基于复杂网络的零部件用量预测方法

在采用有向网络对产品族主结构进行描述的基础上,根据网络中结点入度的双对数坐标分布曲线及其分布规律,提出了产品族零部件关系网络结点入度演化规律,并推导出了结点入度概率密度的解析表达式。通过对零部件种类随产品数变化历史数据的分析,建立了零部件种类与产品数之间的表达关系式。提出了基于给定产品数增量的零部件用量预测方法,以某工业汽轮机产品族数据为例,对部件用量进行了预测,并对预测结果进行了验证,对预测误差进行了分析。

FORECAST METHOD OF USED NUMBER OF PARTS AND COMPONENTS BASED ON COMPLEX NETWORK

Applying directed complex network to model the main structure of product family, according to in-degree bi-logarithmic coordinate distribution curve and distribution rule of nodes of the network, in-degree evolving rule of nodes of the network is presented, analytic expression of in-degree probability density of nodes is derived. Through the analysis of relation between existing kinds of component and existing product number, an expression of the relation between kinds of component and product number is built. A forecast method of increment of component number and parts based on the increment of product is presented. As an example, the component number of industry stream turbine product family is forecasted, forecast result is verified and forecast error is analyzed.