JZJishi Zhang

Jishi Zhang

*This is an algorithmically generated author record based on publicly available information.

Jishi Zhang has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Adaptive Dynamic Programming and Reinforcement Learning for Optimal Control of Complex Nonlinear Systems: From Single-Agent Foundations to Multi-Agent Safe Autonomy

    by Jishi Zhang

    Adaptive Dynamic Programming (ADP) and Reinforcement Learning (RL) have emerged as transformative frameworks for solving optimal control problems in complex nonlinear systems, effectively circumventing the curse of dimensionality and the intractability of the Hamilton-Jacobi-Bellman (HJB) equation. This survey provides a comprehensive and systematic review of the theoretical foundations and practical advancements in ADP and RL spanning over a decade of intensive research. We trace the intellectual evolution from foundational online algorithms for single discrete-time and continuous-time nonlinear systems to advanced decentralized, game-theoretic, and event-triggered control strategies for interconnected and multiagent systems. Special emphasis is placed on the integration of safety constraints via Control Barrier Functions (CBFs), robust handling of asymmetric input saturation and unmatched disturbances, fault-tolerant architectures, and the pursuit of resource efficiency through dynamic event-triggered mechanisms. The survey synthesizes rigorous convergence guarantees, Lyapunovbased stability analyses that explicitly account for neural network approximation errors, and algorithmic innovations such as experience replay, concurrent learning, and integral reinforcement learning. By establishing systematic connections between singleagent and multi-agent frameworks, this review provides a unified perspective on the field and outlines promising future directions, including meta-reinforcement learning, resilient control against cyber-physical attacks, and large-scale human-robot collaboration.

  • Open Access

    A Survey of Learning-Based Optimal and Safe Control of Complex Nonlinear Systems with Adaptive Critic Architectures and Event-Triggered Mechanisms

    by Jishi Zhang, Chunbin Qin

    Adaptive Dynamic Programming (ADP), synergizing reinforcement learning, neural networks, and optimal control theory, has emerged as a powerful data-driven methodology for solving complex decision-making and control problems in nonlinear systems. This survey provides a comprehensive and unified overview of the core principles and recent advances in ADP-based control, structured around an integrated "Intelligent Critic-Decision-Communication" framework. We dissect this framework into four foundational pillars: (1) the theoretical underpinnings rooted in the Hamilton-Jacobi-Bellman/Hamilton-Jacobi-Isaacs (HJB/HJI) equations and the mitigation of the curse of dimensionality; (2) the algorithmic core revolving around the adaptive critic-actor architecture and its enhanced learning mechanisms, including experience replay, concurrent learning, meta-reinforcement learning, and transformer architectures; (3) the augmentation of control objectives to handle real-world constraints such as safety, input saturation, and external disturbances through barrier functions and game theory; and (4) the resource-aware implementation strategies with a particular emphasis on the deep integration of event-triggered mechanisms-from static and dynamic event-triggering to periodic, learning-based, and self-triggered control-along with output feedback observers for multi-agent systems. By synthesizing a vast body of literature spanning foundational theory to industrialgrade applications, this paper offers a clear and structured perspective on the state of the art, elucidating the deep interconnections between algorithmic design, constraint satisfaction, and practical deployment.

  • Open Access

    Event-Triggered Adaptive Dynamic Programming: A Comprehensive Review of Methodologies, Triggering Strategies, and Emerging Frontiers

    by Chunbin Qin, Jishi Zhang, Dehua Zhang

    Adaptive dynamic programming (ADP) has emerged as a powerful framework for solving optimal control problems of nonlinear systems by circumventing the "curse of dimensionality" inherent in conventional dynamic programming. However, the iterative nature of ADP algorithms and the frequent updates required in networked control systems impose substantial computational and communication burdens. Event-triggered control (ETC) offers an efficient remedy by executing control updates only when necessary, thereby significantly reducing resource consumption while preserving closed-loop stability and performance. This paper provides a comprehensive review of the integration of event-triggered mechanisms with ADP. We begin by presenting the foundational principles of ADP and event-triggered control, followed by a systematic taxonomy of triggering strategies, including static event-triggering, dynamic event-triggering, and self-triggered mechanisms. The theoretical frameworks, design methodologies, and stability guarantees for each category are examined in detail. We then survey the application of eventtriggered ADP across diverse domains, including robust control under uncertainties and constraints, multi-agent consensus, datadriven control, fault-tolerant control, and safety-critical systems. Throughout the review, we highlight the progressive shift from periodic time-triggered paradigms toward intelligent, adaptive sampling strategies that balance control performance with resource efficiency. Finally, we identify open challenges and outline promising directions for future research, including integration with advanced learning techniques, resilience against cyberphysical threats, and certification for real-world deployment.

  • Direction-aware and boundary-enhanced hybrid Mamba-CNN network for building footprint extraction

    by Cancong Dai, Xiaoli Liang, Jiayao Wang, Bing Li, Jishi Zhang, Man Wang, Qizhen Wei