基于CART决策树的调度算法研究  

Research on Scheduling Algorithm Based on CART Decision Tree

作　　者：杨松[1] 王艳红[1] 

机构地区：[1]沈阳工业大学人工智能学院,辽宁沈阳110870

出　　处：《工业控制计算机》2024年第11期152-154,共3页Industrial Control Computer

摘　　要：以往的作业车间存在大量的离线加工数据。基于数据挖掘、调度规则和算法优化相关知识,提出基于贝叶斯优化的改进CART算法来对车间数据挖掘利用,根据车间数据的属性逐步划分节点,生成树状结构,剪枝,最后生成加工规则。通过不同数据算例的实验结果表明,经过贝叶斯优化后的CART算法相较于传统CART算法提高了对数据划分的能力并提升了生成的决策树的准确度。In the past,there were a lot of off-line processing data in the job shop.Based on the knowledge of data mining,scheduling rules and algorithm optimization,this paper proposes an improved CART algorithm based on Bayesian optimization to make use of workshop data mining.According to the attributes of workshop data,nodes are gradually divided,tree structures are generated,branches are pruned,and processing rules are generated finally.The experimental results of different data examples show that compared with traditional CART algorithms,the CART algorithm after Bayesian optimization can improve the ability of data partitioning and improve the accuracy of the generated decision tree.

关 键 词：数据挖掘 贝叶斯优化 CART算法 加工规则 决策树 

分 类 号：TB497[一般工业技术] TP18[自动化与计算机技术—控制理论与控制工程]