系统工程与电子技术 ›› 2019, Vol. 41 ›› Issue (7): 1536-1543.doi: 10.3969/j.issn.1001-506X.2019.07.14

• 系统工程 • 上一篇    下一篇

基于自适应SGD-多智能体的防空资源部署优化

张杰,王刚,宋亚飞,姜浩博,赵方正   

  1. 空军工程大学防空反导学院, 陕西 西安 710054
  • 出版日期:2019-06-28 发布日期:2019-07-09

Optimization of air defense resource deployment based on adaptive SGD-multi-Agent

ZHANG Jie,WANG Gang,SONG Yafei,JIANG Haobo,ZHAO Fangzheng   

  1. Air and Missile Defense College, Air Force Engineering University, Xi’an 710054, China
  • Online:2019-06-28 Published:2019-07-09

摘要: 针对分布式环境下的战场指挥资源部署存在的效率低、速度慢、无法达到预期战略、数据集过大导致计算资源损耗过大等问题,提出了一种分布式环境下多智能体联盟的指挥控制资源部署优化算法。通过对深度学习中的梯度下降算法进行学习率的改进,将原本设定的学习率改为自适应的学习率,进而对指挥控制资源部署进行多智能体联盟的设计。仿真证明了该算法对此问题具有优越的适应性,可以高效地解决分布式环境下的多智能体联盟的指挥控制资源部署优化问题。

关键词: 深度学习, 分布式多智能体, 资源部署优化, 梯度下降算法, 智能体联盟

Abstract: For the deployment of battlefield command resources in a distributed environment, there are problems such as low efficiency, slow speed, failure to reach the expected strategy, and excessive data set, resulting in excessive computing resource loss, a multi-agent alliance command in distributed environment is proposed. The control resource deployment optimization algorithm, by improving the learning rate of the gradient descent algorithm in deep learning, changing the originally set learning rate to adaptive learning rate, and then designing a multi-agent alliance for command and control resource deployment. It proves that the algorithm has superior adaptability to this problem, and can effectively solve the problem of command and control resource deployment optimization of multi-agent alliance in the distributed environment.

Key words: deep learning, distributed multi-agent, resource deployment optimization, gradient descent algorithm, agent alliance