失效概率函数求解的高效算法

结合基于分数矩约束的极大熵方法和替代模型法,发展了一种失效概率函数求解的高效算法。所提算法的基本思路是利用自主学习的迭代Kriging方法来构造失效概率函数,即采用较少的训练样本来构造粗糙的失效概率函数,在此基础上通过添加新的违反学习函数约束的样本来更新失效概率函数,直到达到精度要求。对于每一个分布参数的训练样本点,所提方法采用分数矩约束的极大熵法来求解相应的失效概率样本。由于分数矩的计算采用了高效的降维积分,并且由于分数矩约束下极大熵法中优化策略高效地逼近了响应的概率密度函数,从而使得失效概率样本能够被高效高精度地估计出来。为了检验所提方法的精度及效率,给出了两个算例,对比了所提方法与已有的失效概率函数求解的Bayes公式法及Monte Carlo法等,结果表明,所提方法适用于求解复杂的功能函数问题,且在满足精度要求的基础上大大降低了计算量。     

Modern liberal wars,illiberal allies,and peace as the failure of policy

ABSTRACT

The post-Cold War period nearly up to the present has been characterised as the age of liberal wars, yet key facets of the liberal guidance of war remain under appreciated. This article seeks to address this wider gap with regard to the particular concern of war termination and the fulfilment or failure of policy. First, it develops characterisations of liberal wars based on the existing literature, identifying three broad types through consideration of context—defensive versus offensive—and of political and strategic agency, particularly regarding the motives for and intents of action. Three types of liberal wars result: defensive liberal wars, offensive liberal wars with humanitarian motive and geopolitical intent, and offensive liberal wars with geopolitical motive and humanitarian intent. The article then presents one exemplary case for each liberal war with an emphasis on how liberal strategy required an illiberal ally and that ally's effect on the subsequent peace.