用神经网络分辨"膝"区原初质子成分  

IDENTIFICATION OF PRIMARY PROTON COMPONENT AT THE KNEE WITH ARTIFICIAL NEURAL NETWORKS

作　　者：冯存峰[1] 孔繁敏[1] 张学尧[1] 何瑁[1] 戴志强[1] 张乃健[1] 

机构地区：[1]山东大学物理系高能物理研究室,济南250100

出　　处：《计算物理》1999年第6期568-572,共5页Chinese Journal of Computational Physics

基　　金：山东省自然科学基金;国家自然科学基金

摘　　要：利用不同的相互作用模型，对羊八井空气簇射阵列和乳胶室联合实验进行了Ｍｏｎｔｅ Ｃａｒｌｏ模拟，并利用模拟数据研究了利用人工神经网络方法进行原初宇宙线成分分辨的可行性。分析表明，利用该方法可很好地挑选出由原初质子引起的簇射，但所得结果与作用模型有关。A Monte Carlo simulation for the hybrid experiment of air shower array and emulsion chambers at Yangbajing is done with different interaction models. The feasibility of distinguishing primary cosmic ray components by use of the method of artificial neural networks is studied with the simulation data. The analysis indicates that the showers induced by primary protons can be efficiently selected by using this method, but the results obtained appear to be dependent on interaction model. The corresponding systematic error is also estimated.

关 键 词：神经网络 原初质子分辨 宇宙线能谱 膝结构 

分 类 号：O572.12[理学—粒子物理与原子核物理]