An Improved Branch & Bound Method for the Uncapacitated Competitive Location Problem |
| |
Authors: | Stefano Benati |
| |
Affiliation: | (1) Dipartimento di Informatica e Studi Aziendali, Universitá di Trento, Via Inama 5, 38100 Trento, Italy |
| |
Abstract: | In this paper, the problem of locating new facilities in a competitive environment is considered. The problem is formulated as the firm expected profit maximization and a set of nodes is selected in a graph representing the geographical zone. Profit depends on fixed and deterministic location costs and, since customers are independent decision-makers, on the expected market share. The problem is an instance of nonlinear integer programming, because the objective function is concave and submodular. Due to this complexity a branch & bound method is developed for solving small size problems (that is, when the number of nodes is less than 50), while a heuristic is necessary for larger problems. The branch & bound is called data-correcting method, while the approximate solutions are obtained using the heuristic-concentration method. |
| |
Keywords: | competitive location models random utility theory submodular functions heuristic concentration data-correcting method |
本文献已被 SpringerLink 等数据库收录! |
|