Scholarlink

<p>This paper studies distributed resource allocation in multi-agent systems that in practice are often subject to unknown dynamics and external disturbances as well as limited communication resources. In order to solve the issue, this paper adopts an adaptive neural network to compensate the unknown dynamics and external disturbances based on which a novel distributed resource allocation strategy is developed. In addition, for reducing the amount of data transmitted among agents, a dynamic encoding/decoding scheme is introduced to quantize the transmitted data. A dynamic event-triggered communication mechanism is developed to determine the instants when the quantized information is sent to neighboring agents, thereby effectively reducing the communication frequency. By constructing an appropriate Lyapunov function, it is analyzed that all agents’ decisions can converge to optimal resource allocation with a small error. Furthermore, the proposed event-triggered quantization communication mechanism is free of Zeno behavior. Finally, numerical examples are given to illustrate the results.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A neural network-based distributed algorithm for resource allocation with event-triggered quantization communication

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Jian Ma,
  • Xin Cai,
  • Bingpeng Gao,
  • Mingzhe Dai

摘要

This paper studies distributed resource allocation in multi-agent systems that in practice are often subject to unknown dynamics and external disturbances as well as limited communication resources. In order to solve the issue, this paper adopts an adaptive neural network to compensate the unknown dynamics and external disturbances based on which a novel distributed resource allocation strategy is developed. In addition, for reducing the amount of data transmitted among agents, a dynamic encoding/decoding scheme is introduced to quantize the transmitted data. A dynamic event-triggered communication mechanism is developed to determine the instants when the quantized information is sent to neighboring agents, thereby effectively reducing the communication frequency. By constructing an appropriate Lyapunov function, it is analyzed that all agents’ decisions can converge to optimal resource allocation with a small error. Furthermore, the proposed event-triggered quantization communication mechanism is free of Zeno behavior. Finally, numerical examples are given to illustrate the results.