Qingshan Liu’s research while affiliated with Southeast University and other places

... This leads to the wellposedness of the problem, which means that optimal solutions always exist. Quadratic programming models have been successfully applied to find optimal decisions for various problems in many fields from power/energy systems management [15,16,17,18,19,20,21], robotics JOURNAL OF FUNDAMENTAL MATHEMATICS AND APPLICATIONS (JFMA) VOL. 8 NO. 1 (2025) Available online at www.jfma.math.fsm.undip.ac.id [22,23] to microgrid [24], showing its superiority and versatility. ...

Reference:

COORDINATING AND OPTIMIZING TWO-WAREHOUSE INVENTORY SYSTEMS: A MATHEMATICAL PROGRAMMING APPROACH

... The uncertainty and complexity of the control system varies from patient to patient, so it is important to design a controller that meets the requirements of the unknown dynamics of the exoskeleton. Neural networks (NNs) are extensively applied to the control of uncertain, nonlinear, and complex systems due to their outstanding ability to approximate functions with high levels of accuracy [35][36][37][38][39]. Ref. [40] employed a neural network for disturbance estimation, significantly enhancing the robustness of the control approach. ...

Reference:

Adaptive Impedance Control of a Human–Robotic System Based on Motion Intention Estimation and Output Constraints

... Specifically, the Hopfield networks are developed for combinatorial optimization [14], [15]. Since then, a variety of neurodynamic optimization models have been developed for solving numerous optimization problems [16]. Despite the progress, it is acknowledged that an individual neurodynamic model faces challenges in effectively addressing combinatorial optimization problems because gradient-driven neurodynamic models are prone to be trapped in local minima. ...

Reference:

Bi-clustering of Binary Data via Neurodynamics-driven Binary Matrix Factorization

... These applications have fueled the creation of decentralized algorithms, allowing a group of nodes to jointly optimize the cumulative local cost functions. Within decentralized optimization, each agent possesses knowledge of its individual objective function and performs computations locally [10][11][12]. Moreover, they engage in communication with neighboring agents. ...

Reference:

Decentralized non-convex online optimization with adaptive momentum estimation and quantized communication

... To ease the oscillation, a damping term −k vṗ1 with k v = 0.5 is added to the control law as expressed in (9). We also compare our method with two other DRLbased approaches, namely PPO [36] and SAC [37]. The hyperparameters of the training are listed in Table I. ...

Reference:

Learning Resilient Formation Control of Drones with Graph Attention Network

... Related Work: The encirclement problem has been mainly addressed in the literature by framing the task as a formation control problem in which robots implement a distributed protocol to place themselves around the target [2]- [6]. In these formulations, the team is not able to adapt its configuration to multi-objective tasks as, e.g., monitoring sensitive spots and taking into account density functions in the environment. ...

Reference:

Multi-Robot Target Monitoring and Encirclement via Triggered Distributed Feedback Optimization

... Given the challenges of the problem, some of the works give approximate solutions or alternatives that are feasible but not strictly predefined-time optimal. For instance, the method in [20] can achieve predefined-time consensus but asymptotic optimization; the algorithms in [21]- [23] can drive the system to reach a feasible domain in a given time but the optimal solution is only approachable asymptotically; the approximate solutions in [24]- [26] can converge to the neighborhood of the optimizer, but not exact optimization in a given time. ...

Reference:

Corrigendum to "Balance of Communication and Convergence: Predefined-time Distributed Optimization Based on Zero-Gradient-Sum"

... Event-triggered control and communication is a trending research area, which is being applied to diverse fields such as self-driving vehicles [15], [21], cyber-physical systems [22], robotic manipulators [23], multi-agent systems [24], hard-disk drives [25], and so on. In the present work, this technique will be used from the point of view of periodic event-triggered communication (PETC) [26], where the event-triggered conditions are periodically evaluated. In this way, the PETC mechanism ensures a minimum inter-event time, which avoids the well-known Zeno behaviour, and makes easier the digital implementation [27]. ...

Reference:

Resource-Efficient Path-Following Control for a Self-Driving Car in a Networked Control System

... In order to further cope with the time-varying inequality constraints with complex geometry, [21] and [35] developed several kinds of effective dynamical systems by means of penalty method. While [53] developed a dynamical system with specified-time convergence for unconstrained time-varying distributed optimization problems. These methods use real-time data and dynamic control technology to improve the ability to cope with local constraints and objective functions, but they cannot be directly used to deal with coupling constraints involving global decision variables. ...

Reference:

Time-varying distributed fuzzy resource allocation for multi-agent system over switching topology

... Nonlinear filtering is a very important aspect of target tracking. In practical applications, almost all control systems are nonlinear, and linearity is an approximate description of nonlinearity to a certain extent [22]. Therefore, studying nonlinear filtering algorithms to reliably and accurately track targets is the main purpose of designing target tracking systems. ...

Reference:

The GM-JMNS-CPHD Filtering Algorithm for Nonlinear Systems Based on a Generalized Covariance Intersection