In view of the relatively backward situation of automation information technology in manufacturing enterprises, this paper analyzes the manufacturing process of industry, designs and constructs a HD_MAS (Hierarchical Distributed Multi-Agent Systems(MAS)) architecture for collaborative control of process industry process so as to achieve collaborative control optimization for the whole manufacturing process. Based on MAS technology, this paper constructs a collaborative control model for manufacturing system that integrates multiple production units. This model combines several intelligent modules and physical entities organically. Based on the cooperation between Multi-agents, the improved MADDPG (Multi-Agent Deep Deterministic Policy Gradient) algorithm is proposed. A new algorithm framework called DS_MADDPG (Distributed Multi-Agent Deep Deterministic Policy Gradient) is applied to this model. Through the cooperation of Multi-agents, a certain intelligent control system is formed so that it can accurately complete complex production tasks and achieve distributed collaborative control of the manufacturing process, which will drive manufacturing companies to transform into the smart manufacturing.
[1]
Lotfi A. Zadeh.
Toward a perception-based theory of probabilistic reasoning with imprecise probabilities
,
2003
.
[2]
Yuval Tassa,et al.
Continuous control with deep reinforcement learning
,
2015,
ICLR.
[3]
Paulo Leitão,et al.
Agent-based distributed manufacturing control: A state-of-the-art survey
,
2009,
Eng. Appl. Artif. Intell..
[4]
Liu Jian-ying.
Research on Intelligent Decision Support System Based on MAS
,
2008
.
[5]
Alex Graves,et al.
Asynchronous Methods for Deep Reinforcement Learning
,
2016,
ICML.
[6]
Zhang Lieping.
Research on production scheduling problems in process industry based on particle swarm optimization algorithm
,
2012
.
[7]
Min Wu,et al.
A Collaborative Multi-Agent Model with Knowledge-Based Communication for the RoboCupRescue Simulation
,
2006,
International Symposium on Collaborative Technologies and Systems (CTS'06).