Novel Monte Carlo weight window generator for global tallying problems based on maximization of global efficiency indicator

Weight window method is powerful in seeking precise global tallies or single target tally of particle transport problems within a reasonable time cost when its parameters are suitable. But how to get excellent parameters is not trivial, especially for complex transport models. Various strategies are proposed to set numerous parameters of this method based on various considerations. But this challenging problem is still open for new exploration despite all the progress which have been achieved. This paper aims to solve the global tallying problem, which is more difficult in the vast majority of all cases, by a novel weight window generator which supplies excellent parameters of weight window method. This generator relies on the maximization of global efficiency indicator and a natural adaptive strategy. The efficiency and reliability of this generator are validated on the Winfrith Water benchmark and AP1000 full-core model, respectively. Compared to the classical MAGIC method, this proposed method can enhance global efficiency by 2-3 times magnitude when measured by the same global efficiency indicator.