Multi-agent graphical games with input constraints: an online learning solution

Wang, Tianxiang; Wang, Bing‐Chang; Liang, Yong

doi:10.1007/s11768-020-0013-6

Cited by 8 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Remark Constraint matrix

{\bar{U}}_i

indicates that the input has different constraints in each dimension. In contrast to the methods for dealing with input constraints in [36–41], the constraint space here can be viewed as a cube rather than a sphere and thus better suited to practical requirements. Moreover, compared with the output synchronization problem in [29, 31–33], the proposed Problem 2 considers the input saturation phenomenon by inserting a nonquadratic input term into the performance function.…”

Section: Preliminaries and Problem Formulationmentioning

confidence: 99%

“…Therefore, the optimal control policy in each state can be obtained by (30) without any dynamics knowledge of the augmented system. Compared with the VFA methods in [36][37][38][39][40][41], wherein the Bellman Equation ( 16) and the improved control policy (20) are utilized to iteratively find the optimal solution, Algorithm 1 uses the Q-function approximation with ( 29) and (30) to find the input-constrained optimal solution without involving any dynamics knowledge of the augmented system, especially the input dynamics Bi in (20).…”

Section: Lemma 7 Consider Arbitrary Initial Admissible Control Policy Umentioning

confidence: 99%

“…An effective approach to deal with the input constraint is to pre-embed it into the performance function of the optimization problem [35] and iteratively find the opti-mal constrained control input using the actor-critic neural network (NN) structure of RL. In fact, based on this approach, several studies have been presented to derive input-constrained optimal control protocols for both CT [36][37][38] and DT systems [39][40][41]. By introducing a nonquadratic input term, the Hamilton-Jacobi-Bellman (HJB) equation can be solved by value function approximation (VFA) and policy iteration (PI) algorithm, and the corresponding optimal control input is thereby obtained.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Input‐constrained optimal output synchronization of heterogeneous multiagent systems via observer‐based model‐free reinforcement learning

Zhang

Jia

2023

Asian Journal of Control

View full text Add to dashboard Cite

This paper presents a new formulation of input‐constrained optimal output synchronization problem and proposes an observer‐based distributed optimal control protocol for discrete‐time heterogeneous multiagent systems with input constraints via model‐free reinforcement learning. First, distributed adaptive observers are designed for all agents to estimate the leader's trajectory without requiring its dynamics knowledge. Subsequently, the optimal control input associated with the optimal value function is derived based on the solution to the tracking Hamilton‐Jacobi‐Bellman equation, which is always difficult to solve analytically. To this end, motivated by reinforcement learning technique, a model‐free Q‐learning policy iteration algorithm is proposed, and the actor‐critic neural network structure is implemented to iteratively find the optimal tracking control input without knowing system dynamics. Moreover, inputs of all agents are constrained in the permitted bounds by inserting a nonquadratic function into the performance function, where input constraints are encoded into the optimization problem. Finally, a numerical simulation example is provided to illustrate the effectiveness of the proposed theoretical results.

show abstract

“…Remark Constraint matrix

{\bar{U}}_i

Section: Preliminaries and Problem Formulationmentioning

confidence: 99%

Section: Lemma 7 Consider Arbitrary Initial Admissible Control Policy Umentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Input‐constrained optimal output synchronization of heterogeneous multiagent systems via observer‐based model‐free reinforcement learning

Zhang

Jia

2023

Asian Journal of Control

View full text Add to dashboard Cite

show abstract

“…La aparición del COVID 19 en China -y que se propagase en todo el mundo durante el 2020llevó a una modificación sin precedentes sobre la forma de vivir y, en específico de aprender, mismas que se tuvieron que adaptar a un confinamiento impuesto por cada gobierno nacional (Bojović et al, 2020;José et al, 2020;Luna-Nemecio & Tobón, 2021;Nassr et al, 2020;Villegas-Ch et al, 2020;Wang et al, 2020). En situaciones inesperadas como ésta, la educación basada en herramientas virtuales puede ser de gran ayuda (Romero-Rodriguez et al, 2020).…”

Section: Introductionunclassified

El software como apoyo didáctico en la enseñanza de la electrónica digital en la educación superior en México [Software as a didactic support in the teaching of digital electronics in higher education in Mexico]

Pérez

2021

runas

View full text Add to dashboard Cite

La educación actual ha sufrido diversas modificaciones, sobre todo con la aparición del COVID-19 en el mundo. Entre las herramientas tecnológicas que apoyan la actividad didáctica, están los softwares educativos; éstos permiten la simulación de prácticas en laboratorios virtuales que complementan el aprendizaje y permiten demostrar principios teóricos. En el presente artículo se llevó a cabo un análisis documental sobre el uso del software en la electrónica digital, se elaboró un temario común de la materia electrónica digital de las licenciaturas que oferta el Tecnológico Nacional de México. Finalmente se llevó a cabo un análisis del uso del software determinado hacia la materia en cuestión y se estableció una relación que indica el nivel de adecuación de este. Current education has undergone various modifications, especially with the appearance of COVID-19 in the world. Among the technological tools that support the didactic activity is simulation software; These allow to carry out virtual laboratory practices that complement the learning and allow demonstrating theoretical principles. In this article, a documentary analysis was carried out on the use of software in digital electronics, a common agenda of digital electronic matter of the degrees offered by the Tecnológico Nacional de México was elaborated. Finally, an analysis of the use of the software determined towards the subject in question was carried out and a relationship was established that indicates the level of adequacy. Learning styles were identified in the group of surveyed students, and their relationship with the level of satisfaction and achievement of the software.

show abstract

Neural network-based adaptive decentralized learning control for interconnected systems with input constraints

Luo

et al. 2021

Control Theory Technol.

View full text Add to dashboard Cite

Multi-agent graphical games with input constraints: an online learning solution

Cited by 8 publications

References 20 publications

Input‐constrained optimal output synchronization of heterogeneous multiagent systems via observer‐based model‐free reinforcement learning

Input‐constrained optimal output synchronization of heterogeneous multiagent systems via observer‐based model‐free reinforcement learning

El software como apoyo didáctico en la enseñanza de la electrónica digital en la educación superior en México [Software as a didactic support in the teaching of digital electronics in higher education in Mexico]

Neural network-based adaptive decentralized learning control for interconnected systems with input constraints

Contact Info

Product

Resources

About