ABSTRACT This paper focuses on the multi‐UAV encirclement problem in the presence of obstacles by proposing an improved method that integrates the extended Kalman filter (EKF) into the multi‐agent deep deterministic policy gradient (MADDPG) algorithm. Firstly, the EKF is employed to accurately estimate the target position, providing position information for the subsequent encirclement strategy. Then, based on the estimated target position, the hunting points are calculated and determined. Subsequently, the hunting points are allocated to each UAV in a reasonable manner, ensuring that the UAVs can arrive at the estimated positions efficiently and simultaneously in the shortest time. Moreover, a composite reward function is designed to guide the UAVs to make optimal decisions in the encirclement task, where a segmented reward function is used to train the UAV to perform smooth obstacle avoidance. Through extensive training experiments, the convergence and effectiveness of the proposed improved algorithm are significantly verified, providing strong technical support for the efficient execution of the UAV encirclement task.

ABSTRACT The paper describes a new neural network design, which is known as the Local Recurrent Sigma‐Pi Artificial Neural Network (LRSPANN), to model and control nonlinear dynamical systems. The given model includes local recurrent self‐feedback links in the hidden layer, which contribute to the dynamic memory and make it possible to effectively represent the behavior of the temporal system. The backpropagation learning algorithm is a gradient‐descent‐based method that is effectively used in updating network parameters and reducing modeling error. A Lyapunov‐based stability analysis is conducted to achieve reliable learning and closed‐loop stability. The effectiveness of the proposed LRSPANN is tested with the help of comparative simulations with Sigma‐Pi Artificial Neural Network (SPANN), Elman Recurrent Neural Network (ERNN), and Feed‐Forward Neural Network (FFNN). The proposed model has the lowest mean squared error (MSE) of in Example 1 and a mean squared error of in Example 2, which is better than any other compared network. These findings illustrate that the proposed methodology is the most accurate and efficient for modeling and controlling nonlinear systems.

ABSTRACT This study is concerned with the security control problem of multi‐agent systems (MASs). First, the multi‐agent model with unknown actuator faults and disturbances is considered, and the adverse effects of these are estimated by fuzzy logic systems (FLSs). Then, the Hamiltonian–Jacobi–Bellman equation is derived under a performance cost function with a discount term. Then, two FLSs are adopted to build an actor–critic structure to approximate the cost function and the optimal controller, respectively. Furthermore, to reduce the energy consumption, a new RL and improved artificial potential field (APF) based optimal fault‐tolerant formation controller with collision avoidance is obtained. Finally, the effectiveness of the research method is verified by the comparison of simulation results.

ABSTRACT This paper studies stochastic Runge‐Kutta (SRK) approximation for solving stochastic optimal control problems where the state process is governed by Merton's jump‐diffusion model. We propose a practical numerical scheme based on the SRK method to approximate the solutions of the resulting equations. Moreover, strong order conditions of the proposed scheme are provided. Following the presentation of the main scheme of Merton's optimal consumption‐investment problem, solution of a controlled pure jump model is introduced as a variant. Numerical experiments demonstrate the efficiency of the SRK method in solving the optimal consumption‐investment problem, highlighting its potential for practical applications in financial decision‐making under discontinuous dynamics.

ABSTRACT This paper is concerned with the theory, construction and application of variable‐stepsize implicit Peer two‐step methods that are super‐convergent for variable stepsizes, that is, preserve their classical order achieved for uniform stepsizes when applied in a gradient‐based solution algorithm to solve ODE‐constrained optimal control problems in a first‐discretize‐then‐optimize setting. Gradients of the objective function can be computed most efficiently using approximate adjoint variables. High accuracy with moderate computational effort can be achieved through time integration methods that satisfy a sufficiently large number of adjoint order conditions for variable stepsizes and provide gradients with higher‐order consistency. In this paper, we enhance our previously developed variable implicit two‐step Peer triplets constructed in [J. Comput. Appl. Math. 460, 2025] to get ready for large‐scale dynamical systems with varying time scales without losing efficiency. A key advantage of Peer methods is their use of multiple stages with the same high stage order, which prevents order reduction—an issue commonly encountered in semi‐discretized PDE problems with boundary control. Two third‐order methods with four stages, good stability properties, small error constants, and a grid adaptation by equi‐distributing global errors are constructed and tested for a 1D boundary heat control problem and an optimal control of cytotoxic therapies in the treatment of prostate cancer.

ABSTRACT The research and application of power inverters and related technologies represent the mainstream development of modern power electronics technology. For the unknown inverter system with measurement noise, a model‐free adaptive iterative learning control (MFAILC) scheme with a low‐pass filter is proposed. The purpose of this work is to achieve high tracking performance in the output voltage even when measurement noise exists. This method not only effectively overcomes the time‐varying parameter uncertainty in the inverter system by using the input/output data of the controlled plant, but also suppresses the measurement noise through the introduced filter. In order to verify the effectiveness of the convergence, tracking ability, and robustness of the proposed method, a simulation is conducted.

ABSTRACT Dengue fever remains a major public health concern in tropical and subtropical regions due to its complex transmission dynamics and lack of specific antiviral treatment. In this study, an age‐structured delay differential model is developed to investigate dengue transmission between juvenile and adult human populations and mosquito vectors. Biologically meaningful time delays are incorporated to represent incubation periods and age progression. Both deterministic and stochastic formulations are analyzed to capture average disease dynamics and random environmental effects. Global stability of the disease‐free equilibrium for the deterministic model is established using Lyapunov functional techniques, while conditions for disease extinction and the existence of a stationary distribution are derived for the stochastic system. Numerical simulations based on the Milstein method validate the theoretical findings and illustrate the influence of noise intensity on disease persistence. Sensitivity analysis identifies key parameters governing transmission dynamics. Additionally, optimal control strategies targeting juvenile and adult humans and mosquito populations are formulated using Pontryagin's Maximum Principle (PMP) and solved via a forward–backward sweep algorithm. The results demonstrate that optimal interventions significantly reduce infection levels and control costs, highlighting the effectiveness of integrated modeling and control strategies for dengue prevention.