convex ones.

The skewed distribution of data brings difficulties to classify minority and majority samples in the imbalanced problem. The balanced bagging randomly undersampes majority samples several times and combines the selected majority samples with minority samples to form several balanced subsets, in which the numbers of minority and majority samples are roughly equal. However, the balanced bagging is the lack of a unified learning framework. Moreover, it fails to concern the connection of all subsets and the global information of the entire data distribution. To this end, this article puts several balanced subsets into an effective learning framework with a criterion function. In the learning framework, one regularization term called$R_{S}$establishes the connection and realizes the collaborative learning of all subsets by requiring the consistent outputs of the minority samples in different subsets. Besides, another regularization term called$R_{W}$provides the global information to each basic classifier by reducing the difference between the direction of the solution vector in each subset and that in the entire dataset. The proposed learning framework is called globalized multiple balanced subsets with collaborative learning (GMBSCL). The experimental results validate the effectiveness of the proposed GMBSCL.

Conventional multiobjective optimization algorithms (MOEAs) with or without preferences are successful in solving multi- and many-objective optimization problems. However, a strong hypothesis underlying their performance is that MOEAs are able to find a representative solution set to cover the entire Pareto-optimal front (PF) and decision makers are able to conveniently and precisely articulate their preference, which is not always easy to fulfill in practice. Accordingly, it is suggested that representative solutions in the naturally interesting regions of the PF rather than the whole PF should be targeted. A large body of research has been proposed to search or identify the knees or knee regions over the past decades. Therefore, this article aims to provide a comprehensive survey of the research on knee-oriented optimization. We start with a discussion of the importance and basic concepts of the knees, followed by a summary of knee-oriented benchmarks and indicators. After that, knee-oriented frameworks and techniques, and real-world applications are presented. Finally, potential challenges are pointed out and a few promising future lines of research are suggested. The survey offers a new perspective to develop MOEAs for solving multi- and many-objective optimization problems.

Multiagent systems (MASs) provide an effective means for coordinating spatially distributed and networked agents (or nodes, subsystems) such that the desired cooperative tasks can be accomplished with promising reliability, manipulability, scalability, and efficiency. One key issue in the study of MASs is the design of distributed cooperative control protocol and algorithm that depend on only local and real-time information exchanges among interacting agents over networks. However, network-enabled information sharing and increasing connectivity in practical MASs present several attack factors for malicious adversaries, thereby rendering secure control of MASs fundamentally significant. This article provides a brief survey of systems and control technologies that have been available for addressing different secure control problems of MASs in the face of various malicious attacks. First, attacks on MASs are classified based on different configuration layers. Then, the existing attack models and strategies on communication layer and agent layer are systematically examined, respectively. Furthermore, some typical secure control techniques for MASs that have been employed to handle these attacks are surveyed. Finally, several challenging issues are envisioned for potential future research.

In model-based strategies, such as real-time optimization (RTO) and model predictive control (MPC), the process model plays an important role. However, model-plant mismatch is inevitable because of various uncertainties in the real application. Therefore, it is highly desirable to detect the mismatch online with measurements of the closed-loop system. This article proposes a mismatch detection approach based on a Gaussian process (GP) model, which is integrated with RTO and MPC to improve economic performance. The GP model is obtained by training a dataset, including parameters and measurements. An ideal system with no model-plant mismatch is designed to compare to the real control system having mismatch. The outputs error of the two systems is used to build a feedback mechanism to adjust the estimation. The re-estimated parameters are used to update the model RTO and MPC. The proposed approach is nonintrusive and can be used in multiinput–multioutput nonlinear systems. Based on the benchmark model of a continuous-stirred tank reactor, a detailed 4-in-4-out nonlinear model of the Fischer esterification reaction is provided to demonstrate the performance of the proposed approach. The result shows that the relative biases of estimates decrease from 5 to 1%, and the economic profit increase by 1.5%

Multi-instance learning (MIL) has garnered considerable attention in recent years due to its favorable performance in various scenarios. Nonetheless, most previous studies have implicitly expressed the correlation between instances and bags. Moreover, the importance of negative instances has been largely overlooked. Hence, we seek to present an explicit and intuitively understandable method that can compensate for these deficiencies. In this article, we creatively introduce a metric-based multiconcept MIL approach based on two aspects. First, the triplet-based bag embedding method identifies instance categories and builds attention weights for every instance explicitly. Accordingly, bag embedding is accomplished under the limitation of weak supervision. Second, the developed instance correlation metric approach in the superbag considers the multiconcept issue to boost the model generalization performance. We have designed a rich variety of experiments to demonstrate the performance of our algorithm. The artificial data experiment reveals the interpretability of the proposed network. The results of the comparison experiment confirm that our method shows favorable performance in multiple tasks. Finally, we illustrate the motivation of the presented method by the ablation experiments.

Learning over massive data stored in different locations is essential in many real-world applications. However, sharing data is full of challenges due to the increasing demands of privacy and security with the growing use of smart mobile devices and Internet of thing (IoT) devices. Federated learning provides a potential solution to privacy-preserving and secure machine learning, by means of jointly training a global model without uploading data distributed on multiple devices to a central server. However, most existing work on federated learning adopts machine learning models with full-precision weights, and almost all these models contain a large number of redundant parameters that do not need to be transmitted to the server, consuming an excessive amount of communication costs. To address this issue, we propose a federated trained ternary quantization (FTTQ) algorithm, which optimizes the quantized networks on the clients through a self-learning quantization factor. Theoretical proofs of the convergence of quantization factors, unbiasedness of FTTQ, as well as a reduced weight divergence are given. On the basis of FTTQ, we propose a ternary federated averaging protocol (T-FedAvg) to reduce the upstream and downstream communication of federated learning systems. Empirical experiments are conducted to train widely used deep learning models on publicly available data sets, and our results demonstrate that the proposed T-FedAvg is effective in reducing communication costs and can even achieve slightly better performance on non-IID data in contrast to the canonical federated learning algorithms.

Optimal iterative learning control (OILC) has been recognized as an excellent model-based means for regulating batch process with abundant successful applications reported in the past decades but also received considerable criticisms for its poor robustness against model mismatch that is common for many industrial situations. Despite numerous attempts to address the issue, many of them are still not able to yield satisfactory control performance particularly in the presence of a possible combination of time-varying uncertainties and conservatively designed controllers, which may compromise the learning mechanism, hence rendering the robustness issue of OILC far from well explored. This article intends to investigate the aforementioned issue by proposing a new OILC method resting upon the minimization of a dynamic upper bound on tracking error which is distilled from better exploitation of the time variation of uncertainties. We also show that the problem can be formulated in the framework of convex–concave game that can be efficiently solved by a subgradient method with an excellent balance of optimality and computation time. Such a formulation enables us to gain: 1) guaranteed monotonic convergence on tracking error; 2) remarkably reduced conservatism on controller synthesis; and 3) controllable computation complexity. It is further shown that the proposed method is capable of handling nonlinearity, for example Volterra system, a classic representation of nonlinear process. The efficacy of the method is verified by numerical experiments on a continuous stirred tank reactor model.

Sampling method is one of the most commonly used techniques in dealing with imbalanced data. Most of the existing undersampling methods randomly select samples from negative class with replacement. However, it may lose some important information of the training data. Moreover, increasing the positive data by oversampling in high imbalanced situations may cause the overlapping problem. To overcome these problems, this paper proposes a hybrid sampling method. The method takes the distributions of the training data into consideration by the information entropy, thus distinguishing the important samples in the undersampling procedure. Meanwhile, since the positive data only extend to the size of each subset of the negative class in the oversampling, the overlapping problem is relieved. Further, the method retains all the data in the training procedure and generates various data views from the original training data. Then each view is handled with an individual basic classifier. Finally, all the basic classifiers are combined by the ensemble method. The newly proposed method is named as Entropy-based Hybrid Sampling Ensemble Learning (EHSEL). In addition, the EHSEL is applied to three different kinds of basic classifiers to validate its robustness. Experiments results show the great effectiveness of the EHSEL on real-world imbalanced data sets.

This article presents a novel target detection method, which is named as circular target feature detection framework based on a deep convolutional neural network (DCNN). The central proposition of this method uses the optimized DCNN architecture to detect the target and locate the position of the circle accurately in the image field of view. In this article, a Hough transform based on threshold processing (HTP) is embedded into the optimized DCNN architecture, which calculates the center positions and radius of all circles by training the circular samples for each detected rectangular frame. It can efficiently identify small circular target materials in the industry and screen out unqualified particles. The experimental results show that the boundary information of the circles is obtained clearly from the complex noise background images, thereby accurately determining the location of the circle. It has some advantages over only using a specific circular recognition algorithm. We proposed the new study on HTP-DCNN, which has extremely high accuracy in the field of machine vision positioning with circles for industrial applications.

Based on the current research on EEG emotion recognition, there are some limitations, such as hand-engineered features, redundant and meaningless signal frames and the loss of frame-to-frame correlation. In this paper, a novel deep learning framework is proposed, named the frame-level distilling neural network (FLDNet), for learning distilled features from the correlations of different frames. A layer named the frame gate is designed to integrate weighted semantic information on multiple frames to remove redundant and meaningless signal frames. A triple-net structure is introduced to distill the learned features net by net to replace the hand-engineered features with professional knowledge. Specifically, one neural network is normally trained for several epochs. Then, a second network of the same structure will be initialized again to learn the extracted features from the frame gate of the first neural network based on the output of the first net. Similarly, the third net improves the features based on the frame gate of the second network. To utilize the representation ability of the triple neural network, an ensemble layer is conducted to integrate the discriminative ability of the proposed framework for final decisions. Consequently, the proposed FLDNet provides an effective method for capturing the correlation between different frames and automatically learn distilled high-level features for emotion recognition. The experiments are carried out in a subject-independent emotion recognition task on public emotion datasets of DEAP and DREAMER benchmarks, which have demonstrated the effectiveness and robustness of the proposed FLDNet.

The wide range of time scales in chemical reaction systems has become an important problem in reactive flow simulations. This work proposes an intelligent time-scale operator-splitting (OS) chemistry integration method, which is effective in reduction of numerical stiffness and model complexity. Different from most existing publications, a pretrained backpropagation neural network is used to identify the slow and fast reactions and detect the sources of model stiffness on the fly, which replaces the expensive eigendecomposition of Jacobian matrix. With the fast-slow decomposition, the chemical source term can be represented as the sum of a stiff part and a nonstiff part. A stable time-scale OS integration is performed to solve the stiff chemical ordinary differential equations, which balances the computational cost with accuracy. In the simulation, a favorable comparison of the proposed integration method with the existing ODE solvers, such as implicit Euler, explicit Euler, and Runge-Kutta, is included to show its effectiveness and merits.

The classification on imbalanced data sets is a great challenge in machine learning. In this paper, a geometric structural ensemble (GSE) learning framework is proposed to address the issue. It is known that the traditional ensemble methods train and combine a series of basic classifiers according to various weights, which might lack the geometric meaning. Oppositely, the GSE partitions and eliminates redundant majority samples by generating hyper-sphere through the Euclidean metric and learns basic classifiers to enclose the minority samples, which achieves higher efficiency in the training process and seems easier to understand. In detail, the current weak classifier builds boundaries between the majority and the minority samples and removes the former. Then, the remaining samples are used to train the next. When the training process is done, all of the majority samples could be cleaned and the combination of all basic classifiers is obtained. To further improve the generalization, two relaxation techniques are proposed. Theoretically, the computational complexity of GSE could approach O(nd log(nmin) log(nmaj)). The comprehensive experiments validate both the effectiveness and efficiency of GSE.

Near-infrared (NIR) spectroscopy is widely used to estimate product quality and other key variables. Eliminating redundant variables is very important in constructing a high-quality NIR model. This article proposes a new wavelength-selection method for NIR spectroscopy based on joint mutual information and weighted bootstrap sampling (WBS). The method considers the combination effect of variables and involves the dynamic selection of wavelength in each iteration to increase the model prediction accuracy. The index based on joint mutual information is used to determine the importance of variables and thus accurately reflects the variable-combination effect. WBS is further used to dynamically adjust the importance of candidate variables, i.e., to increase the weights of samples with poor prediction results and decrease those of samples with accurate prediction. This process ensures that the subsequently selected wavelength focuses on inaccurately estimated samples. The performance of this method is demonstrated through three NIR datasets of gasoline, shootout, and diesel fuels. The proposed method is found to have better accuracy than the traditional partial-least-squares method, variable iterative space shrinkage approach, and several other wavelength-selection methods.

This article presents a secure communication scheme based on the quantized synchronization of master-slave neural networks under an event-triggered strategy. First, a dynamic event-triggered strategy is proposed based on a quantized output feedback, for which a quantized output feedback controller is formed. Second, theoretical criteria are derived to ensure the bounded synchronization of master-slave neural networks. With these criteria, an explicit upper bound is given for the synchronization error. Sufficient conditions are also provided on the existence of quantized output feedback controllers. A Chua's circuit is chosen to illustrate the effectiveness of our theoretical results. Third, a secure communication scheme is presented based on the synchronization of master-slave neural networks by combining the basic principle of cryptology. Then, a secure image communication is studied to verify the feasibility and security performance of the proposed secure communication scheme. The impact of the quantization level and the event-triggered control (ETC) on image decryption is investigated through experiments.

This paper tackles the high-dimensional robust order scheduling problem. A multi-objective evolutionary algorithm called constrained nondominated sorting differential evolution based on decision variable classification is developed to search for robust order schedules. The decision variables are classified into highly and weakly robustness-related variables according to their contributions to the robustness of candidate solutions. The experimental results reveal that the performance of robust evolutionary optimization can be greatly improved via analyzing the properties of decision variables and then decomposing the high-dimensional robust optimization problem. It is also unveiled that the order scheduling is greatly affected by the uncertain daily production quantities. The robust order schedules are able to provide more information on earliness/tardiness of the orders, which enhances the flexibility of the production.

Modern power grid is increasingly integrated with battery energy storage systems (BESSs). This paper deals with the problem of state-of-charge (SoC) balance control for multiple distributed BESSs in smart grid. The BESSs are expected to work cooperatively to not only fulfil the overall power requirement but also meet the constraints of the same relative SoC variation rate. To achieve this objective, a distributed SoC balance control approach is presented with event-triggered signal transmissions. It is designed with the dynamic average consensus (DAC) mechanism for parameter estimations. The DAC enables distributed control of each BESS through communicating with its neighboring BESSs. Different from traditional periodic signal transmission, the event-triggered signal transmission embedded in our approach allows each BESS to transmit signal to its neighboring BESSs only when needed, thus reducing the communication traffic. Theoretical lower bounds are established for consecutive interevent intervals such that the Zeno behavior is excluded. Case studies are conducted to demonstrate the effectiveness of the presented approach.

An ethylene plant typically consists of multiple cracking furnaces in parallel to process various feeds. For tackling this problem, it is convenient to formulate it as a cyclic scheduling problem that can be modeled as a large-scale mixed integer optimization nonlinear programming (MINLP). However, due to the existence of mixed variables and many constraints, the problem is hard to be solved efficiently by conventional deterministic algorithms or stochastic algorithms. To solve this problem, we propose a novel two-level optimization framework based on real-coded genetic algorithms (GA) and sequential quadratic programming (SQP). Our approach is based on reformulating the MINLP as a nested optimization with two loops. In the outer layer, to avoid wasting computation time, the GA is used first to filter out infeasible integer solution candidates and pass the feasible ones to the inner loop for fitness evaluation. In the inner loop, by fixing feasible integer solutions, the problem is simplified to a nonlinear programming problem (NLP), which is then solved by the SQP algorithm. A real-world case study demonstrates the efficacy of the developed methodology compared with existing MINLP solvers.

This study focuses on the classification and pathological status monitoring of hyper/hypo-calcemia in the calcium regulatory system. By utilizing the Independent Component Analysis (ICA) mixture model, samples from healthy patients are collected, diagnosed, and subsequently classified according to their underlying behaviors, characteristics, and mechanisms. Then, a Just-in-Time Learning (JITL) has been employed in order to estimate the diseased status dynamically. In terms of JITL, for the purpose of the construction of an appropriate similarity index to identify relevant datasets, a novel similarity index based on the ICA mixture model is proposed in this paper to improve online model quality. The validity and effectiveness of the proposed approach have been demonstrated by applying it to the calcium regulatory system under various hypocalcemic and hypercalcemic diseased conditions.

The Success History-based Adaptive Differential Evolution algorithm with Linear Population Size Reduction (L-SHADE) is a highly competitive continuous optimizer. In this paper, we extend L-SHADE towards handling mixed variable optimization problems, where some of the variables may be categorical, ordinal, integer or continuous. For handling the discrete variables, we consider three different options. The first is to adapt the mechanism for handling categorical variables introduced in an earlier proposed ant colony optimization algorithm, called ACOMVto use it with L-SHADE (L-SHADEMV). The second is to apply a simpler scheme that relies on the mechanisms how ant colony optimization algorithms traditionally tackle discrete problems (L-SHADEACO). The third is to introduce a mechanism that exploits the ideas of DE within the generation of new values for the categorical variables, by introducing specific rules to be applied in the mutation (L-SHADERULE). We use an automatic algorithm configuration tool for optimizing the algorithms' parameters to allow an unbiased comparison of the three schemes. The performance of the proposed algorithms was evaluated on artificial mixed-variable optimization problems and compared with L-SHADE that incorporates a simple rounding scheme to handle categorical variables, which is used as a baseline method. The experimental results show that all three categorical handling methods outperform the baseline and L-SHADEACOis the best performing scheme in terms of accuracy and robustness.

An online monitoring method, which aims to deal with the high order non-Gaussian characteristics in chemical process, is proposed in this paper. In the framework of the proposed method, kernel based independent component analysis is utilized to identify the operation status of the chemical process and statistics pattern analysis is employed to combine with independent component analysis to extract high order information from the process so as to improve the monitoring performance. The modified statistics pattern analysis introduces the Mahalanobis distance into statistics pattern to analyze the inner structure relationship between the samples. Then, the validity and effectiveness of our proposed method is illustrated by applying to a representative non-Gaussian process, Continuous Stirred Tank Reactor (CSTR). The results show that the proposed method has its advantages when compared to other conventional Eigen-decomposition monitoring algorithms.

In this brief, we propose a new data-driven model-free adaptive control (MFAC) method with dual radial basis function neural networks (RBFNNs) for a class of discrete-time nonlinear systems. The main novelty lies in that it provides a systematic design method for controller structure by the direct usage of I/O data, rather than using the first-principle model or offline identified plant model. The controller structure is determined by equivalent-dynamic-linearization representation of the ideal nonlinear controller, and the controller parameters are tuned by the pseudogradient information extracted from the I/O data of the plant, which can deal with the unknown nonlinear system. The stability of the closed-loop control system and the stability of the training process for RBFNNs are guaranteed by rigorous theoretical analysis. Meanwhile, the effectiveness and the applicability of the proposed method are further demonstrated by the numerical example and Aspen HYSYS simulation of distillation column in crude styrene produce process.

This paper is concerned with the synchronization of multiagent systems connected via different types of interactions, known as multilayer networks. Additive coupling and Markovian switching coupling are proposed to capture the layered connections with two kinds of mathematical models constructed. First, based on simultaneously diagonalization of multiple Laplacian matrices, a general criterion is derived, ensuring that the synchronization problem with additive coupling can be decoupled. Then, an alternative condition is presented, which is related to the number of layers, regardless of the number of agents. With the derived criteria, a concept of joint synchronization region is introduced and further discussed as a network design problem. Synchronization with Markovian switching layers is also analyzed in parallel, exemplified by some special cases of two-layer networks. Finally, a group of cellular neural networks coupled by two-layer connections are chosen to illustrate the effectiveness of the theoretical results.

This paper proposes a novel scheme of nonuniform discretizetion-based control vector parameterization (ndCVP, for short) for dynamic optimization problems (DOPs) of industrial processes. In our ndCVP scheme, the time span is partitioned into a multitude of uneven intervals, and incremental time parameters are encoded, along with the control parameters, into the individual to be optimized. Our coding method can avoid handling complex ordinal constraints. It is proved that ndCVP is a natural generalization of uniform discretization-based control vector parameterization (udCVP). By integrating ndCVP into hybrid gradient particle swarm optimization (HGPSO), a new optimization method, named ndCVP-HGPSO for short, is formed. By application in four classic DOPs, simulation results show that ndCVP-HGPSO is able to achieve similar or even better performances with a small number of control intervals; while the computational overheads are acceptable. Furthermore, ndCVP and udCVP are compared in terms of two situations: given the same number of control intervals and given the same number of optimization variables. The results show that ndCVP can achieve better performance in most cases.

This study proposes an online monitoring technique for nonlinear multiple-mode problems in industrial processes. The contributions of the proposed technique are summarized as follows: 1) Lazy learning (LL), a new adaptive local modeling method, is introduced for multiple-mode process monitoring. In this method, multiple modes are separated and accurately modeled online, and the between-mode dynamic process is considered. 2) The modified receptor density algorithm (MRDA) exhibiting superior nonlinear ability is introduced to analyze the residuals between the actual system output and the model-predicted output. The simulation of the Tennessee Eastman process with multiple operation modes shows that compared with other techniques mentioned in this study, the proposed technique performs more accurately and is more suitable for nonlinear processes with multiple operation modes.