Font Size: a A A

Research And Development Of Production Planning And Scheduling System Based On Multi-Agent

Posted on:2007-03-06Degree:MasterType:Thesis
Country:ChinaCandidate:X L WangFull Text:PDF
GTID:2132360182470971Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Multi-Agent technology is multi-subject crossed research field and its application shows the high value. It is the hotspot in the recent study on artificial intelligence. As the middle layer in the CMIS, production scheduling is the key of the CIMS. Under the open and dynamic environment of real manufacture system, it is the NP hard combinatorial optimization problem and it also behaves greatly dynamic characteristics. With the autonomous and cooperative ability of the Multi-Agent, it is possible to solve the complex and dynamic scheduling problem.In this paper, the Multi-Agent technology and its application in the production scheduling were introduced. The plans distributing and assignments scheduling were integrated into the whole system by Multi-Agent technology. The main research work is described as following:1. The popular contract net protocol was analyzed. Then the disadvantages were presented, which were crowding communication andthe low negotiating efficiency. Finally a new improved contract net model was proposed to solve the problem of distributed shop planning.2. Ant colony algorithm was used to solve the job-shop scheduling problem. A new way of ant crawling was proposed. In term of the characteristic of the solutions, an adaptive adjustment process of the volatility coefficient was introduced.3. Combining ant colony algorithm and reinforcement learning, a new job-shop scheduling algorithm based on an adaptive agent was proposed. When the production environment changed, the artificial ant can make decision by the past and the immediate encouragement. Then the distribution of the assignments on the machines was completed.4. The former research achievements were summarized and algorithms library including genetic algorithms, neural network algorithms and the algorithm in this paper was built. Then it was combined in the whole system by the improved contract net protocol. In the end, the production planning and scheduling system was developed.
Keywords/Search Tags:agent, production scheduling, ant colony algorithm, reinforcement learning
PDF Full Text Request
Related items