Optimality conditions for invex nonsmooth optimization problems with fuzzy objective functions (Q6102829)

The author considers the following optimization problem with fuzzy-valued objective function: \[ \tilde{f}(x) \rightarrow \min \text{ s. t. }g_j(x) \leq 0, \, j\in J,\, h_i(0) = 0,\, i \in I, \] where \(I,J\) are finite index sets and \(g_j:\mathbb{R}^n \rightarrow \mathbb{R}\), \(h_i(x): \mathbb{R}^n \rightarrow \mathbb{R}\) are non-smooth locally Lipschitz functions. The fuzzy-valued objective function is described by \(\alpha\)-cuts defined as follows: \[ \tilde{f}^{\alpha} = [\min_x f^L(x,\alpha), \min_x f^R(x,\alpha) ]. \] Clarke generalized derivative and Clarke generalized gradient are defined and used to introduce the concept of invexity. The representation the objective function by \(\alpha\)-cuts leads to introduction an associated bi-objective problem and establishing a connection between Pareto optimal solution of the bi-objective problem and nondominated solutions of the originally considered fuzzy-objective problem. Besides, the bi-objective problem has been used to formulate the Karush-Kuhn-Tucker optimality conditions. Some open problems connected with the presented research are outlined in the concluding part of the paper.