针对有向无环图结构的多版本分布模式优化

为提高多版本协同设计的整体存储及查询性能，针对具有有向无环图结构、结合完整和增量存储的多版本模型，在研究多版本演化过程、版本间继承关系的基础上，提出可双向恢复的增量版本定义方式，将有向无环图结构转换成带权无向图结构。以版本差异为指标计算版本恢复成本，使用改进的Floyd算法，以计算所有版本最小恢复成本为目标设计多版本分布模式优化算法。分析和实例结果验证，该方法可以提高协同设计多版本的数据存储可靠性和提取效率，并且优化算法具有较高的计算效率。

Optimization of Multi-version Distribution Mode for Directed Acyclic Graph Structure

To improve search efficiency and storage performance of multi-version in collaborative design,this paper presents a definition of incremental version based on analyzing the evolution of versions and relationship between the versions in structure of directed acyclic graph which some versions are saved as entirety and others saved as incremental.After using it,the related versions can be restored in two ways and the structure of multi-version Directed Acyclic Graph（DAG） is changed into the one undirected graph.The algorithm is designed to optimize the distribution of multi-version.In the algorithm,versions recovery cost is calculated from differences between adjacent versions,and the best distribution mode of versions is selected based on the minimum versions recovery cost calculated using the ideas of improved Floyd algorithm.Analysis and examples results demonstrate that this method increases the safety and reliability of version data and improves searching efficiency of all the versions,and the algorithm has higher calculating efficiency at the same time.