1: \begin{abstract}

2: We consider a two-stage stochastic optimization problem, in which

3: a long-term optimization variable is coupled with a set of short-term

4: optimization variables in both objective and constraint functions.

5: Despite that two-stage stochastic optimization plays a critical role

6: in various engineering and scientific applications, there still lack

7: efficient algorithms, especially when the long-term and short-term

8: variables are coupled in the constraints. To overcome the challenge

9: caused by tightly coupled stochastic constraints, we first establish

10: a two-stage primal-dual decomposition (PDD) method to decompose the

11: two-stage problem into a long-term problem and a family of short-term

12: subproblems. Then we propose a PDD-based stochastic successive convex

13: approximation (PDD-SSCA) algorithmic framework to find KKT solutions

14: for two-stage stochastic optimization problems. At each iteration,

15: PDD-SSCA first runs a short-term sub-algorithm to find stationary

16: points of the short-term subproblems associated with a mini-batch

17: of the state samples. Then it constructs a convex surrogate for the

18: long-term problem based on the deep unrolling of the short-term sub-algorithm

19: and the back propagation method. Finally, the optimal solution of

20: the convex surrogate problem is solved to generate the next iterate.

21: We establish the almost sure convergence of PDD-SSCA and customize

22: the algorithmic framework to solve two important application problems.

23: Simulations show that PDD-SSCA can achieve superior performance over

24: existing solutions. 

25: \end{abstract}

26: