Abstract: Aiming at the problems that the uncertain disturbance factors affect the material distribution time node in complex product assembly shops, a novel forecasting method was proposed to schedule material distribution time node ground on information entropy analysis. The types of uncertain disturbance factors in the complex product assembly lines were defined and quantified by the integrated time requirement factors,and the status of assembly station was defined. Then the state transition probability matrix was also established. The forecasting model of material distribution time node was built, which related with the Markov chain characteristics of station state transition for complex product assembly workshops.A dynamic error compensation method was utilized to amend the initial prediction values through calculating average prediction errors. An evaluation system was constructed to evaluate the forecasting results using information entropy, that the maximum delivery feasibility time and distribution accuracy were selected as the evaluation index. Finally, the forecasting method was validated by an example of a companys grinding machine spindle assembly station. The results show that the method has a significant effect on increasing the feasibility of the maximum distribution and improving the accuracy of the material distribution time.

Key words: material distribution, time node prediction, information entropy, error compensation

摘要： 针对不确定性干扰因素影响复杂产品装配车间物料配送时间准确性的问题，提出了一种基于信息熵评价的物料配送时间节点预测方法。分析了复杂产品装配车间不确定性干扰因素的种类，采用综合时间需求因子量化不确定性干扰因素，定义装配工位状态，并建立工位状态转移概率矩阵。基于工位状态变化的马尔可夫链特性，建立复杂产品装配车间物料配送时间节点预测模型。提出平均预测误差的动态误差补偿方法修正预测值，并选取最大配送可行性时限和配送准确度为评价指标，构建基于信息熵的物料配送系统评价体系，对物料配送时间节点预测方法的有效性进行评价。最后，以某公司磨床主轴装配工位的物料配送历史数据为例，对提出的预测方法进行验证。结果表明，所提出的预测方法有助于增大物料配送系统的最大配送可行性时限、提升物料配送时间节点的准确性。

关键词: 物料配送, 时间节点预测, 信息熵, 误差补偿