The Estimator Of Rare Event From The Perspective Of Heavy-tailed Distribution In SRAM

At the present day and age, static random access memory (SRAM) unit is widelyused in the manufacture of a large number of semiconductor devices. People’s need fora stable SARM unit becomes higher and higher. At the same time, people hope toestimate its validity accurately and efficiently. But because of its small failure rate, it isdifficult to estimate its validity. Therefore, the present thesis put forward the simulationof rare event with a view to giving valid estimators so as to solve the problem.The current thesis first introduces the related background and incoming problemsof this issue. Also, the extensive application of the simulation of rare event is illustratedin this thesis. Where after, the thesis gives the estimators by different simulationmethods under the given assumption (rare event conforms to the heavy-taileddistribution and it belongs to Dominatedly varying and Subexponential). Meanwhile,the author here makes a comparison of the efficiency with the purpose of finding out avalid and efficient simulation method. The present thesis ends up with a simulatedexperiment which is carried out with SAS, and the results of the experiments verify theproposed theory.