首页 | 本学科首页   官方微博 | 高级检索  
     

三层虚拟工作流模型的非线性制造工艺多目标优化算法研究EI北大核心CSCD
引用本文:罗智勇,王静远,谢志强,孙广路,杨旭.三层虚拟工作流模型的非线性制造工艺多目标优化算法研究EI北大核心CSCD[J].自动化学报,2022,48(3):896-908.
作者姓名:罗智勇  王静远  谢志强  孙广路  杨旭
作者单位:1.哈尔滨理工大学计算机科学与技术学院 哈尔滨 150080
基金项目:国家自然科学基金面上项目(61772160);;黑龙江省自然科学基金项目(LH2021F030)资助~~;
摘    要:时间、生产质量和成本是加工制造中相互制约的重要参数,平衡此参数使制造工艺最优是一个NP(Non-deterministic polynomial)难题,对此出现了许多优秀的调度方法.然而这些方法的优化对象均为线性工艺,对于普遍存在的非线性工艺却无法调度优化.针对此不足,本文以非线性工艺为优化对象提出了三层虚拟工作流模型Three-VMG(Three-virtual model graph)及其优化算法Three-OVMG(Three-optimal virtual model graph).该模型和算法首先建立非线性工作流,采用虚拟技术寻找虚拟结点进行重构,将其改造为虚拟线性工作流;其次结合工艺特点对模型进行分段,采用逆向分层串归约来实现段内最优解,采用累积最优解来衔接各段间的值;最后根据优化结果自顶向下完成各层资源的优化调度.实验表明,该过程较传统时间最小化优化调度算法具有显著的优化效果,其性能及可操作性也能满足工程要求.

关 键 词:工作流  制造工艺  优化调度  虚拟技术  生产质量
收稿时间:2019-02-27

Multi-objective Optimization Algorithm for Non-linear Manufacturing Process Based on Three-tier Virtual Workflow Model
Affiliation:1.School of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080
Abstract:Time, production quality and cost are important parameters of mutual constraints in manufacturing. Balancing these parameters to make the manufacturing process more reasonable is an NP (non-deterministic polynomial) problem, and there are many excellent scheduling methods to solve this problem. However, the optimization objects of these methods are linear processes, so they can not be scheduled and optimized for the prevalent non-linear processes. To overcome the deficiency, a three-tier virtual workflow model, Three-VMG (three-virtual model graph), and its optimization algorithm, Three-OVMG (three-optimal virtual model graph), are proposed to optimize the non-linear processes in this paper. The model and algorithm first establish a non-linear workflow, using virtual technology to find virtual nodes for reconstruction, thus transforming it into virtual linear workflow. Secondly, according to the process characteristics, the model is segmented. Uses serial reduction strategy to calculate reverse hierarchical solution in each segment, and then cumulative values of solutions between segments. Finally, according to the optimization results, the optimal scheduling of resources at each level is top-down accomplished. Experiments show that the process has a significant optimization effect compared with the traditional time minimization optimal scheduling algorithms. In addition, the performance and operability of the algorithm can meet the engineering requirements.
Keywords:
本文献已被 维普 等数据库收录!
点击此处可从《自动化学报》浏览原始摘要信息
点击此处可从《自动化学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号