This paper discusses some novel strategies to tackle the difficulty problem of finding a good logical network topology with minimum traffic congestion. We apply our strategy on solving two typical topologies structures: the Mesh and the Self-Healing Rings. The typical strategies for solving the logical topology design for both structures are those based on the use of mixed-integer linear programming. However the literature shows that these approaches can be frustrating, time consuming and costly. As an alternative to these strategies, our approach combines the capability of meta-heuristics of finding good solutions in a very short computational time and provides the mixed-integer linear programming with good upperbounds in order to pruning great chunk of the searching space. In this work we show that our approach is promising as we are able to solve large problems in a reasonable amount of time for both type of topologies we studied.