Efficient Predictive Regulation Algorithms for AGV System in Industrial Internet

Abstract

In recent years, the industrial Internet has developed rapidly. In order to improve the reliability, real-time, and economy, Automated Guided Vehicle (AGV) in intelligent manufacturing system becomes an indispensable technology. However, the current AGV system relies too much on the fixed network bandwidth environment in information transmission and management. When the traffic demand changes frequently, this form of network configuration lacks network resource management mechanism. Further, it leads to the problems of delay, waste of network flow, and inability to dynamically allocate network resources. So it is vital to improve the AGV system. Therefore, this paper proposes three predictive control algorithms and a Network Cable Scheduling algorithm to manage the network resources. They are Markov Chain Linear Programming Regulation (MCLPR) algorithm, Prophet Linear Programming Regulation (PLPR) algorithm, and Machine Learning Linear Programming Regulation (MLLPR) algorithm. The experimental results show that PLPR and MLLPR algorithm have high efficiency in the aspect of regulation. MLLPR algorithm has the lowest cost. MLLPR algorithm has the strongest leakage limitation ability, followed by PLPR algorithm. The balance regulation efficiency of MLLPR in none “4 + 1” mode is the highest in different network cable modes.