针对综采工作面液压支架跟机自动化过程中移架动作存在的丢架、推移不到位等问题，提出了基于遗传算法（GA）与BP神经网络组合模型的控制方法。通过建立BP神经网络控制器为主体的反馈控制，将支架的运动参数作为模型的输入，神经网络控制器用来计算实际输出与理想输出之间误差，判别是否需要回调控制，并添加遗传算法来优化更新模型的各层阈值和权值，从而得到网络模型的最优解，最终由执行部分来完成输出动作。组合网络模型具有良好的非线性特性，可以更好的满足非线性环境，利用神经网络的预测值与实际输出的差值来得到拟合曲线。通过对BP神经网络模型、GA模型、GA-BP组合模型的均方误差（mse）分析，判断出GA-BP组合模型具有更快的训练速度和更高的预测准确率。相比较于单一的BP神经网络模型和GA模型，GA-BP组合模型可以很大程度地提高液压支架跟机过程中的推移精度，从而更好地适应综采工作面的环境和设备变化。基于对模型稳定性的分析，绘制组合网络的适应度曲线，种群在第5次迭代后趋于收敛，在第5次到第15次迭代的适应度值就已基本达到稳定，在迭代第15次后种群已达到最优参数集且恒定不变。采用上述方案的液压支架电液控制系统能够自主感知设备各项运动参数的变化，实现支架自身的静态调整和动态演化，可为综采工作面无人化建设提供技术支撑。

In view of the problems of frame loss and improper support movement in the automatic process of hydraulic support following machine in fully mechanized mining working, a control method was proposed based on Genetic Algorithm (GA) and BP neural network combined model. Through the establishment of BP neural network controller as the main feedback control, the motion parameters of the support are used as the input of the model. The neural network controller  is used to calculate the error between the actual output and the ideal output, to determine whether callback control is required. In order to optimize the thresholds and weights of each layer of the updated model to obtain the optimal solution of the network model, and finally get the optimal solution of the network model, and the execution part completes the output action. The combined network model has good nonlinear characteristics and can better meet the nonlinear environment. The difference between the predicted value of the neural network and the actual output is used to obtain the fitting curve. By analyzing the mean square error (mse) of the BP neural network model, GA model, and BP-GA combined model, it is justified that the GA-BP combined model has faster training speed and higher prediction accuracy. Compared with a single BP neural network model and GA model, the GA-BP combined model can greatly improve the accuracy of the hydraulic support in the process of following the machine, so as to better adapt to the changes in the environment and equipment in the fully mechanized mining working. Based on the analysis of model stability, the fitness curve of the combined model was drawn. The population tends to converge after the 5th iteration, and the fitness value from the 5th to 15th iteration is basically stable, and after the 15th iteration the population has reached the optimal parameter and became constant. The hydraulic support electro-hydraulic control system adopting the above schemecan autonomously sense the changes of various motion parameters of the equipment, realizethe static adjustment and dynamic evolution of the support itself, and provide technical support for the unmanned operation of the fully mechanized mining faces.