Search

Cavity flame holder is a well-organized method for fuel mixing inside the combustion chamber of supersonic vehicles. In this study, the combined machine learning technique of proper orthogonal decomposition (POD) combined with Long Short-Term Memory network (LSTM) is used for prediction of fuel jet penetration inside the cavity flame holder at free stream Mach=2.2 is investigated. Computational Fluid Dynamic is applied for the three-dimensional model of a cavity flame holder with extruded multi-nozzles. A comparison of single and extruded multi-nozzles for the fuel mixing at the combustor with cavity configuration is also done. The presented results show that the use of multi-jet efficiently increases the fuel mixing inside the cavity since the vortex structure is enhanced in this configuration. The usage of a shock creator pushes the shear layer into the cavity and limits the fuel penetration inside the cavity. Therefore, the shock creator is not recommended for the fuel mixing of extruded multi-jet injection systems. [ABSTRACT FROM AUTHOR]

The article discusses predictors of congenital cytomegalovirus (CMV) infection outcomes, highlighting a good prognosis if the mother's infection occurred after the first trimester and if the newborn has normal hearing, platelet count, and head ultrasound. Topics include predictors of CMV sequelae at 2 years, such as maternal infection timing, neonatal symptoms, and early diagnostic tests like cranial ultrasound.

The path planning and tracking control problem of autonomous vehicle with model matched uncertainty and external disturbance is studied in this paper. A safe lane change zone is designed by considering lane change time, distance, and driving speed. Meanwhile, the candidate paths are generated using cubic quasi-uniform B-spline curve, and the optimal path is selected based on feasibility, comfort, and efficiency criteria. The proposed planning method ensures an efficient and comfortable lane change path without collision with surrounding vehicles. Model uncertainty and external disturbance may affect lane change tracking control of autonomous vehicle. Hence, a fault-tolerant path tracking controller is implemented by using deep neural networks and adjustable zonotope-tube model predictive control. A deep neural network-based controller is proposed to compensate for model matched uncertainty, thereby enhancing the fault tolerance of the control system. Then, an adjustable zonotope-tube model predictive controller is designed to enclose the actual trajectory within a zonotope-tube with the nominal state as the center and the disturbance set as the radius by using the feedback control to achieve the asymptotic stabilization of the closed-loop system. The effectiveness of this path planning and tracking control method is finally verified by numerical simulation. • A lane change path planning method suitable for structured roads is proposed, which is based on the utilization of a safe lane change zone and cubic quasi-uniform B-spline curves. The proposed method ensures an efficient and comfortable lane change path. • A fault-tolerant path tracking controller is presented for autonomous vehicle, which incorporates a deep neural network and adjustable zonotope-tube model predictive control. The controller effectively handles model uncertainty and external disturbances. • The computation of robust positive invariant set is performed using the adjustable zonotope-tube method, which enhances scalability and reduces computational cost. [ABSTRACT FROM AUTHOR]

We consider the inherent attack tolerance properties of resilient model predictive control (R-MPC) for cyber–physical systems (CPSs) modeled by discrete linear time-invariant (LTI) systems subjected to limited disruptions. In this paper, the relationship between the maximum allowable duration of Denial-of-Service (DoS) attacks, namely attacks that disrupt both sensor to controller (S–C) and controller to actuator (C–A) communication channels, and the upper bound of disturbances is deduced at length. Moreover, to achieve robust recursive feasibility and closed-loop stability of the system, we discuss that MPC with state, control and terminal constraints has a certain degree of inherent attack tolerance concerning the DoS attack duration and parameters like system matrices and the nominal terminal set. Finally, numerical simulation is given to substantiate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

The solar thermal collectors are widely used for several applications nowadays. The improvement of their thermal performance helps to mitigate the initial costs and sizes of these components. In addition, the robust prediction of the solar thermal collectors' performance is a promising technique to eliminate the future numerical simulations of these devices. To this end, the present numerical research is performed to investigate how placing two similar twisted tape turbulators side by side inside the absorber tube of a solar water heater affects the second-law aspects of the device. The pattern of changes in outlet water temperature, useful thermal exergy (Q u,ex), and rate of entropy production due to heat transfer (S g , t h) and fluid flow (S g , f r) in different values of Reynolds number (Re) and twist pitch (H) were presented and analyzed. Validation of the numerical method was done with the experimental results available in the literature. The negative effect of boosting the Reynolds number and twist pitch on the entropy was reported. In addition, the downward (9.31–11.43 %) and upward (2.51–4.97 %) influence of Re and H on the S g , t h were reported. Moreover, it was found that the growth in Re and H is associated with the growth (654.97–661.27 %) and reduction (2.39–3.20 %) of S g , f r. Finally, the nonlinear calibration using the Chaos Control Method (CCM) was employed to determine a predictive model for the Q u,ex in terms of Re and H. The results demonstrated that two predictive models can estimate Q u,ex with high degree of accuracy with the correlation coefficient of R2=0.999 and RMSE of 0.089. [ABSTRACT FROM AUTHOR]

The control input derived from model predictive control (MPC) scheme depends on the receding horizon whose length needs to be determined. Design length is also a trade-off between computational complexity and optimization metrics. To this end, a framework for adaptive selection of the horizon length has been developed for discrete systems in previous works. In this paper, this framework is extended to continuous nonlinear systems. A new algorithm is presented to solve the problem of how to adaptively select the horizon length. Then, according to the traditional method and the robust MPC scheme based on tubes, the recursive feasibility and robustness of the algorithm (the adaptive horizon nonlinear model predictive control, called AH-NMPC) for quasi-infinite time domain adaptive horizontal continuous-time nonlinear systems are proved. Finally, the NMPC controller is used to perform path-tracking experiments on an autonomous vehicle to verify the good effects of the algorithm (AH-NMPC). [ABSTRACT FROM AUTHOR]

This paper is concerned with model predictive control (MPC) for multisensor linear systems. In order to fully utilize information and achieve higher control accuracy, two fusion MPC algorithms are proposed. The first algorithm is named estimation-fusion model predictive controller (E-F MPC), which local estimations are fused and then fusion control input is obtained by the fusion estimation. The second algorithm is named control-fusion model predictive controller (C-F MPC), which fuses the control inputs obtained from local systems. The relationships, differences and complexity of the two algorithms are analyzed. The advantages of E-F MPC are that it is simple and easy to implement with high control accuracy, while the disadvantage is the bad robustness because of the heavy estimation, prediction and control tasks for the fusion node. For C-F MPC, with a small loss of accuracy, network throughput and computational cost of its fusion node will be greatly reduced, which makes it good robustness and flexibility. Simulation analysis verifies the effectiveness and correctness of the conclusion. [ABSTRACT FROM AUTHOR]

We present a dynamic output feedback fuzzy model predictive control (FMPC) approach for the Takagi–Sugeno model with bounded disturbance in this paper. The proposed approach relies on the formulation, at each sampling instant, of a lexicographic optimization problem that takes into account multiple objectives including improving the control performance, enlarging the region of attraction (ROA) and tightening the estimation error set (EES). Since the controller parameters and Lyapunov matrices are kept as degrees of freedom for the lower-ordered optimization problems, larger ROA and less conservative EES are obtained. As a result, the control performance can be greatly improved with respect to previous works. Furthermore, we prove the recursive feasibility of the approach and guarantee the stability of the closed-loop system. The result is verified through a numerical example. [ABSTRACT FROM AUTHOR]

Automotive coatings often suffer from high-speed stone impact wear damage. Currently, it remains a significant challenge to evaluate the stone-chipping resistance performance of automotive coatings. The main purpose of this work is to develop a computational approach combining a numerical model with a prediction method to achieve this end. The numerical model is developed based on the Computational Fluid Dynamics (CFD) and the Discrete Element Method (DEM) to account for the moving behaviors of large, irregularly shaped stone chips driven by high-pressure airflow in a multifunctional anti-impact tester. In this model, stone chips are represented using multi-spheres. To address the mismatch issue in the inlet boundary conditions, this study utilizes high-fidelity modeling of the tee pipe in a multi-functional impact tester. Additionally, it takes into account the influence of the high-pressure airflow compressibility on particle behaviors. A wear prediction method is further developed. Within this method, a particle erosion model is utilized to account for short-term coating damage evaluation based on CFD-DEM simulation results. Subsequently, a prediction method is proposed to efficiently estimate long-term damage for a DIN standard multi-particle impact case by using mean-range random functions. With the aid of the developed computational method, single- and multi-particle impact simulations are conducted, and numerical results are found in good agreement with experimental data in terms of the damage distribution trend and impact position distributions. [ABSTRACT FROM AUTHOR]

Purpose: In a boundary layer ingestion (BLI) propulsion system, the fan operates continuously under distorted inflow conditions, leading to an increment of aerodynamic loss and in turn impacting the potential fuel burn reduction of the aircraft. Usually, in the preliminary design stage of a BLI propulsion system, it is essential to assess the impact of fuselage boundary layer fluids on fan aerodynamic performances under various flight conditions. However, the hub region flow loss is one of the major loss sources in a fan and would greatly influence the fan performances. Moreover, the inflow distortion also results in a complex and highly nonlinear mapping relation between loss and local physical parameters. It will diminish the prediction accuracy of the commonly used low-fidelity computational approaches which often incorporate traditional physics-based loss models, reducing the reliability of these approaches in evaluating fan performances. Meanwhile, the high-fidelity full-annulus unsteady Reynolds-averaged Navier–Stokes (URANS) approach, even though it can give rather accurate loss predictions, is extremely time-consuming. This study aims to develop a fast and accurate hub loss prediction method for a BLI fan under distorted inflow conditions. Design/methodology/approach: This paper develops a data-driven hub loss prediction method for a BLI fan under distorted inflows. To improve the prediction accuracy and applicability, physical understandings of hub flow features are integrated into the modeling process. Then, the key physical parameters related to flow loss are screened by conducting a sensitivity analysis of influencing parameters. Next, a quasi-steady assumption of flow is made to generate a training sample database, reducing the computational time by acquiring one single sample from the highly time-consuming full-annulus URANS approach to a cost-efficient single-blade-passage approach. Finally, a radial basis function neural network is used to establish a surrogate model that correlates the input parameters and the output loss. Findings: The data-driven hub loss model shows higher prediction accuracy than the traditional physics-based loss models. It can accurately capture the circumferentially and radially nonuniform variation trends of the losses and the associated absolute magnitudes in a BLI fan under different blade load, inlet distortion intensity and rotating speed conditions. Compared with the high-fidelity full-annulus URANS results, the averaged relative prediction errors of the data-driven hub loss model are kept less than 10%. Originality/value: The originality of this paper lies in developing a new method for predicting flow loss in a BLI fan rotor blade hub region. This method offers higher prediction accuracy than the traditional loss models and lower computational time cost than the full-annulus URANS approach, which could realize fast evaluations of fan aerodynamic performances and provide technical support for designing high-performance BLI fans. [ABSTRACT FROM AUTHOR]

This article uses the neural network method and numerical data for predicting the performance of an integrated energy system. The studied system is an integrated energy system consisting of a gas turbine with an air preheater and subsystems such as the Kalina cycle, supercritical carbon oxide cycle and hydrogen production. In this unit, a photovoltaic thermal (PVT) module has been used to generate electric power, as well as heating and starting the organic Rankine cycle. In addition, a Proton exchange membrane (PEM) electrolyzer to produce hydrogen is integrated. To propose a model that can predict the behavior of the integrated system by changing the parameters, the important outputs of the system such as energy efficiency, exergy efficiency, exergy destruction rate and net output power have been extracted in the form of functions according to the input parameters namely T 11 , P 11 , P 18 , R p. The results show that the suggested functions of four outputs have the acceptable value for mean square error and regression coefficient which show high accuracy. The influence of each variable on each other and their interaction on the responses are drawn and adequately explained using three-dimensional response surface diagrams. The analysis represented that the amount of exergy destruction in the high-pressure ratio is at its maximum value with the increase in the temperature of point 4. In addition, it is at its maximum value in the high-pressure ratio with different changes in the pressure of point 11. Generally, the p-values for predicted values less than 0.001 represent that the model is statistically considerable at the 99 % confidence level. The p-values for suggested functions show a high level of confidence. [ABSTRACT FROM AUTHOR]

This work focuses on the output tracking control for discrete-time Markov jump systems (MJSs) via a model predictive control (MPC) approach. Not only the mean-square stability (MSS), but also the H 2 performance and H ∞ performance is considered in the MPC controller design. To save communication resources, a version of adaptive event-triggered mechanism (AETM) is utilized. Different from the traditional static event-triggered mechanism, its threshold coefficient is time-varying and can be adjusted in real-time to fit the system evolution. Moreover, the asynchronous switching between the controller modes and the plant modes is taken into account. The design objective is to synthesis an adaptive event-triggered asynchronous MPC controller such that not only the augmented closed-loop system is MSS with a certain level of H 2 / H ∞ performance, but also the communication burden can be reduced to some extent. It is the first attempt to investigate the output tracking control of MJSs via an MPC approach. By constructing a plant model-dependent Lyapunov function, a set of conditions is derived to determine a desired MPC controller. In addition, two MPC related issues, namely recursive feasibility and closed-loop MSS, are also studied. At last, an illustrative example is given to verify the availability of the theoretical findings. [ABSTRACT FROM AUTHOR]

The investigation of the mixed time/event-triggered quantized model predictive security control (MPSC) for switched systems with DoS attacks is the emphasis of our study. To lessen the DoS attacks' impact, a unique structure that combines a network loop and a local loop is adopted. The operation sequence of permissible controllers coordinated by the permissible type-switching mechanism (PTM) is further designed. To handle the effect of limited bandwidth and save network communication resources, a time-varying uniform quantization rule and a mixed time/event-triggered scheme are designed. Meanwhile, a buffer is introduced in the observer, and the stored predicated observation data in which will be used to update the observer state with different weighting coefficients when DoS attacks occur. By introducing a new coefficient into the performance index, the upper bound of which is obtained. Then the model predictive control (MPC) design is transformed into an optimization problem minimizing such constant upper bound. On this basis, the sufficient conditions are deduced for the exponential convergence of closed-loop systems. Ultimately, two systems are adopted to testify the effectiveness of the proposed method, and the optimal weighting coefficients which guarantees the fast system convergence speed are obtained for the specific system. [ABSTRACT FROM AUTHOR]

In this paper, the dynamic output-feedback model predictive control (MPC) problem is investigated for systems with polytopic uncertainties and amplify-and-forward (AF) relays, where the signal transmission with random power is adopted to facilitate the remote transmission with the limited transmission capacity of sensors. With respect to the parameter uncertainties of the system and the randomness of AF relays, a new objective function is constructed, which is defined by the mathematical expectation of the conventional quadratic function over the infinite time horizon. Then, the singular value decomposition technique is applied to handle the inherent couplings of variables to formulate a solvable optimization problem. Furthermore, sufficient conditions are provided to guarantee the recursive feasibility of the proposed MPC algorithm, meanwhile the underlying system can be stabilized in the mean-square sense by the designed controllers. Finally, two simulation examples are used to show the effectiveness of the proposed MPC strategy. [ABSTRACT FROM AUTHOR]

This paper presents an additive-state-decomposition-based model predictive tracking control and disturbance rejection method for a permanent magnet synchronous motor (PMSM) servo system subject to unknown parameter perturbations, unmodeled dynamics, and time-varying load torque. The basic idea of this method is to equivalently decompose the original system into a primary system for handling the tracking control subproblem and a secondary system for dealing with the robust stabilization subproblem. A model predictive controller is designed for the primary system to achieve high-accuracy tracking of the reference speed. As for the secondary system, a novel high-order generalized extended state observer (HGESO) is constructed to estimate the multiple disturbances simultaneously, and a state feedback control law incorporating a disturbance compensator is developed to eliminate the adverse effect of the multiple disturbances on the system output. By combining the control inputs of the two subsystems together, the control objectives of the original system can be achieved. Both the stability criterion and design procedure of the closed-loop control system are developed. Finally, hardware-in-the-loop-based comparative experiments are conducted to demonstrate that the proposed method effectively suppresses the influence of the multiple disturbances on motor speed tracking accuracy and that the control system has both satisfactory dynamic performance and robustness. [ABSTRACT FROM AUTHOR]

To achieve an accurate remaining useful life (RUL) prediction for lithium-ion batteries (LIBs), this study proposes an adaptive self-attention long short-term memory (SA-LSTM) prediction model. The innovations of the designed prediction model include the following. (1) It features an optimized local tangent space alignment algorithm, which allows the extraction of an indirect health indicator (HI) that can precisely describe battery degeneration from charge data. The extracted HI exhibits a high correlation with the standard capacity, thus facilitating RUL estimation. (2) By introducing a masked multi-head self-attention module into the time-series prediction model based on LSTM, critical information in the sequences is captured and the prediction performance is improved. (3) An online self-tuning mechanism for the weights and biases of neural networks is designed to correct cumulative estimation errors in long-term predictions and reduce the effects of local fluctuations and regeneration. The proposed prediction model enables the HI values in future cycles to be iteratively estimated using the one-step-ahead method, and the RUL can be forecast once the predicted signal falls. Experimental results indicate the effectiveness and superiority of the proposed prediction method. [ABSTRACT FROM AUTHOR]

The prediction of traffic flow is a challenging yet crucial problem in spatial-temporal analysis, which has recently gained increasing interest. In addition to spatial-temporal correlations, the functionality of urban areas also plays a crucial role in traffic flow prediction. However, the exploration of regional functional attributes mainly focuses on adding additional topological structures, ignoring the influence of functional attributes on regional traffic patterns. Different from the existing works, we propose a novel module named POI-MetaBlock, which utilizes the functionality of each region (represented by Point of Interest distribution) as metadata to further mine different traffic characteristics in areas with different functions. Specifically, the proposed POI-MetaBlock employs a self-attention architecture and incorporates POI and time information to generate dynamic attention parameters for each region, which enables the model to fit different traffic patterns of various areas at different times. Furthermore, our lightweight POI-MetaBlock can be easily integrated into conventional traffic flow prediction models. Extensive experiments demonstrate that our module significantly improves the performance of traffic flow prediction and outperforms state-of-the-art methods that use metadata. • The functional attributes of various regions are utilized to produce distinct predictions. • The traffic characteristics are dynamically generated to analyze various functional attributes. • Our approach can be seamlessly integrated with traditional traffic prediction models. • An irregularly partitioned dataset is proposed. • The effect of our module is relatively strong. [ABSTRACT FROM AUTHOR]

Traffic accident prediction based on dashboard cameras is of great significance for ensuring the guaranteed-safety of self-driving systems and reducing the occurrence of accidents. Due to complexity of traffic scenarios and wide range of object's motion, early prediction of accidents remains a challenge. However, the existing methods process all traffic objects and all video frames indiscriminately, thus leading to poor prediction performance. To address this issue, we propose a novel traffic accident prediction framework, namely two-layer hidden state aggregation based two-stream network (THAT-Net). The proposed method first fuses the spatial and temporal flow to capture complementary spatio-temporal information, which filters out irrelevant objects in traffic scenes to reduce the influence of the object's motion state. Furthermore, a two-layer hidden state aggregation structure is designed to reintegrate the hidden state weights of gated recurrent units. It captures contextual information through frame-level and segment-level aggregation and decreases the influence of irrelevant frames to reduce the complexity of traffic scenes. Experiments based on two real-world datasets show the state-of-the-art performance of THAT-Net. Our proposed method achieves the highest accuracy by predicting accidents 0.48 sec to 2.8 sec earlier compared to the baseline methods in more challenging situations. Our code is available at: https://github.com/redeyezt/THAT-Net. • We propose a novel traffic accident prediction framework named THAT-Net. • We introduce the optical flow information of the video to reduce the influence of irrelevant objects on accident prediction. • We use a CTA module to reduce the impact of irrelevant frames in videos on traffic accident prediction. • Compared with other accident prediction models, THAT-Net can forecast accidents earlier and with greater precision. [ABSTRACT FROM AUTHOR]

• New scenario-based MPC strategy developed for linear systems subject to correlated parametric and/or additive uncertainties. • The proposed algorithm adapts the conditional scenario reduction method to the MPC framework. • Two numerical examples validate the higher probability of constraint satisfaction of the proposed MPC. • The high probability of constraint satisfaction of the proposed MPC allows solving a OCP with fewer scenarios when a shorter solution time is needed. This paper proposes a novel MPC approach called conditional scenario-based model predictive control (CSB-MPC), developed for discrete-time linear systems affected by parametric uncertainties and/or additive disturbances, which are correlated and with bounded support. At each control period, a primary set of equiprobable scenarios is generated and subsequently approximated to a new reduced set of conditional scenarios in which each has its respective probabilities of occurrence. This new set is considered for solving an optimal control problem in whose cost function the predicted states and inputs are penalised according to the probabilities associated with the uncertainties on which they depend in order to give more importance to predictions that involve realisations with a higher probability of occurrence. The performances of this new approach and those of a standard scenario-based MPC are compared through two numerical examples, and the results show greater probabilities of not transgressing the state constraints by the former, even when considering a smaller number of scenarios than the scenario-based MPC. [ABSTRACT FROM AUTHOR]

The main results of this paper are concentrated on the nonlinear model predictive control (MPC) tracking optimization based on high-order fully actuated (HOFA) system approaches. The proposed HOFA MPC strategy makes full use of full-actuation property to eliminate the nonlinear dynamics of the system, and then the nonlinear optimization problem is equivalently transformed into a series of easy-solve linear convex optimization problems. Different from general nonlinear MPC methods and the current optimal control of the HOFA system approach, an analytical controller with smooth and less energy is obtained by the moving horizon optimization. And it is proven that the proposed controller can stabilize the corresponding tracking error closed-loop system. Finally, not limited to FA systems, as examples, a nonlinear numerical under-actuated model in the mathematical sense and a benchmark nonlinear under-actuated mechanical system are transformed into corresponding equivalent HOFA systems, the simulation results are given to verify the effectiveness of the proposed strategy. [ABSTRACT FROM AUTHOR]

This paper explores the trajectory tracking control problem for a wheeled mobile robot (WMR) in an environment with obstacles and unknown disturbances. A fixed-time extended state observer is presented, which is utilized to estimate unknown disturbances and improve the convergence speed of estimation errors. By introducing the obstacle avoidance cost, a model predictive controller with disturbance compensation is proposed to guarantee desired tracking performance in the presence of obstacles. The proposed method is analyzed for recursive feasibility and closed-loop system stability subject to unknown disturbances and obstacles. Finally, both simulation and experiment are conducted to express the satisfactory tracking effect of the developed approach. [ABSTRACT FROM AUTHOR]

Self-driving vehicles must be equipped with path tracking capability to enable automatic and accurate identification of the reference path. Model Predictive Controller (MPC) is an optimal control method that has received considerable attention for path tracking, attributed to its ability to handle control problems with multiple constraints. However, if the data acquired for determining the reference path is contaminated by non-Gaussian noise and outliers, the tracking performance of MPC would degrades significantly. To this end, Correntropy-based MPC (CMPC) is proposed in this paper to address the issue. Different from the conventional MPC model, the objective of CMPC is constructed using the robust metric Maximum Correntropy Criterion (MCC) to transform the optimization problem of MPC to a non-concave problem with multiple constraints, which is then solved by the Block Coordinate Update (BCU) framework. To find the solution efficiently, the linear inequality constraints of CMPC are relaxed as a penalty term. Furthermore, an iterative algorithm based on Fenchel Conjugate (FC) and the BCU framework is proposed to solve the relaxed optimization problem. It is shown that both objective sequential convergence and iterate sequence convergence are satisfied by the proposed algorithm. Simulation results generated by CarSim show that the proposed CMPC has better performance than conventional MPC in path tracking when noise and outliers exist. [ABSTRACT FROM AUTHOR]

A novel robust hierarchical multi-loop composite control scheme is proposed for the trajectory tracking control of robotic manipulators subject to constraints and disturbances. The inner loop based on inverse dynamics control is used to reduce the nonlinear tracking error system to a set of decoupled linear subsystems to alleviate the computational effort during the sequel optimization. The feasible regions of the equivalent state and control input of each subsystem can be computed efficiently by choosing an appropriate inertia matrix estimate. The external loop, relying on a set of separate disturbance-observer-based tube model predictive composite controllers, is used to robustly stabilize the decoupled subsystems. In particular, the disturbance observers are designed to compensate for the disturbances actively, while the tube model predictive controllers are used to reject the residual disturbances. The robust tightened constraints are obtained by calculating the outer-bounding-tube-type residual disturbance invariant sets of the closed-loop subsystems. Furthermore, the recursive feasibility and input-to-state stability of the closed-loop system are investigated. The effectiveness of the proposed control scheme is verified by the simulation experiment on a PUMA 560 robotic manipulator. [ABSTRACT FROM AUTHOR]

Ship speed optimization can reduce fuel consumption and protect the environment. The existing offline speed optimization model cannot ensure the optimal ship speed due to the changing weather and sea condition during the voyage. We propose an online speed optimization model which consists of an offline speed optimization model and a fuel consumption prediction model. The online speed optimization model is to plan the average speed for each segment of the entire voyage. The fuel consumption prediction model is used to continuously monitor the fuel consumption during the voyage. For fuel consumption prediction, we propose a bidirectional long short-term memory (BiLSTM) based encoder-decoder with temporal attention mechanism. By the comparative experiments with three baseline models, the accuracy of the proposed prediction model is verified. [ABSTRACT FROM AUTHOR]

This paper proposes an approach for the analysis and comparison of stochastic solvers based on the statistical distribution of their variables. The observed variables of the stochastic solvers are the runtime and number of function evaluations required to reach a (sub)-optimal solution. These variables were measured using a target approach. We extended the conventional approach, which usually predicts the average value of variables, by predicting their statistical distributions. If possible, we can predict, not only the average value, but also all the values of the observed variables according to the identified distribution and given probability. We can also predict the probability of reaching (sub)-optimal solutions according to the variable's values. The approach was empirically validated by comparing solvers for discrete and continuous problems. In our experiment, the differences between predicted and measured values of runtime, number of function evaluations and probability are 10% or less. This proves that the proposed approach is useful and can be used for the analysis and comparison of stochastic solvers. Although, the approach has some limitations, it can mitigate the issue of stochastic solvers being unable to provide (sub)-optimal solutions, and can be used to determine the stopping optimization criteria. • An approach to the comparison of stochastic solvers. • The predictive modeling based on optimal and sub optimal solutions. • The analysis of the distribution of runtime and number of function evaluations. • The predictive models based on distribution of solvers' variables are established. • The predictive models provide probability of reaching the optimal solutions. [ABSTRACT FROM AUTHOR]

Multi-label learning (MLL) copes with applications such as text classification and image recognition where each instance is associated with multiple labels. Currently, some label-specific feature conversion approaches such as LIFT serve each individual label well, however do not take advantage of label correlation. Some label correlation exploitation approaches such as GLOCAL have independent model prediction and label correlation stages, thus limiting their learning ability. This paper introduces LSTC, a new algorithm that generates label-specific features and exploits third-order label correlation. On the one hand, for each label, we choose the same number of positive and negative representative instances inspired by density peak. These instances are used to generate a new data matrix with label-specific features. On the other hand, we train a paired output prediction network for each label based on these matrices of its own and that of two auxiliary labels. In this way, third-order label correlations are implicitly exploited. Specifically, the two auxiliary labels are the most similar and least similar, respectively, to avoid getting stuck in local optima. Experiments are conducted on seventeen benchmark datasets in comparison with ten popular algorithms. Results on eight measures demonstrate the superiority of LSTC on data from various domains except the text-domain. The source code is available at github.com/fansmale/lstc. • We adopt density peaks under Gaussian kernel to extract label-specific features. • We train the model for each label using the data of two auxiliary labels. • We use the softmax function to obtain numerical predictions from paired outputs. • Our algorithm shows superiority in ranking-based measures of non-text data. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the impact of optimal hyper-parameter configuration in relational topic models. The main goal is to validate the hypothesis that single-objective Bayesian Optimization (BO) can discover a hyper-parameter setting that leads a set of relational topic models to simultaneously ensure good prediction capabilities and significant topics from a qualitative perspective. Our research, as a result of a comparative analysis performed on 7 state-of-the-art models, 5 performance measures and 3 datasets, has highlighted three main findings: (1) the majority of relational topic models are not able to offer a good trade-off between classification capabilities and topic interpretability; (2) single-objective optimization of hyper-parameters, targeted on maximizing the F1-Measure, is able to create topics that are also optimal with respect to the Kullback Leibler divergence measure; (3) the Pareto frontiers across several performance metrics reveals that the most promising trade-off between the performance metrics can be obtained by Constrained Relational Topic Models. [ABSTRACT FROM AUTHOR]

Next-location prediction is a special task of the next POIs recommendation. Different from general recommendation tasks, next-location prediction is highly context-dependent: (1) sequential dependency, i.e., the sequential locations checked in by a user have high correlation; (2) temporal dependency, i.e. check-in preferences are identified as different days or nights; (3) spatial dependency, i.e. users prefer to visit closer locations. Recent studies have been very successful in predicting users' next location by comprehensively considering user preferences. Nonetheless, these methods not only fail to capture temporal dependencies but also fail to capture location topology information. To fill this gap, we propose a GNN-based model which converts POIs into a low-dimensional metric and integrates users' long-term and short-term preferences to comprehensively represent dynamic preferences. The model consists of graph neural networks for long-term preference modeling and LSTM for short-term preference modeling. Comprehensive experiments are conducted on two real-world datasets, and results demonstrate the effectiveness of our approach over state-of-the-art methods for next-location prediction. • A long-term and short-term preference learning model considering contextual information and sequence information is proposed. • The long-term preference module captures spatial dependencies using graph neural networks. • Geographic accessibility probability is used to adjust predictions. [ABSTRACT FROM AUTHOR]

Machine learning models, especially deep neural network (DNN) models, have widespread and valuable applications in business activities. Training a deep learning model for commercial use requires plenty of private data, expert knowledge, and computing resources. The huge commercial value of such trained models has attracted the attention of attackers. Attackers can construct a dataset by repeatedly querying the target model for the output of the requested samples and then train a substitute model on this dataset that functions similarly to the target model. In this paper, we propose a defense scheme based on physical unclonable function (PUF) against such black-box model extraction attacks. We deploy a PUF on the user side and the corresponding PUF model on the service provider side to ensure that only legitimate users can obtain the correct model predictions. Our experimental results show that by choosing a suitable fuzzy extractor threshold d , legitimate users can recover more than 99.5% of the prediction results with a little additional computational overhead to the service provider. We perform a model extraction attack in the most favorable case for the attacker, and the prediction accuracy of the obtained substitute model is only about 10%, which demonstrates the effectiveness of our proposed scheme. Compared to existing defenses, our scheme not only effectively prevents black-box model extraction attacks but also ensures that the accuracy of the prediction service for legitimate users is not affected. • Defense based on PUF obfuscation against black-box model extraction attacks. • A PUF and the corresponding PUF model ensure that legitimate users can obtain correct model predictions. • Legitimate users can correctly recover more than 99.5% of the predictions with a little additional computational overhead. • The prediction accuracy of the model extracted by the attacker is only about 10%. [ABSTRACT FROM AUTHOR]

An adaptive sliding mode-model predictive control for the path following of intelligent unmanned vehicle is given in this paper. On account of excellent performances of the sliding mode structure, this algorithm can not only effectively estimate the uncertainty of the vehicle system to further improve the following accuracy, but minish the amount of calculation in comparision with model predictive control. Then, the following accuracy between the real system and the theoretical model can be compensated by the fractional order coefficient of controller. Therefore, an adaptive fractional order sliding mode-fractional order model predictive control is designed to follow the path of the intelligent unmanned vehicle. Meanwhile, the robust stability and control accuracy of the associated control algorithm are proved. Finally, different paths are designed to verify the theoretical analysis of the control performance in the controllers. [ABSTRACT FROM AUTHOR]

Multi-label learning aims to solve classification problems where instances are associated with a set of labels. In reality, it is generally easy to acquire unlabeled data but expensive or time-consuming to label them, and this situation becomes more serious in multi-label learning as an instance needs to be annotated with several labels. Hence, semi-supervised multi-label learning approaches emerge as they are able to exploit unlabeled data to help train predictive models. This work proposes a novel approach called Self-paced Multi-label Co-Training (SMCT). It leverages the well-known co-training paradigm to iteratively train two classifiers on two views of a dataset and communicate one classifier's predictions on unlabeled data to augment the other's training set. As pseudo labels may be false in iterative training, self-paced learning is integrated into SMCT to rectify false pseudo labels and avoid error accumulation. Concretely, the multi-label co-training model in SMCT is formulated as an optimization problem by introducing latent weight variables of unlabeled instances. It is then solved via an alternative convex optimization algorithm. Experimental evaluations are carried out based on six benchmark multi-label datasets and three metrics. The results demonstrate that SMCT is very competitive in each setting when compared with five state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Multi-label learning aims to solve classification problems where instances are associated with a set of labels. In reality, it is generally easy to acquire unlabeled data but expensive or time-consuming to label them, and this situation becomes more serious in multi-label learning as an instance needs to be annotated with several labels. Hence, semi-supervised multi-label learning approaches emerge as they are able to exploit unlabeled data to help train predictive models. This work proposes a novel approach called Self-paced Multi-label Co-Training (SMCT). It leverages the well-known co-training paradigm to iteratively train two classifiers on two views of a dataset and communicate one classifier's predictions on unlabeled data to augment the other's training set. As pseudo labels may be false in iterative training, self-paced learning is integrated into SMCT to rectify false pseudo labels and avoid error accumulation. Concretely, the multi-label co-training model in SMCT is formulated as an optimization problem by introducing latent weight variables of unlabeled instances. It is then solved via an alternative convex optimization algorithm. Experimental evaluations are carried out based on six benchmark multi-label datasets and three metrics. The results demonstrate that SMCT is very competitive in each setting when compared with five state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Aiming at the problems of unstable batch control of key crystal quality parameters and susceptibility to batch-to-batch non-repetitive disturbances during repeated operation of single crystal furnaces, this paper proposes a data-driven iterative learning model predictive control method based on an adaptive iterative extended state observer (IESO) for designing melt temperature and crystal diameter learning controllers with disturbance suppression. By applying dynamic linearization techniques and model predictive control strategies along the iterative axis, an ILMPC scheme with disturbance compensation terms using only input and output data of the system is designed. Among them, adaptive IESO is used to estimate the disturbance compensation terms. Then, the theoretical analysis shows that the tracking error of the ILMPC scheme can converge to a bounded range as the number of iterations increases. The experimental results verify the effectiveness of the proposed control method, which not only ensures that the control system has learning ability, but also achieves stable and accurate control of crystal quality parameters. [ABSTRACT FROM AUTHOR]

For stochastic nonlinear systems with time-varying delays, the existing robust control approaches are unnecessarily conservative in most practical scenarios. Within this context, a mathematically rigorous and computationally tractable tube-based model predictive control scheme is proposed in the framework of contraction theory. A contraction metric is systematically constructed via convex optimization by forming a differential LyapunovKrasovskii function on tangent space. It guarantees the perturbed actual solution trajectories to be contained within a robust positive invariant tube centered along the reference trajectories and results in an explicit exponential bound on the deviation. The application scenarios of the control contraction metric controller are extended from constant delay systems into time-varying delay systems thereby. Compared with the existing robust mechanism for time-delay systems based on min-max optimization formulation with a linear feedback controller, the proposed scheme greatly reduces the design conservativeness and yields a larger region of attraction. A sparse multi-dimensional Taylor network (MTN) is designed to parameterize the family of the geodesic. Compared to conventional NNs and MTN surrogates, sparse MTN features a more concise topology that enhances its computational efficiency conspicuously. Results of the numerical simulations verify the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

This paper proposes a robust switched model-based predictive controller design for discrete linear systems with state constraints, inputs, and disturbances limited in norm. Modeled via linear matrix inequalities, the online and offline designs of the proposed control aim at minimizing the upper bound of the quadratic performance index for a horizon of infinite prediction associated with the state estimator and the switching rule, seeking to guarantee the robust stability for closed-loop systems. To this end, three theorems are formulated. To demonstrate the effectiveness of the control strategy, a comparative analysis is performed between the performance of the proposed model and a benchmark method. From the results, it is possible to conclude that the proposed method is promising in the scope of control of linear systems subject to switching, being more efficient than the benchmark for the stabilization and control of both numerical examples. [ABSTRACT FROM AUTHOR]

Previous models for predicting rumor-forwarding trends were primarily focused on feature generation and model prediction in two independent directions: message text and user association features. However, the abstraction of user awareness, text contextual feature extraction limitation, and inefficiency of traditional hyperparameter search methods still pose numerous challenges. This study proposes a rumor-forwarding trend prediction model that combines user awareness and multi-type rumor to address such challenges. First, considering the abstraction of user awareness under multi-type rumors, we extract features by cascading user behavior, historical activities, interactions, and activity levels and by fusing features using a two-layer fully connected network to effectively quantify the relevant features of user awareness. Second, considering the limitations of traditional text representation in semantic context understanding, we use the Bidirectional Encoder Representation from Transformers (BERT) pre-training model to characterize the text in the topic, obtain text representation sequence with contextual relationships, and propose an Improved Cuckoo Search (ICS) method that optimizes the hyperparameters of the temporal convolutional network (TCN) model. Finally, an Improved Cuckoo Search-TCN-based rumor-forwarding trend prediction model is constructed based on user awareness features and text representation sequences to predict the rumor-forwarding trend. Certain rumors with a large potential impact range can be monitored at the early dissemination stage. [ABSTRACT FROM AUTHOR]

• Temporal information is effectively integrated into the representation model. • A new prediction model is proposed to gain temporal information. • The memory capacity of the replay buffer is smaller than the existing methods. • The policy lag is proven to be decreased quickly in maximum-entropy reinforcement learning. Partially observable Markov decision process (POMDP) is a key challenging problem in the application of reinforcement learning since it comprehensively describes real agent-environment interactions. Recent works mainly utilize conventional reward signals to train a representation that converts POMDPs to MDPs. However, rewards alone are not enough for a good representation without temporal information. In this paper, we first introduce a novel Recurrent Prediction Model to integrate temporal information into the representation that solves POMDP problems by training three additional unsupervised prediction models, named transition model, reward recovery model, and observation recovery model. This paper secondly makes a modification of the data structure of vanilla replay buffer to reduce the memory usage and thirdly proposes an off-policy correction algorithm to decrease the policy lag in POMDPs. The experiments show that our model achieves better performance in partially observable environments on both stand-alone and distributed training systems. [ABSTRACT FROM AUTHOR]

This paper proposes a high-precision three-dimensional nonlinear optimal computational guidance law in the terminal phase of an interceptor that ensures near-zero miss-distance as well as the desired impact angle. Additionally, it achieves these ambitious objectives while ensuring that the lead angle and lateral acceleration constraints are not violated throughout its trajectory. This ensures (i) the target does not escape the field of view of its seeker at any point in time (a state constraint) and (ii) it does not demand unreasonable lateral acceleration that cannot be generated (a control constraint). The guidance problem is formulated and solved using newly proposed Path-constrained Model Predictive Static Programming (PC-MPSP) framework. All constraints, both equality and inequality, are equivalently represented as linear constraints in terms of the errors in the control history, thereby reducing the complexity and dimensionality of the problem significantly. Coupled with a quadratic cost function in control, the problem is then reduced to a standard quadratic optimization problem with linear constraints, which is then solved using the computationally efficient interior-point method. Results clearly demonstrate the advantage of the proposed guidance scheme over the conventional Biased PN as well as the recently proposed GENeralized EXplicit (GENEX) guidance techniques. Numerical simulations with variation in initial conditions and Monte–Carlo simulations with parametric uncertainty demonstrate the robustness of the proposed guidance scheme. [ABSTRACT FROM AUTHOR]

In this paper, a distributed control architecture is presented for addressing the cooperative perimeter surveillance of rectangular areas for multi-agent system whose dynamics is described by double integrator models subject to exogenous bounded disturbances. In particular, a novel methodology to generate proper state trajectories of a swarm of agents is provided and then exploited as references for the underlying model predictive controllers bank. Specifically, the swarm is first driven along the perimeter of a rectangular strip and next imposed to enter a given containment region while remaining simultaneously outside a forbidden region around a prescribed target. Steady-state conditions are analyzed and sufficient conditions derived in terms of the control law parameters. Simulation results put in light the main properties of the control architecture that is designed to adequately switch between two different distributed algorithms in order to guarantee constraints satisfaction within the capturing region despite any disturbance realization. [ABSTRACT FROM AUTHOR]

This paper studies model predictive security control (MPSC) for networked switched systems under denial-of-service (DoS) attacks. Most of existing works only adjust the triggering scheme when being attacked. Different from them, this paper proposes a novel timing-response event-triggering scheme (TR-ETS) to reduce the impact of attacks on system performance, which can not only configure system resources adaptively, but also accurately detect attack information and compensate the attacked data. Specifically, the proposed scheme includes two event-based triggers, which can dynamically and jointly regulate the communication/calculation ability, generate virtual attack sequences and acquire the number of passive packet loss. Then, based on the triggered states, a class of model predictive controllers is designed to optimize the control action. Due to possible strong attacks, a security control framework including network and local loops be introduced and a permissable type-switching mechanism (PTM) is used. Under the permissable controllers (i.e., network and local controllers), sufficient conditions for the stability of closed-loop switched systems are derived. In addition, a set of model predictive optimization algorithm using linear matrix inequalities (LMIs) technique is addressed. Finally, the effectiveness of the proposed method is verified by illustrative examples. [ABSTRACT FROM AUTHOR]

Reasonable prediction of natural gas consumption and production in China is helpful for decision-makers formulate policies to solve the increasingly prominent contradiction between supply and demand of natural gas in China. Taking into account the data environment of China's small sample, this paper considers using the grey prediction model with characteristic of small sample modeling to study the consumption and production of natural gas in China. To achieve accurate prediction, this paper developed a novel discrete fractional nonlinear grey Bernoulli model with power term (DFNGBM(1,1, α)), and used the latest equilibrium optimizer to solve the hyper-parameters of the model. The proposed model has the advantages of most existing grey prediction models, such as fractional-order accumulation operation and time power term. In addition, the proposed model is a unified expression form of many other grey forecasting models. The consumption and production of natural gas in China from 2003 to 2020 are used as examples to verify the feasibility and validity of the proposed model compared with other competitive models that have been used to study natural gas. It turns out that the prediction ability of the proposed model is better than other models. Therefore, the feasibility and effectiveness of the proposed model have been confirmed. Finally, the DFNGBM (1,1, α) model is used to predict China's natural gas consumption and production in the next few years, and some corresponding recommendations are given based on the prediction results. [ABSTRACT FROM AUTHOR]

In this paper, the problem of observer-based model predictive control (MPC) for a multi-channel cyber-physical system (CPS) with uncertainties and hybrid attacks is investigated via interval type-2 Takagi-Sugeno (IT2 T-S) fuzzy model. Both denial-of-service (DoS) and false data injection (FDI) attacks are studied due to the vulnerability of wireless transmission channels. The objective of the addressed problem is to improve the security performance of the multi-channel CPS under malicious attacks, which has not been adequately investigated in the existing MPC algorithms. Moreover, uncertainties which appear not only in the membership functions but in both state and input matrices are considered. In this paper, different from the method that FDI attacks are handled by the bounded functions, an off-line observer is designed to actively defend against the FDI attacks. Meanwhile, an on-line MPC optimization algorithm, which minimizes the upper bound of objective function respecting input constraints, is presented to obtain the secure controller gains. Finally, an illustrative example is provided to verify the effectiveness and superiority of presented approach. [ABSTRACT FROM AUTHOR]

• Model Predictive Control (MPC) algorithms for GRU and LSTM networks are described and evaluated for two benchmarks. • It is shown that the typical approach to MPC that hinges on successively linearized networks does not give precise predictions and satisfactory control quality. • The presented MPC control schemes utilize online advanced trajectory linearization, which yields simple quadratic optimization programs. • The discussed approaches give excellent prediction accuracy and control quality. • The performance of MPC based on LSTM and GRU networks is compared; simpler GRU ones are recommended. This article is concerned with Model Predictive Control (MPC) algorithms that use Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) neural networks for prediction. For two benchmark processes, it is shown that the typical approach to MPC that hinges on successively linearized LSTM or GRU models do not give precise predictions and satisfactory control quality. The presented MPC control schemes utilize online advanced trajectory linearization, which yields simple quadratic optimization programs. It is shown that the discussed approaches give excellent prediction accuracy and control quality, very similar to that possible in MPC with full nonlinear prediction and nonlinear optimization done online. It is also demonstrated that the described MPC algorithms are a few times faster than the MPC method with nonlinear optimization. Moreover, the performance of MPC based on LSTM and GRU networks is compared, and simpler GRU networks are recommended. [ABSTRACT FROM AUTHOR]

A new event-triggered model predictive control (ET-MPC) framework is proposed for perturbed linear time-invariant (LTI) systems with input and state constraints. Firstly, we propose a novel aperiodic sampling strategy, which determines the upcoming update instants based on the proportion, integral and differential (PID) of the error between the actual state and its optimal prediction to effectively reduce the computational and communication burden of the system. Then, a PID based ET-MPC algorithm is developed and its feasibility as well as the closed-loop stability are strictly proved. Finally, simulation and comparison studies are conducted based on the cart-damper-spring system to illustrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the use of Model Predictive Control (MPC) applications for quasi-Linear Parameter Varying (qLPV) systems subject to faults along the input channels. We propose a Fault Tolerant Control (FTC) mechanism based on a robust state-feedback MPC synthesis, considering polytopic inclusions. In order to alleviate the numerical burden of the robust min-max procedure, we use small prediction horizons, in such a way that the solution becomes viable for real-time systems. The FTC system is able to tolerate time-varying saturation of the actuator, which may happen due to malfunctions. Recursive feasibility and poly-quadratic stability guarantees are ensured through the synthesis of adequate terminal ingredients. Accordingly, we present a catalogue of three different LMI remedies, considering: (a) parameter-independent ingredients, (b) a parameter-dependent terms and (c) a parameter-dependent maps that take into account bounded rates of parameter variation. An autonomous driving car example is used to illustrate the performances of the proposed technique, which is compared to other MPCs from the literature. The proposed FTC method is able to ensure good performances, obtained with reduced computational demand. [ABSTRACT FROM AUTHOR]

In this paper, the optimal driving torques of four wheels in an electric vehicle (EV) are obtained by minimizing the losses of four in-wheel motors. In order to slightly change these optimal torques for vehicle stability recovery, a new constrained active front steering (AFS) system is analytically designed and integrated with the torque vectoring (TV) system in a novel decentralized structure. In the proposed structure, the required external yaw moment is applied by the TV controller only when the constrained AFS is filled to capacity determined by a stability index extracted from the nonlinear phase plane analysis. As a result of this integration strategy, the external yaw moment is used as low as possible. Consequently, the torques in electrical motors are used near the optimal values consistent with optimal energy consumption. Comparative simulation studies with the standalone TV are conducted in the CarSim software environment to show the efficiency of the proposed decentralized control structure in terms of energy consumption and stability. Moreover, the suitability of the constrained control method used in the integration structure is shown in comparison with the well-known nonlinear model predictive control method in terms of practical implementation. [ABSTRACT FROM AUTHOR]

Model predictive controllers use dynamics models to solve constrained optimal control problems. However, computational requirements for real-time control have limited their use to systems with low-dimensional models. Nevertheless, high-dimensional models arise in many settings, for example, discretization methods for generating finite-dimensional approximations to partial differential equations can result in models with thousands to millions of dimensions. In such cases, reduced-order models (ROMs) can significantly reduce computational requirements, but model approximation error must be considered to guarantee controller performance. In this article, a reduced-order model predictive control (ROMPC) scheme is proposed to solve robust, output feedback, constrained optimal control problems for high-dimensional linear systems. Computational efficiency is obtained by using projection-based ROMs, and guarantees on robust constraint satisfaction and stability are provided. The performance of the approach is demonstrated in simulation for several examples, including an aircraft control problem leveraging an inviscid computational fluid dynamics model with dimension 998 930. [ABSTRACT FROM AUTHOR]

Various efforts have been devoted to developing stabilizing distributed model predictive control (MPC) schemes for tracking piecewise constant references. In these schemes, terminal sets are usually computed offline and used in the MPC online phase to guarantee recursive feasibility and asymptotic stability. Maximal invariant terminal sets do not necessarily respect the distributed structure of the network, hindering the distributed implementation of the controller. On the other hand, ellipsoidal terminal sets respect the distributed structure, but may lead to conservative schemes. In this article, a novel distributed MPC scheme is proposed for reference tracking of networked dynamical systems, where the terminal ingredients are reconfigured online depending on the closed-loop states to alleviate the aforementioned issues. The resulting nonconvex infinite-dimensional problem is approximated using a quadratic program. The proposed scheme is tested in simulation, where the proposed MPC problem is solved using distributed optimization. [ABSTRACT FROM AUTHOR]

Approximating model predictive control (MPC) policy using expert-based supervised learning techniques requires labeled training datasets sampled from the MPC policy. This is typically obtained by sampling the feasible state space and evaluating the control law by solving the numerical optimization problem offline for each sample. Although the resulting approximate policy can be cheaply evaluated online, generating large training samples to learn the MPC policy can be time-consuming and prohibitively expensive. This is one of the fundamental bottlenecks that limit the design and implementation of MPC policy approximation. This technical article aims to address this challenge, and proposes a novel sensitivity-based data augmentation scheme for direct policy approximation. The proposed approach is based on exploiting the parametric sensitivities to cheaply generate additional training samples in the neighborhood of the existing samples. [ABSTRACT FROM AUTHOR]

Two stochastic model predictive control algorithms, which are referred to as distributionally robust model predictive control algorithms, are proposed in this article for a class of discrete linear systems with unbounded noise. Participially, chance constraints are imposed on both of the state and the control, which makes the problem more challenging. Inspired by the ideas from distributionally robust optimization (DRO), two deterministic convex reformulations are proposed for tackling the chance constraints. Rigorous computational complexity analysis is carried out to compare the two proposed algorithms with the existing methods. Recursive feasibility and convergence are proven. Simulation results are provided to show the effectiveness of the proposed algorithms. [ABSTRACT FROM AUTHOR]