Action strategy method based on cascade mode

An action and pattern technology, applied in the field of artificial intelligence, can solve problems such as complex decision-making and multi-agent system dimensions, and achieve the effects of reducing parameter dimensions, saving parameter space, and high decision-making efficiency

Pending Publication Date: 2020-05-12
INFORMATION SCI RES INST OF CETC
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Aiming at the technical problems of complex dimensions and complex decision-making of multi-agent systems, the present invention provides an action strategy method based on cascade mode, which greatly reduces the parameter dimension of the strategy space, reduces a large number of useless parameter spaces, and accelerates the convergence speed at the same time

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Action strategy method based on cascade mode
  • Action strategy method based on cascade mode
  • Action strategy method based on cascade mode

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0034] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in combination with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are exemplary only, and are not intended to limit the scope of the present invention. Also, in the following description, descriptions of well-known structures and techniques are omitted to avoid unnecessarily obscuring the concept of the present invention.

[0035] For N agents, the strategy parameter space is θ={θ 1 ,...,θ N}, the policy set is π={π 1 ,...,π N}, each agent has its own reward mechanism, for agent i, its cumulative reward J(θ i )=E[R i ] gradient is

[0036]

[0037] in is the centralized action-value function of agent i, whose input includes the actions of all agents and related state information, figure 1 Shown is the policy function in the class...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a large-scale action strategy method based on a cascade mode. The large-scale action strategy method comprises the steps of collecting state information vectors of multiple units in real time; calculating decision features based on a neural network function; dividing action modules according to the spatial attribute of actions; constructing a mapping network from the decision features to spatial actions for each action module, and calculating a spatial action probability; constructing a mapping network from the decision features to behavior actions for each action module, and calculating a behavior action probability; and deciding a behavior action based on the behavior action probability, and deciding the spatial position of executing the behavior actions based onempty behavior actions. According to the method, the parameter space is reduced from O (n*m) to O (n+m), the parameter dimension of the strategy space is greatly reduced, a large amount of useless parameter space is reduced, and meanwhile the convergence speed is increased.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to an action strategy method based on a cascade mode. Background technique [0002] A multi-agent system consists of a group of autonomous, interactive entities that share the same environment, perceive the environment through perceptrons and take actions through actuators. According to the structure of the agents in the system, it can be divided into homogeneous multi-agent systems and heterogeneous multi-agent systems. The models of heterogeneous multi-agent systems are not uniform among individuals, which makes there are certain differences in the way individuals perceive the environment or the decision-making space. . The multi-agent game has the characteristics of real-time confrontation, group cooperation, incomplete information game, huge search space, multiple complex tasks, and time-space reasoning. It is a very challenging problem in the field of artificia...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/04G06N3/08
CPCG06N3/08G06N3/045
Inventor 李明强唐思琦陈思高放黄彬城
Owner INFORMATION SCI RES INST OF CETC
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap