Gangshan Jing

• Doctor of Philosophy
• Assistant Professor at Chongqing University

Recently introduced distributed zeroth-order optimization (ZOO) algorithms have shown their utility in distributed reinforcement learning (RL). Unfortunately, in the gradient estimation process, almost all of them require random samples with the same dimension as the global variable and/or require evaluation of the global cost function, which may i...

Achieving distributed reinforcement learning (RL) for large-scale cooperative multi-agent systems (MASs) is challenging because: (i) each agent has access to only limited information; (ii) issues on scalability and sample efficiency emerge due to the curse of dimensionality. In this paper, we propose a general distributed framework for sample effic...

Origami is known as a traditional art of paper folding. It has attracted extensive attention due to its self-folding mechanism, shape-morphing capability, and deployable structures. This article develops network-based methods for designing and controlling a three-dimensional (3D) triangulated origami tessellation to approximate multiple surfaces. T...

Angle-constrained formation control has attracted much attention from control community due to the advantage that inter-edge angles are invariant under uniform translations, rotations, and scalings of the whole formation. However, almost all the existing angle-constrained formation control methods are limited to undirected triangulated sensing grap...

This paper investigates the fuel-optimal guidance problem of the end-to-end human Mars entry, powered descent, and landing (EDL) mission. It applies a unified modeling scheme and develops a computationally efficient new optimization algorithm to solve the multi-phase optimal guidance problem. The end-to-end EDL guidance problem is first modeled as...

The main challenge of large-scale cooperative multi-agent reinforcement learning (MARL) is two-fold: (i) the RL algorithm is desired to be distributed due to limited resource for each individual agent; (ii) issues on convergence or computational complexity emerge due to the curse of dimensionality. Unfortunately, most of existing distributed RL ref...

Existing distributed cooperative multi-agent reinforcement learning (MARL) frameworks usually assume undirected coordination graphs and communication graphs while estimating a global reward via consensus algorithms for policy evaluation. Such a framework may induce expensive communication costs and exhibit poor scalability due to requirement of glo...

Recently introduced distributed zeroth-order optimization (ZOO) algorithms have shown their utility in distributed reinforcement learning (RL). Unfortunately, in the gradient estimation process, almost all of them require random samples with the same dimension as the global variable and/or require evaluation of the global cost function, which may i...

Designing the optimal linear quadratic regulator (LQR) for a large-scale multi-agent system (MAS) is time-consuming since it involves solving a large-size matrix Riccati equation. The situation is further exasperated when the design needs to be done in a model-free way using schemes such as reinforcement learning (RL). To reduce this computational...

We address model-free distributed stabilization of heterogeneous multi-agent systems using reinforcement learning (RL). Two algorithms are developed. The first algorithm solves a centralized linear quadratic regulator (LQR) problem without knowing any initial stabilizing gain in advance. The second algorithm builds upon the results of the first alg...

We address the problem of model-free distributed stabilization of heterogeneous multi-agent systems using reinforcement learning (RL). Two algorithms are developed. The first algorithm solves a centralized linear quadratic regulator (LQR) problem without knowing any initial stabilizing gain in advance. The second algorithm builds upon the results o...

This paper studies angle-based sensor network localization (ASNL) in a plane, which is to determine locations of all sensors in a sensor network, given locations of partial sensors (called anchors) and angle constraints based on bearings measured in the local coordinate frame of each sensor. We firstly show that a framework with a non-degenerate bi...

This paper investigates the six-degree-of-freedom (6-DoF) entry guidance problem for the Human Mars exploration mission. For the Human-scale entry, powered descent, and landing mission, it is required to use aerodynamic forces to decelerate the vehicle during the entry phase. Instead of assuming the entry vehicle as a point mass, we consider both t...

Individual agents in a multi-agent system (MAS) may have decoupled open-loop dynamics, but a cooperative control objective usually results in coupled closed-loop dynamics thereby making the control design computationally expensive. The computation time becomes even higher when a learning strategy such as reinforcement learning (RL) needs to be appl...

Designing the optimal linear quadratic regulator (LQR) for a large-scale multi-agent system (MAS) is time-consuming since it involves solving a large-size matrix Riccati equation. The situation is further exasperated when the design needs to be done in a model-free way using schemes such as reinforcement learning (RL). To reduce this computational...

This paper proposes two reinforcement learning (RL) algorithms for solving a class of coupled algebraic Riccati equations (CARE) for linear stochastic dynamic systems with unknown state and input matrices. The CARE are formulated for a minimal-cost variance (MCV) control problem that aims to minimize the variance of a cost function while keeping it...

This paper studies angle-based sensor network localization (ASNL) in a plane, which is to determine locations of all sensors in a sensor network, given locations of partial sensors (called anchors) and angle constraints based on bearings measured in the local coordinate frame of each sensor. We firstly show that a framework with a non-degenerate bi...

Sensor network localization (SNL) is to determine physical coordinates of all sensors in a network given global coordinates of anchors and available measurements among sensors and anchors. Two challenges related to SNL are to find conditions leading to a uniquely localizable network and develop effective and efficient methods to solve SNL problems....

In this paper, the mixed equilibrium problem is solved by a multi-agent network. The objective for agents is to cooperatively find a point in a convex set, at which the sum of some local bifunctions with a free variable is non-negative. To address this problem, we propose a distributed extragradient algorithm based on a consensus strategy. By imple...

This technical note aims to design a decentralized control strategy for multiple double-integrator agents to achieve flocking while maintaining an angle-constrained triangulated formation shape in the plane. We design a novel decentralized formation controller, which can (i) steer agents to achieve common velocity while meeting some angle constrain...

In this paper, the problem of online distributed optimization is investigated, where the sum of locally dynamic cost functions is considered to be strongly pseudoconvex. To address this problem, we propose an online distributed algorithm based on an auxiliary optimization strategy. The algorithm involves each agent minimizing its own cost function...

Distributed Algorithm for Solving Convex Inequalities

This paper investigates distributed coordination control problems under a state-dependent communication graph with several inherent links. In the network, an interaction arises between two agents if either their states differ by less than a fixed range or an inherent communication link exists between them. By considering that each agent has a state...

In this paper, the distributed Nash equilibrium (NE) searching problem is investigated, where the feasible action sets are constrained by nonlinear inequalities and linear equations. Different from most of the existing investigations on distributed NE searching problems, we consider the case where both cost functions and feasible action sets depend...

This paper introduces the notion of weak rigidity to characterize a framework by pairwise inner products of inter-agent displacements. Compared to distance-based rigidity, weak rigidity requires fewer constrained edges in the graph to determine a geometric shape in an arbitrarily dimensional space. A necessary and sufficient graphical condition for...

This paper introduces the notion of weak rigidity to characterize a framework by pairwise inner products of inter-agent displacements. Compared to distance-based rigidity, weak rigidity requires fewer constrained edges in the graph to determine a geometric shape in an arbitrarily dimensional space. A necessary and sufficient graphical condition for...

This paper presents an angle-based approach for distributed formation shape stabilization of multi-agent systems in the plane. We develop an angle rigidity theory to study whether a planar framework can be determined by angles between segments uniquely up to translations, rotations, scalings and reflections. The proposed angle rigidity theory is ap...

This paper presents an angle-based approach for distributed formation shape stabilization of multi-agent systems in the plane. We develop an angle rigidity theory to study whether a planar framework can be determined by angles between segments uniquely up to translations, rotations, scalings and reflections. The proposed angle rigidity theory is ap...

In this paper, a class of convex feasibility problems (CFPs) are studied for multi-agent systems through local interactions. The objective is to search a feasible solution to the convex inequalities with some set constraints in a distributed manner. The distributed control algorithms, involving subgradient and projection, are proposed for both cont...

In this paper, a distributed subgradient-based algorithm is proposed for continuous-time multi-agent systems to search a feasible solution to convex inequalities. The algorithm involves each agent achieving a state constrained by its own inequalities while exchanging local information with other agents under a time-varying directed communication gr...

In this paper, a distributed subgradient-based algorithm is proposed for continuous-time multi-agent systems to search a feasible solution to convex inequalities. The algorithm involves each agent achieving a state constrained by its own inequalities while exchanging local information with other agents under a time-varying directed communication gr...

In this paper, we study the consensus problem for continuous-time and discrete-time multi-agent systems in state-dependent switching networks. In each case, we first consider the networks with fixed connectivity, in which the communication between adjacent agents always exists but the influence could possibly become negligible if the transmission d...

In this paper, we consider the flocking problem of multi-agent systems with multiple groups. First, some algorithms using local information are designed to divide the agents into any pre-assigned number of groups in fixed and switching heterogeneous networks, respectively. Based on algebraic graph theory and Barbalat’s lemma, convergence criteria a...

In this paper, we consider group flocking of multi-agent systems in which agents are dispersed to different subgroups. By using local information, the flocking algorithms are proposed to solve group flocking problem in heterogeneous and homogeneous networks, respectively. For each given algorithm, the corresponding criterions are established to ens...