Discrete optimization
In this paper, we consider the generalized prize-collecting Steiner forest problem, extending the prize-collecting Steiner forest problem. In this problem, we are given a connected graph G = (V, E) and a set of vertex sets V = {V1, V2, · · · , Vl }. Every edge in E has a nonnegative cost, and every vertex set in V has a nonnegative penalty cost. For a given edge set F ⊆ E, vertex set Vi ∈ V is said to be connected by edge set F if Vi is in a connected component of the F-spanned subgraph. The objective is to find such an edge set F such that the total edge cost in F and the penalty cost of the vertex sets not connected by F is minimized. Our main contribution is to give a 3-approximation algorithm for this problem via the primal-dual method.
This paper deals with a general variant of the reverse undesirable (obnoxious) center location problem on cycle graphs. Given a ‘selective’ subset of the vertices of the underlying cycle graph as location of the existing customers, the task is to modify the edge lengths within a given budget such that the minimum of distances between a predetermined undesirable facility location and the customer points is maximized under the perturbed edge lengths. We develop a combinatorial O(n log n) algorithm for the problem with continuous modifications. For the uniform-cost model, we solve this problem in linear time by an improved algorithm. Furthermore, exact solution methods are proposed for the problem with integer modifications.
This paper addresses a geometric programming problem, where the objective function and constraints are interval-valued functions. The concept of acceptable feasible region is introduced, and a methodology is developed to transform this model to a general optimization problem, which is free from interval uncertainty. Relationship between the solution of the original problem and the transformed problem is established. The methodology is illustrated through numerical examples. Solutions by the proposed method and previous methods are analyzed.