Direct Method for Solving Bilinear Programming Problem

The bilinear programming problem is considered, where a column, which corresponds to one of the variables, is not fixed but can be chosen from a convex set. This problem is known as the Dantzig – Wolfe problem. Earlier, a modified support method was proposed to solve the problem, using the decomposition of the problem constraints of the Dantzig – Wolfe method. The author of the paper has developed a direct exact method for solving the formulated problem. The method is based on the idea of the solving a linear programming problem with generalized direct constraints and a general concept of an adaptive solution method. The notions of support, support plan, optimal and suboptimal (e-optimal) plan are introduced which is a given approximation of the objective function to the optimal plan of the problem. Criteria for optimality and suboptimality of the support plan have been formulated and have been proved in the paper. The search for the optimal solution is based on the idea of maximizing the increment of the objective function. This approach allows more fully to take into account the main target and structure of the problem. Improving a support plan consists of two parts: replacing the plan and replacing the support. To find a suitable direction, a special derived problem is solved while taking into account the main constraints of the problem. The replacement of the support is based on the search for the optimal plan of the dual problem. The method leads to an optimal solution to the problem in a finite number of iterations (in the case of a non-degenerate value).

Linearized Robust Counterparts of Two-Stage Robust Optimization Problems with Applications in Operations Management

In this article, we discuss an alternative method for deriving conservative approximation models for two-stage robust optimization problems. The method mainly relies on a linearization scheme employed in bilinear programming; therefore, we will say that it gives rise to the linearized robust counterpart models. We identify a close relation between this linearized robust counterpart model and the popular affinely adjustable robust counterpart model. We also describe methods of modifying both types of models to make these approximations less conservative. These methods are heavily inspired by the use of valid linear and conic inequalities in the linearization process for bilinear models. We finally demonstrate how to employ this new scheme in location-transportation and multi-item newsvendor problems to improve the numerical efficiency and performance guarantees of robust optimization.