| Autor: |
Santos, Dorabella, Martins, Lúcia, Gomes, Teresa |
| Zdroj: |
Annals of Telecommunications; Feb2023, Vol. 78 Issue 1/2, p1-12, 12p |
| Abstrakt: |
In this paper, we revisit a biobjective optimization problem arising in software-defined networking with availability guarantees. The problem jointly optimizes the number of controllers and their locations, as well as optimizes a set of links that can be upgraded to achieve intercontroller availability targets. This problem is known to be NP-complete, includes generalized nonlinear flow constraints and includes nonlinear and nonconvex path pair availability constraints. In this paper, we propose an iterative approach for solving the optimization problem. The approach consists in a linear approximation that is obtained through convex relaxation. The linear approximated problem is solved iteratively, and can be tailored to be as close to the original problem as possible at the expense of a greater computational effort. The results obtained using this approach are compared with previously obtained ones. We show how the iterative approach is able to significantly improve most of the tested instances (or find the same solution as previously obtained) within feasible computational effort. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
Full text from SpringerLink