Abstract: In order to improve the accuracy of the man-hour calculation under the environment of mass customization, a method for determining man-hour based on error correction coefficient is proposed due to the complexity of man-hour influence factors. In this paper, the evaluation system of man-hour influence factors is established. The fuzzy linguistic questionnaire method is used to evaluate the influence factors and the fuzzy comprehensive evaluation of key influence factors is obtained. By using BP neural network method, the mapping relationship between influence factors and error correction coefficient is determined after training network through Matlab. Then the error correction coefficient is obtained and the precise calculation of man-hour is achieved. This method is verified with strong operability and high accuracy by an example.

Key words: mass customization, man-hour, error correction coefficient, Back Propagation（BP） neural network, fuzzy linguistic questionnaire, matlab

摘要： 为了提高大规模定制环境下工时计算的准确度，针对工时影响因素的复杂性，提出了一种基于误差修正系数的工时计算方法。建立工时影响因素的评价体系，运用了模糊语言量表法对影响因素进行模糊评价，得出了各关键影响因素的模糊综合评价值。采用BP神经网络方法，通过MATLAB进行网络训练，确定了各影响因素与误差修正系数之间的映射关系，得出了误差修正系数，实现了工时的精确计算。通过实例验证了该方法可操作性强、准确度高等优点。

关键词: 大规模定制, 工时, 误差修正系数, 反向传播（BP）神经网络, 模糊语言量表, matlab