A hybrid approach for MILP partitioning problem
R. Farfar Boudour, D. Kimour, T. Mohamed, Hichem Arioui, Rochdi Merzouki, Hadj Ahmed Abbassi · AIP conference proceedings · 2008
Splitting a large system into smaller and more manageable units has become an important problem and a challenging task for some fields of research. The question of generating a good partitioning into smaller modules becomes a minimization problem for the number of parts being called by other parts. Two different and complementary approaches turned out to be promising to tackle the problem. First, we used a new approach, known as extreme partitioning, where hw/sw partitioning is based on profiling. Nevertheless, this one didn't guarantee hard timing constraints. To overcome this weakness, we coupled it with a MILP partitioning approach and used to reach this solution strategy, by efficient reformulations and a clever implementation. The advantage of the first one is to reduce particularly the time to market and the second one is to verify timing constraints by scheduling and to keep the cost down by clustering the nodes. We illustrate this method on two examples. The results are encouraging.