大规模非相似混杂制造系统缓冲区容量优化分配技术研究

The large-scale and non-homogeneous manufacturing system, which is with a network organizational structure and in which the machines are disturbed by stochastic breakdown events, is studied and an optimal buffer allocation technique is offered with the limitation that the productivity demand of the system is satisfied and the minimum total buffer capacity is gotten. Firstly, a hierarchical analysis model is set up by the decomposition method of hybrid production lines. Secondly, the homogenized transformation techniques of non-homogeneous manufacturing systems as well as error analysis are studied by conversely applying the system aggregation method of hybrid production lines. Thirdly, the buffer allocation techniques for large-scale non-homogeneous flow lines are offered, and then the buffer allocation techniques for other non-homogeneous and large-scale manufacturing units such as parallel lines, assembly lines, cross/closed lines etc. are studied and offered based on the decomposition technique. Finally, for large-scale manufacturing systems consisting of manufacturing units, the problem of initial solutions is studied, meanwhile, a so-called ‘hierarchical bidirectional recursive hybrid synchronous’ buffer allocation technique, which is mixed together with and proceeds simultaneously with system performance analysis, is proposed. The study lays the foundation of solving similar problems of large-scale complex manufacturing systems disturbed by many stochastic events. Meanwhile, when the modern enterprises which are commonly disturbed by stochastic events, under the huge market competition pressure and with complex collaborative organization structure of arts and crafts optimize the systematic design and especially do globally-optimal scheduling or management of buffer resources, they may be guided by the study. The study standing in the researching frontier of interdisciplinary domain has great value in scientific researching and has great potential power in application.

针对具有网络结构、设备存在随机故障的大规模非相似型制造系统，在满足系统必要生产力要求、总缓冲区容量最小条件下，拟建立其缓冲区容量最优分配理论与技术。首先，基于混杂生产线分解技术原理构造递阶分析模型；逆向应用混杂生产线系统综合方法，研究包括误差分析的非相似型制造系统相似转换技术；在解决规模化非相似型流水线缓冲区容量优化分配问题基础上，基于分解技术原理，研究解决规模化非相似型平行线、装配线、交叉/闭合线等其它典型制造单元的缓冲区容量优化分配问题；最后对制造系统整体，探讨其初解确定方法，并提出与系统性能分析相互融合、同步进行的递阶双向递推混合同步缓冲区容量优化分配技术。项目为解决多种随机事件影响下的复杂制造系统同类问题奠定基础，为面临激烈市场竞争、生产经营常受随机事件影响、具有复杂工艺协作组织结构的现代企业开展缓冲区资源全局性优化管理、调度及相关设计工作提供指导，具有较强科学研究价值和应用潜力。

项目以具有网络结构、设备存在随机故障的大规模非相似型制造系统为研究对象，研究制造网络递阶分解拆分问题，提出了一种生产线递阶分解建模并行寻优技术，彻底解决了本研究领域由系统规模所带来的生产线系统寻优技术难点问题，并考虑设备随机故障等随机事件的影响，建议了一种生产线模糊聚类递阶建模方法；同时研究非相似型制造系统相似转换技术，发展了一种针对强非相似性生产线的相似转换技术，该技术对强非相似型的生产线较传统技术具有更高分析精度；在满足系统必要生产力要求、总缓冲区容量最小条件下，研究规模化制造单元以及复杂制造系统的缓冲区容量最优分配问题，提出了一种新的规模化制造系统缓冲区容量初解确定方法，其解相比于传统方法更接近于最优解，利于系统寻优；同时依据缓冲区容量优化分配问题特点，提出了多种改进型启发式寻优算法；在此基础之上，结合项目递阶分解建模并行寻优的技术成果与传统重叠分解方法，分别提出了规模化流水线、规模化平行线、规模化装配线等制造单元以及面向规模化制造网络系统的缓冲区容量优化分配技术，开发了相关技术软件模块。同时也将部分启发式算法的研究成果在其它领域进行了推广应用。项目成果不仅解决了不可靠制造系统研究领域中规模化制造系统的缓冲区容量优化分配问题，并且具有技术一般性，亦可应用于规模化生产线其它资源配置或相关优化设计、控制问题。. 本项目公开发表科研论文11篇，其中SCI/EI刊源或收录6篇；获发明专利授权1项；在国内外进行学术交流12次；培养研究生14名。