Multi-Agent-Based Glass Deep Processing Scheduling System

With the deepening and development of artificial intelligence applications and research,industrial manufacturing is gradually changing from traditional resource-intensive production to intelligent,agile and economical production.As an important part of China's basic investment and construction supply chain,glass deep-processing manufacturing industry is also facing a transition from a traditional high energy consumption and high cost production model to an intelligent,efficient and energy-saving production model.As one of the research hotspots in industrial manufacturing,the production scheduling problem is also an important breakthrough in the current glass deep processing manufacturing industry's transition to smart manufacturing.Due to the complexity of the glass deep processing manufacturing process and the uncertainty of the actual production environment,the glass deep processing production scheduling problem is more complex.Multi-agent can decompose the complex production scheduling problem into a series of related sub-problems.The agents in the system form a coordinated and interacted logical whole through a unified negotiation and communication mechanism,cooperating to process system tasks.This paper adopts multi-agent technology and models the glass deep processing production scheduling environment to design a multi-agent scheduling system that conforms to the characteristics of glass deep processing production and studies from the following aspects:1.Adopt multi-agent system modeling method combined with the production characteristics of glass deep processing,construct a multi-agent glass deep processing scheduling system framework to achieve direct and indirect interaction of each agent in the system to ensure that all agents can fully grasp and use the system Global information to enhance the responsiveness of the glass deep-processing workshop to processing and production tasks.2.Analyze the common problems in the production scheduling of glass deep processing workshops,establish multi-scheduling optimization goals for completion time,completion energy consumption and completion cost,and build glass deep processing workshop and equipment-level processing time,energy consumption and cost models.Research a glass deep processing scheduling method based on the combination of time window and multi-agent task negotiation,construct a processing task resource allocation model of glass deep processing workshop,and implement global optimization of scheduling through multi-agent negotiation method and introducing time window.3.Based on the construction of multi-agent scheduling system,due to the difference in performance of glass deep processing heterogeneous equipment,and the problem of equipment selection for parallel processing in large-scale deep processing workshop scheduling,the characteristics of equipment are classified based on fuzzy membership,based on equipment Feature classification uses genetic algorithm to improve the relevant steps in the algorithm to achieve multi-agent scheduling scheme for glass deep processing with production scheduling target preference.