In this paper,the author extends directly adaptive multicut aggregation method of Svyatoslav Trukhanov,Lewis Ntaimo and Andrew Schaefer to solving two stage problems of stochastic convex programming with recourse.

A strong deviation theorem for arbitrary discrete random sequence;

The randomness of the number sequence was ensured by generated from binary system random sequence.

Random sequence was often used in empirical study of sort as input,but two correlative problems,that whether the random sequence s generation method and number of input groups were suitable,were generally not attached importance or ignored.

A New Method for Correlated Random Series Simulation;

Sample relative entropy and some small deviation theorems for dependent stochastic sequence;

The paper studys the strong convergence theorems for the arbitrary stochastic sequences by the convergence theorem for martingale-difference sequences.

A convergence of the section sum in stochastic sequence under finite state is produced with the help of same distribute and comparative function of stochastic sequenct.