| Autor: |
Liangwei ZHAO, Xiaowei LI, Xuguang CHAI |
| Předmět: |
|
| Zdroj: |
Academic Journal of Manufacturing Engineering; 2020, Vol. 18 Issue 3, p98-105, 8p |
| Abstrakt: |
A large amount of sample data is very important for implementing the dispatching rule mining in the job shop scheduling problem (JSP). This paper attempts to study the dispatching rule mining method for JSP based on Multi-Pass simulation. To this end, it introduces the Multi-Pass simulation mechanism and builds an optimal simulation platform, to provide sample data for the dispatching rule mining in JSP. Then, combining neural network (NN) theory and concept lattice theory, a dispatching rule mining algorithm of job shop based on attribute selection was proposed to avoid the redundant scheduling attributes from affecting the performance of this algorithm. On this basis, the authors further studied the proposed dispatching rule mining algorithm suitable for dynamic production environment. Finally, the simulation results showed that the proposed algorithm based on attribute selection has a better mining accuracy than the rule mining method without considering attribute selection; it is superior to others in terms of performance, and can adaptively obtain the optimal dispatching rules. The research findings have important guiding significance and reference value for solving uncertain and dynamic scheduling problems. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
