Abstract:In this paper, we study a robust optimization problem whose constraints include nonsmooth and nonconvex functions and the intersection of closed sets. Using advanced variational analysis tools, we first provide necessary conditions for the optimality of the robust optimization problem. We then establish sufficient conditions for the optimality of the considered problem under the assumption of generalized convexity. In addition, we present a dual problem to the primal robust optimization problem and examine duality relations.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the optimality and duality of robust optimization problems with uncertain parameters and closed - set intersection constraints. Specifically, the research objectives are:
1. **Optimality Conditions**: Provide necessary conditions for the optimal solution of robust optimization problems that include non - smooth, non - convex functions and closed - set intersection constraints.
2. **Sufficient Conditions**: Under the assumption of generalized convexity, establish sufficient conditions for the optimal solution of this type of robust optimization problems.
3. **Dual Problem**: Propose a dual problem corresponding to the original robust optimization problem and study the dual relationship between the two.
### Specific Form of the Problem
Consider the following uncertain optimization problem (UP):
\[
\inf_{x \in \mathbb{R}^n} \left\{ f(x, \tau) \mid x \in \bigcap_{j = 1}^m C_j, g_i(x, u_i)\leq0, i = 1,\ldots, p \right\},
\]
where \(x\) is the decision variable, \(\tau\) and \(u_i\) (\(i = 1,\ldots, p\)) are uncertain parameters, belonging to the uncertainty sets \(T\subset\mathbb{R}^k\) and \(V_i\subset\mathbb{R}^{n_i}\) respectively, and these sets are non - empty compact sets. \(C_j\subset\mathbb{R}^n\) (\(j = 1,\ldots, m\)) are non - empty closed subsets, and \(f:\mathbb{R}^n\times T\rightarrow\mathbb{R}\) and \(g_i:\mathbb{R}^n\times V_i\rightarrow\mathbb{R}\) (\(i = 1,\ldots, p\)) are functions.
The corresponding robust optimization problem (RP) is defined as:
\[
\inf_{x \in \mathbb{R}^n} \left\{ \max_{\tau\in T} f(x, \tau) \mid x \in \bigcap_{j = 1}^m C_j, g_i(x, u_i)\leq0, \forall u_i\in V_i, i = 1,\ldots, p \right\}.
\]
### Significance of the Research
Due to the existence of prediction errors, fluctuations and uncertain data, many practical problems have uncertain data. Therefore, robust optimization, as an effective framework, is used to study mathematical programming problems under data uncertainty. This paper aims to provide a theoretical basis and solutions for such complex problems by introducing advanced variational analysis tools.