遗传神经网络在工时定额计算中的应用

工时定额数据量大、影响因素多，使用常规拟合方法计算工时定额比较困难。为提高工时定额计算的正确性，采用人工神经网络技术，在MATLAB中建立了工时定额计算神经网络模型。针对BP神经网络存在易陷入局部最小值、收敛速度慢等不足，引入标准遗传算法来优化神经网络的权值和阈值。实验结果表明，基于实数编码的遗传算法优化速度快，优化后的神经网络迅速收敛，神经网络模型的测试误差低于5％。遗传神经网络可以克服单独使用神经网络时存在的缺点，训练好的模型在工时定额计算时正确性较高，有较好的实用价值。

The application of genetic neural network in calculating man-hour quota

Man-hour quota has large data size and many influencing factors, it＇s hard to calculate man-hour quota by using conventional fitting methods. In order to improve computation accuracy of man-hour quota, BP （Back-Propagation） neural network was used to build man-hour quota calculation model in MATLAB. For the BP neural network has disadvantages like getting Into local minimum point easily, low convergence speed and so on, standard genetic algorithm （GA） was integrated to optimize the weight and bias of neural network. Experimental result shows genetic algorithm based on real coding has fast velocity as optimizing the parameters of neural network, the optimized neural network converge rapidly, and the test error of the neural network models are under five percent. The genetic neural network can overcome the disadvantages of using neural network independently, and the trained model has relatively high accuracy in calculating man-hour quota. The genetic neural network has good value in the man-hour quota calculating.