基于ARTSOC的预警误差分析方法及应用

Early Warning Error Analysis Based on ARTSOC and Application

日本的地震预警水平比较高,但是仍然存在较大的误差.为了分析误差的大小及特征,通过引入人工社会(artificial societies,ARTISOC)智能技术,开发模拟了一种在各种地震发生条件下的误差分析方法,在既定观测条件下,针对各种发生状态,分析误差的变动范围及规律.利用日本地震观测网络的实际数据,计算出目前预警方法下其误差的变动范围及特征.该方法给出了误差分析的一种途径,为改善地震观测网络的布局和修正预警方法等措施提供了帮助.

Although Japan is good in earthquake early warning （EEW）, the warnmg error still needs to be reduced. A method is proposed based on simulation results using the multi agent technology--artificial societies （ARTISOC）. Under the given conditions, the error characteristics can be examined according to the possibility of different cases in the simulation. The method is tested with actual data from Japan. It can be used to improve the layout of the earthquake observation network.