With the development of society and the improvement of people's living standard, diabetes has gradually become one of the world public health problems that endangers human life safety and affects the development of global economy. Type 1 diabetes is a chronic metabolic disorder that prevents the pancreas from producing insulin and requires lifelong treatment with daily insulin injections to prevent high blood sugar. The lack of insulin leads to the continuous high level of blood sugar in patients. The most important thing for patients is to monitor blood sugar changes and control blood sugar within the normal range. In this paper, a systematic literature search is carried out to study these two aspects, and key information such as the learning model adopted in the literature, the main results, the development of relevant technologies, and the limitations are summarized. From the perspective of glucose-insulin prediction model, due to the complex structure of physiological model, many parameters are difficult to identify, most of the modeling methods are data-driven. There is a great room for improvement in the research on how to mine and utilize existing models to effectively establish accurate glucose prediction models for different objects. From the perspective of control theory, after insulin injection, there is a certain delay in the reduction of blood glucose concentration, and the onset time is different depending on the injection site. The rapid development of deep learning and the increase in available data offer the possibility of addressing these challenges in the near future. When designing closed-loop glucose algorithms, we consider using a variety of approaches to establish a personalized glucose control algorithm for each diabetic patient.

This paper addresses the issue of stochastic stability for continuous-time semi-Markovian jump neural networks. Initially, according to the characteristics of the time-delay interval, the Lyapunov-Krasovskii functional (LKF) with the semi-Markov process is constructed to ensure the stability of the switching networks. Next, the promoted double integral inequality lemma is utilized to estimate the weak infinitesimal operator of the designed LKF, and this paper establishes a time-delay correlation criterion. Moreover, the criterion is combined with the Lyapunov stability theory to make the system reach stability in the mean-square sense. Finally, the paper provides an example to demonstrate the effectiveness of the proposed approach.

Particle swarm optimization (PSO) algorithm has been widely used in large-scale complex problems such as resource allocation in recent years because of its simple implementation and easy operation. However, the slow convergence speed and low solution accuracy of the algorithm also restrict its further applications. To solve the above problems, this paper introduces the chromosome crossing characteristics of genetic algorithm (GA), and proposes a comprehensive learning particle swarm optimization based on optimal particle recombination. With the help of the ability of comprehensive learning strategy to efficiently explore the solution space, this method organically combines the excellent information explored by each particle through the optimal particle recombination, so as to obtain a better individual, speed up the convergence of the algorithm, and improve the solution accuracy of the problem. The experimental results of benchmark function show that the proposed algorithm has faster convergence speed and optimization accuracy than the original algorithm, and the results of Friedman test and Wilcoxon signed-rank test prove the feasibility of the optimal particle recombination operation in particle swarm optimization.

Implementing cooperative scheduling of multi-home microgrid energy and reducing the dependence on the main grid have become the focus of microgrid energy management research. This paper proposes a new multi-agent adaptive dynamic programming (MAADP) method for the cooperative control of distributed home energy. Each home is defined as a learning agent that needs to reasonably schedule the energy storage system to meet the respective load demand while accomplishing cooperative scheduling among the individual homes. In addition, an energy clearing center (ECC) is introduced to complete the energy exchange between each microgrid to protect the benefits of all parties. The proposed method adopts the learning strategy of “centralized learning and decentralized execution” to avoid the leakage of private information. The experimental comparison with the benchmark method verifies that the method can realize the cooperative scheduling of each home and reduce the dependence on the main grid.

In this paper, based on ACP (ACP: artificial societies, computational experiments, and parallel execution) approach, a parallel control method is proposed for zero-sum games of unknown time-varying systems. The process of constructing a sequence of artificial systems, implementing the computational experiments, and conducting the parallel execution is presented. The artificial systems are constructed to model the real system. Computational experiments adopting adaptive dynamic programming (ADP) are shown to derive control laws for a sequence of artificial systems. The purpose of the parallel execution step is to derive the control laws for the real system. Finally, simulation experiments are provided to show the effectiveness of the proposed method.

In this paper, a new bolt fault diagnosis method is developed to solve the fault diagnosis problem of wind turbine flange bolts using one-dimensional depthwise separable convolutions. The main idea is to use a one-dimensional convolutional neural network model to classify and identify the acoustic vibration signals of bolts, which represent different bolt damage states. Through the methods of knock test and modal simulation, it is concluded that the damage state of wind turbine flange bolt is related to the natural frequency distribution of acoustic vibration signal. It is found that the bolt damage state affects the modal shape of the structure, and then affects the natural frequency distribution of the bolt vibration signal. Therefore, the damage state can be obtained by identifying the natural frequency distribution of the bolt acoustic vibration signal. In the present one-dimensional depth-detachable convolutional neural network model, the one-dimensional vector is first convolved into multiple channels, and then each channel is separately learned by depth-detachable convolution, which can effectively improve the feature quality and the effect of data classification. From the perspective of the realization mechanism of convolution operation, the depthwise separable convolution operation has fewer parameters and faster computing speed, making it easier to build lightweight models and deploy them to mobile devices.

This paper presents an online integral reinforcement learning (RL) solution for problems with hierarchy decision makers. Specifically, we reformulate this model as a leader-follower game, in which control input and deterministic disturbance act as decision makers at different levels of hierarchy: The control input plays the role of the leader, while the disturbance plays the role of the follower. The main contributions of this paper can be summarized as follows. First, we introduce online RL to deal with systems that have partially unknown information, meaning that accurate dynamic information is not required. Second, we solve the leader-follower coupled Hamilton-Jacobi (HJ) and Riccati equations approximately online using the derived algorithm. Third, we provide turning laws for cost functions and controllers, which ensure closed-loop stability simultaneously.