Dongshu Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Dongshu Wang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Longkun Tang
Huaqiao University
6 shared publications - Lihong Huang
Department of Mathematics and Computer Science, Changsha University, Changsha 410022, China
5 shared publications Longkun Tang
4 shared publicationsLihong Huang
3 shared publicationsQingwen Yu
3 shared publications- Shifan Luo
School of Mathematical Sciences, Huaqiao University, Quanzhou, China
3 shared publications
Latest Publications
- Open Access
- 19 September 2026
- Published by Elsevier in SSRN Electronic Journal
Predefined-time stabilization of time-varying reaction--diffusion systems with sign-indefinite dynamics via aperiodic intermittent control
This article investigates the predefined-time (PT) stabilization of time-varying reaction--diffusion systems (RDSs) subject to transmission delays and generalized Robin boundary conditions, in which the aperiodic intermittent actuation, viewed as a time-dependent switching between active and resting modes, is scheduled through a tunable activation ratio $\chi$. Departing from classical Lyapunov methods that require strictly negative-definite derivatives at every instant, two original PT stability criteria are established from an integral envelope constructed with the logistic function: the derivative is permitted to be sign-indefinite, so that transient energy divergence during resting intervals and within active control phases is explicitly tolerated, while exact convergence at the preassigned deadline $T_p$ is still guaranteed. On this basis, semi-intermittent and intermittent controllers with bounded control effort are synthesized; in contrast to conventional PT controllers whose gains grow without bound near the deadline, the proposed actuation vanishes as the state converges, eliminating singular gains and peaking phenomena. Exploiting the physical dissipation of the Robin boundaries, the derived scalar algebraic criteria unify Dirichlet, Neumann, and mixed boundary configurations without resorting to high-dimensional LMIs. As a representative nonlinear application, the framework is specialized to spatiotemporal neural networks with chaotic open-loop behavior; numerical simulations confirm exact PT convergence, initial-state independence of the settling-time bound, energy-efficient aperiodic switching, and chattering-free implementation under continuous actuation.- 19 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPredefined-time stabilization of time-varying reaction--diffusion systems with sign-indefinite dynamics via aperiodic intermittent control
This article investigates the predefined-time (PT) stabilization of time-varying reaction--diffusion systems (RDSs) subject to transmission delays and generalized Robin boundary conditions, in which the aperiodic intermittent actuation, viewed as a time-dependent switching between active and resting modes, is scheduled through a tunable activation ratio $\chi$. Departing from classical Lyapunov methods that require strictly negative-definite derivatives at every instant, two original PT stability criteria are established from an integral envelope constructed with the logistic function: the derivative is permitted to be sign-indefinite, so that transient energy divergence during resting intervals and within active control phases is explicitly tolerated, while exact convergence at the preassigned deadline $T_p$ is still guaranteed. On this basis, semi-intermittent and intermittent controllers with bounded control effort are synthesized; in contrast to conventional PT controllers whose gains grow without bound near the deadline, the proposed actuation vanishes as the state converges, eliminating singular gains and peaking phenomena. Exploiting the physical dissipation of the Robin boundaries, the derived scalar algebraic criteria unify Dirichlet, Neumann, and mixed boundary configurations without resorting to high-dimensional LMIs. As a representative nonlinear application, the framework is specialized to spatiotemporal neural networks with chaotic open-loop behavior; numerical simulations confirm exact PT convergence, initial-state independence of the settling-time bound, energy-efficient aperiodic switching, and chattering-free implementation under continuous actuation. - 7 September 2026
- Published by IOP Publishing in Physica Scripta
- Vol. 101 (36) , 365205
Predefined-time synchronization of Kuramoto-oscillator networks via a single-term arctangent-based control protocol
This paper investigates the predefined-time phase agreement (PA) and frequency synchronization (FS) of Kuramoto-oscillator networks using a minimal distributed control protocol. First, an arctangent-based predefined-time stability lemma is established, featuring a single fractional-power arctangent term in the Lyapunov inequality. This arctangent transformation maps the inequality into a fractional-power differential form, ensuring that the settling time is upper-bounded by a user-assignable parameter independent of initial conditions and network topology. A contribution of this work is the identification of a matched-scaling condition between the exponent of the controller nonlinearity and that of the Lyapunov inequality. This condition resolves an asymptotic consistency issue in existing arctangent-based criteria, guaranteeing the validity of the convergence analysis uniformly over the state space. Building upon this lemma, a distributed chattering-free controller is designed for PA under a semi-circle condition on the initial phases. It comprises a single coupling term with a closed-form gain governed by the prescribed settling time, the fractional power, the network size, and the algebraic connectivity of the control graph. For FS, the controller is augmented with a linear frequency-consensus channel that globally neutralizes the cosine-modulated coupling of the differentiated dynamics. The resulting two-term structure eliminates the signum, absolute value, and multiple-exponent components required by conventional polynomial-based protocols. Sufficient conditions for predefined-time PA of identical oscillators and FS of non-identical oscillators are derived. Numerical simulations validate the theoretical results, demonstrating smooth, bounded control signals and scalability to large networks.
- 7 September 2026
- Published by IOP Publishing in Physica Scripta
- Vol. 101 (36) , 365205
Predefined-time synchronization of Kuramoto-oscillator networks via a single-term arctangent-based control protocol
This paper investigates the predefined-time phase agreement (PA) and frequency synchronization (FS) of Kuramoto-oscillator networks using a minimal distributed control protocol. First, an arctangent-based predefined-time stability lemma is established, featuring a single fractional-power arctangent term in the Lyapunov inequality. This arctangent transformation maps the inequality into a fractional-power differential form, ensuring that the settling time is upper-bounded by a user-assignable parameter independent of initial conditions and network topology. A contribution of this work is the identification of a matched-scaling condition between the exponent of the controller nonlinearity and that of the Lyapunov inequality. This condition resolves an asymptotic consistency issue in existing arctangent-based criteria, guaranteeing the validity of the convergence analysis uniformly over the state space. Building upon this lemma, a distributed chattering-free controller is designed for PA under a semi-circle condition on the initial phases. It comprises a single coupling term with a closed-form gain governed by the prescribed settling time, the fractional power, the network size, and the algebraic connectivity of the control graph. For FS, the controller is augmented with a linear frequency-consensus channel that globally neutralizes the cosine-modulated coupling of the differentiated dynamics. The resulting two-term structure eliminates the signum, absolute value, and multiple-exponent components required by conventional polynomial-based protocols. Sufficient conditions for predefined-time PA of identical oscillators and FS of non-identical oscillators are derived. Numerical simulations validate the theoretical results, demonstrating smooth, bounded control signals and scalability to large networks.
- Open Access
- 4 June 2026
- Published by Elsevier in SSRN Electronic Journal
Leader-Follower Multi-Group Consensus Optimal Control of Multi-Agent Systems via Reinforcement Learning: A Multi-Layer Network framework
Multi-task cooperative control is an important frontier research problem in swarm intelligence systems, with theoretical and methodological foundations still largely underdeveloped. This paper proposes a multi-layer network framework to investigate the leader-follower multi-group consensus optimal control of multi-agent systems, in which each layer is equipped with its own leader, and the leaders from different layers are coupled through inter-layer interactions. A novel local tracking error is designed to capture both the intra-layer topological information and the inter-layer coupling one, thereby transforming the optimal control problem into an optimal regulation problem. For the resulting regulation problem, a distributed value iteration (VI) algorithm is proposed to solve the Hamilton-Jacobi-Bellman (HJB) equation, and then two easy-to-use algorithms are developed to obtain the optimal control policy in model-based and model-free framework, respectively based on overall performance index and Actor-Critic networks. Furthermore, the influence of coupling interactions within and across layers on the consensus of multi-agent systems is analyzed both theoretically and numerically, showing that increasing intra-layer coupling strengths accelerates the consensus of multi-agents, and an appropriately large inter-layer coupling strengths similarly promotes the consensus of multi-agents. Finally, some numerical simulations are provided to verify and support the proposed theoretical findings.
- 4 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessLeader-Follower Multi-Group Consensus Optimal Control of Multi-Agent Systems via Reinforcement Learning: A Multi-Layer Network framework
Multi-task cooperative control is an important frontier research problem in swarm intelligence systems, with theoretical and methodological foundations still largely underdeveloped. This paper proposes a multi-layer network framework to investigate the leader-follower multi-group consensus optimal control of multi-agent systems, in which each layer is equipped with its own leader, and the leaders from different layers are coupled through inter-layer interactions. A novel local tracking error is designed to capture both the intra-layer topological information and the inter-layer coupling one, thereby transforming the optimal control problem into an optimal regulation problem. For the resulting regulation problem, a distributed value iteration (VI) algorithm is proposed to solve the Hamilton-Jacobi-Bellman (HJB) equation, and then two easy-to-use algorithms are developed to obtain the optimal control policy in model-based and model-free framework, respectively based on overall performance index and Actor-Critic networks. Furthermore, the influence of coupling interactions within and across layers on the consensus of multi-agent systems is analyzed both theoretically and numerically, showing that increasing intra-layer coupling strengths accelerates the consensus of multi-agents, and an appropriately large inter-layer coupling strengths similarly promotes the consensus of multi-agents. Finally, some numerical simulations are provided to verify and support the proposed theoretical findings.
- 15 April 2026
- Published by Springer Nature in The Journal of Geometric Analysis
- Vol. 36 (5)
Time-Delay Dynamics in Fishery Resources under Periodic Harvesting: Stability and Sustainability Analysis
- 15 April 2026
- Published by Springer Nature in The Journal of Geometric Analysis
- Vol. 36 (5)
Time-Delay Dynamics in Fishery Resources under Periodic Harvesting: Stability and Sustainability Analysis
- 19 March 2026
- Published by Elsevier in Communications in Nonlinear Science and Numerical Simulation
- Vol. 159
Novel fixed/predefined-time stability analysis of discontinuous impulsive systems
- 19 March 2026
- Published by Elsevier in Communications in Nonlinear Science and Numerical Simulation
- Vol. 159
Novel fixed/predefined-time stability analysis of discontinuous impulsive systems
- 10 May 2025
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 113 (16) , 22085-22105
Global dynamics for a class of nonautonomous mosquito models with time delay and switching control strategy
- 10 May 2025
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 113 (16) , 22085-22105
Global dynamics for a class of nonautonomous mosquito models with time delay and switching control strategy
- Open Access
- 24 February 2025
- Published by Springer Nature in Research Square
Global dynamics for a class of nonautonomous mosquito models with time delay and switching control strategy
Mosquitoes serve as the primary vectors for several life-threatening pathogens, including malaria, dengue fever, and Zika virus. To effectively control the transmission of these diseases, a variety of integrated vector management strategies are currently employed to reduce mosquito population densities. This paper focuses on the sterile insect technique (SIT) and incompatible insect technique (IIT), which involves the release of sterile male mosquitoes to suppress wild mosquito populations, and delves into the impact of different release strategies on mosquito population dynamics. We established a switched mosquito population suppression model with time delay. We first investigate the existence and stability of the switching ordinary differential equation model, and derive some criteria for determining whether T-periodic solutions are stable or unstable. By taking the initial function as a solution to the delay-free model, we obtain sufficient conditions for existence of a unique or exactly two periodic solutions. Furthermore, we gain necessary and sufficient conditions for nonexis-tence of periodic solutions. Finally, the numerical results have been shown to further highlighted the effectiveness of our proposed model.
- 24 February 2025
- Published by Springer Nature in Research Square
Open AccessGlobal dynamics for a class of nonautonomous mosquito models with time delay and switching control strategy
Mosquitoes serve as the primary vectors for several life-threatening pathogens, including malaria, dengue fever, and Zika virus. To effectively control the transmission of these diseases, a variety of integrated vector management strategies are currently employed to reduce mosquito population densities. This paper focuses on the sterile insect technique (SIT) and incompatible insect technique (IIT), which involves the release of sterile male mosquitoes to suppress wild mosquito populations, and delves into the impact of different release strategies on mosquito population dynamics. We established a switched mosquito population suppression model with time delay. We first investigate the existence and stability of the switching ordinary differential equation model, and derive some criteria for determining whether T-periodic solutions are stable or unstable. By taking the initial function as a solution to the delay-free model, we obtain sufficient conditions for existence of a unique or exactly two periodic solutions. Furthermore, we gain necessary and sufficient conditions for nonexis-tence of periodic solutions. Finally, the numerical results have been shown to further highlighted the effectiveness of our proposed model.
- 1 April 2024
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 112 (10) , 8739-8760
Global dynamics of a non-smooth SIV system with uncertain effective vaccine protection rate
- 1 April 2024
- Published by Springer Nature in Nonlinear Dynamics
- Vol. 112 (10) , 8739-8760
Global dynamics of a non-smooth SIV system with uncertain effective vaccine protection rate
- 17 November 2023
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 177
Large time behavior in a reaction diffusion epidemic model with logistic source
- 17 November 2023
- Published by Elsevier in Chaos, Solitons & Fractals
- Vol. 177
Large time behavior in a reaction diffusion epidemic model with logistic source