基于改进径向基函数神经网络算法的后勤保障资源消耗预测技术

Consumption Prediction Technologyabout Logistics Support Resources Based on Improved RBF Neural Network

文章选取陆军部队战时保障资源消耗中的油料消耗情况作为典型研究对象,同步考虑战时兵力规模、作战强度、地理环境、行动样式等影响因素,在广义回归神经网络模型良好的容错性和快速学习能力的基础上,独创性地使用基于共轭梯度下降法改进径向基函数(RBF)神经网络算法模型,较传统神经网络运行效率有了较大的提升,在保证油料消耗预测精度的同时,能有效提高预测的实时性与可靠性。

The fuel consumption in the army’s wartime support resources consumption is selected as the typical research object, and the influence factors such as wartime force scale, combat intensity, geographic environment, and action style are simultaneously considered. The generalized regression neural network model has good fault tolerance and fast learning. Based on the ability, the original use of the conjugate gradient descent method to improve the radial basis function(RBF) neural network algorithm model, while ensuring the accuracy of fuel consumption prediction, has a greater improvement in the operating efficiency of the traditional neural network. It can effectively improve the real-time and reliability of prediction.