Yuyang Zhou

A novel model-free hierarchical reinforcement learning (HRL)–based Linear Quadratic Regulator (LQR) control framework with adaptive weight selection is proposed to address the reliance of conventional LQR methods on accurate system models and manual parameter tuning. The proposed approach adopts a two-level learning architecture in which a high-level meta-agent adaptively optimizes the LQR weighting matrices Q and R through entropy-based trajectory evaluation, while a low-level base-agent performs model-free policy iteration to update the state-feedback control law under unknown system dynamics. By decoupling weight optimization from control-law learning, the framework enables simultaneous adaptation of the cost-function parameters and the feedback gain without requiring explicit model information. To enhance learning stability and exploration during weight adaptation, Gaussian noise and an experience replay mechanism are incorporated into the learning process. Numerical simulations on second- and third-order linear systems demonstrate that the proposed HRL-based LQR method achieves effective control performance, reliable convergence, and improved adaptability in model-free environments.

To address the challenges of unbalanced demand and high operational costs in highway port logistics, this study investigates the scheduling of tractors and semitrailers under time window constraints in a networked environment, where geographically distributed ports are interconnected by fixed routes, and tractors dynamically transport semitrailers between ports to balance asymmetric demands. A mathematical optimization model is developed, incorporating multiple car yards, diverse transport demands, and temporal constraints. To solve the model efficiently, an Adaptive Large Neighborhood Search (ALNS) algorithm is proposed and benchmarked against an improved Ant Colony System (IACS). Simulation results show that, compared to traditional scheduling methods, the proposed approach reduces the number of required tractors by up to 61% and operational costs by up to 21%, depending on tractor working hours. The tractor-to-semitrailer ratio improves from 1.00:1.10 to 1.00:2.59, demonstrating the enhanced resource utilization enabled by the ALNS algorithm. These findings offer practical guidance for optimizing tractor and semitrailer configurations in highway port operations under varying conditions.

Tracking control of the output probability density function presents significant challenges, particularly when dealing with unknown system models and multiplicative noise disturbances. To address these challenges, this paper introduces a novel tracking control algorithm based on reinforce-ment Q-learning. Initially, a B-spline model is employed to represent the original system, thereby transforming the control problem into a state weight tracking issue within the B-spline stochastic system model. Moreover, to tackle the challenge of unknown stochastic system dynamics and the presence of multiplicative noise, a model-free reinforcement Q-learning algorithm is employed to solve the control problem. Finally, the proposed algorithm’s effectiveness is validated through comprehensive simulation examples.

CEO power has been extensively studied across various disciplines and country contexts. Despite the exponential growth of research, there has been limited effort to integrate the vast body of literature. Using bibliometric and other analytical techniques we apply to the 580 articles in our review, we identify and discuss the topics and major research streams considered in CEO power research and their evolution over the years. We also highlight several shortcomings in the existing literature, including four pressing challenges concerning unclear conceptualizations, varied measurement and methods, the under-contextualized nature of CEO power across international contexts, and a lack of attention to how the changing corporate governance landscape has affected CEO power. We provide a roadmap for future scholarship by offering suggestions for addressing these pressing challenges. Finally, we provide several new and promising research directions in our discussion.

This study presents a novel event-triggered relearning framework for neural network modeling, designed to improve prediction precision in dynamic stochastic complex industrial systems under non-stationary and variable conditions. Firstly, a sliding window algorithm combined with entropy is applied to divide the input and output datasets across different operational conditions, establishing clear data boundaries. Following this, the prediction errors derived from the neural network under different operational states are harnessed to define a set of event-triggered relearning criteria. Once these conditions are triggered, the relevant dataset is used to recalibrate the model to the specific operational condition and predict the data under this operating condition. When the predicted data fall within the training input range of a pre-trained model, we switch to that model for immediate prediction. Compared with the conventional BP neural network model and random vector functional-link network, the proposed model can produce a better estimation accuracy and reduce computation costs. Finally, the effectiveness of our proposed method is validated through numerical simulation tests using nonlinear Hammerstein models with Gaussian noise, reflecting complex stochastic industrial processes.

In this paper, we present an innovative decentralised control framework, designed to address stochastic dynamic complex systems that are influenced by multiple multiplicative noise factors. Our advanced approach builds upon the foundation of conventional Decentralised Fully Probabilistic Design (DFPD) by refining the Riccati equation to accommodate multiple noise sources effectively. By embracing the inherent stochastic nature of complex systems, our methodology fully characterises their dynamic behaviours using probabilistic state–space models, delivering a comprehensive representation of subsystem components. Importantly, the DFPD approach also incorporates system and input constraints by characterising their corresponding ideal distributions, ensuring optimal functionality and performance while adhering to permissible boundaries. To further enhance system performance, we introduce a probabilistic message passing architecture that enables seamless communication between neighbouring subsystems and promotes harmonised decision-making among local nodes. To demonstrate the efficacy of our proposed framework, we employ a three-inverted pendulum system as a numerical example and compare its performance to that of the conventional DFPD. Through this comparison, we showcase the advantages of our novel decentralised control approach in handling complex systems with multiple noise factors.

For many human body diseases, treatments in the early stages are more efficient and safer than those in the later stages; therefore, detecting the early symptoms of a disease is crucial. One of the most significant early indicators for diseases is bio-mechanical motion. This paper provides a unique way of monitoring bio-mechanical eye motion based on electromagnetic sensing technology and a ferro-magnetic material, ferrofluid. The proposed monitoring method has the advantages of being inexpensive, non-invasive, sensor-invisible and extremely effective. Most of the medical devices are cumbersome and bulky, which makes them hard to apply for daily monitoring. However, the proposed eye-motion monitoring method is designed based on ferrofluid eye make-up and invisible sensors embedded inside the frame of glasses such that the system is wearable for daily monitoring. In addition, it has no influence on the appearance of the patient, which is beneficial for the mental health of some patients who do not want to attract public attention during treatment. The sensor responses are modelled using finite element simulation models, and wearable sensor systems are created. The designed frame of the glasses is manufactured based on 3-D printing technology. Experiments are conducted to monitor eye bio-mechanical motions, such as the frequency of eye blinking. Both the quick blinking behaviour with an overall frequency of around 1.1 Hz and the slow blinking behaviour with an overall frequency of around 0.4 Hz can be observed through experimentation. Simulations and measurements results show that the proposed sensor design can be employed for bio-mechanical eye-motion monitoring. In addition, the proposed system has the advantages of invisible sensor set-up and will not affect the appearance of the patient, which is not only convenient for the daily life of the patient but also beneficial for mental health.

Output probability density function (PDF) tracking control of stochastic systems has always been a challenging problem in both theoretical development and engineering practice. Focused on this challenge, this work proposes a novel stochastic control framework so that the output PDF can track a given time-varying PDF. Firstly, the output PDF is characterised by the weight dynamics following the B-spline model approximation. As a result, the PDF tracking problem is transferred to a state tracking problem for weight dynamics. In addition, the model error of the weight dynamics is described by the multiplicative noises to more effectively establish its stochastic dynamics. Moreover, to better reflect the practical applications in the real world, the given tracking target is set to be time-varying rather than static. Thus, an extended fully probabilistic design (FPD) is developed based on the conventional FPD to handle multiplicative noises and to track the time-varying references in a superior way. Finally, the proposed control framework is verified by a numerical example, and a comparison simulation with the linear–quadratic regulator (LQR) method is also included to illustrate the superiority of our proposed framework.

Non-Gaussian randomness widely exists in complex dynamical systems, in which the traditional mean-variance index cannot fully reflect the systematic characteristics. To improve the performance of control design subjected to non-Gaussian noises, stochastic distribution control (SDC) theory was proposed in the 1990s, where the output probability density function (PDF) has been investigated as an additional system variable. Following this framework, SDC has been extended to other research subjects in control systems such as filter design, fault diagnosis, and so on. It shows that SDC supplies an important solution to enhance the accuracy of system design, which is further beneficial to almost all the topics subject to non-Gaussian randomness. Meanwhile, the theoretical results of the SDC have been applied to several practical industrial applications. As data science raises based on the development of industrial artificial intelligence, SDC has been further developed recently focusing on data-driven design and multi-agent systems. To explore the new challenges with the evolution of SDC, e.g. unknown system models, unknown noise distributions, strong non-stationary transient dynamics, stability analysis and industrial applications, this survey summarises the most recent published results in the last 5 years of SDC work in terms of modelling, control, filtering, fault diagnosis, and industrial applications. Based on the technical analysis, potential future work is discussed in the end.

In this paper, a novel decentralised control framework based on decentralised fully probabilistic design (DFPD) is proposed for a class of stochastic dynamic complex systems with multiple multiplicative noises. Compared with the existing conventional DFPD, the new procedure is improved by modifying the Riccati equation in order to deal with multiple multiplicative noises. Considering the stochastic nature of complex systems, the systems’ dynamical behaviours are fully charaterised by probabilistic state-space models. In this way, a complete description of the components of the subsystems is provided. In addition, probabilistic message passing architecture is introduced to provide communication between neighbouring subsystems and to harmonise the actions between the local nodes. To illuminate the effectiveness of the proposed framework, a three inverted pendulum system numerical example is presented and the results are compared with the conventional DFPD.