A parallel hybrid approach to hard optimization problems

This paper presents a software platform for the distributed solution of complex optimization problems. For such problems there are often no outstanding approaches which always perform well. That is why we propose to use diierent solution methods simultaneously. They can communicate and thus prootate from each other. Some general paradigms are applicable to a wide class of problems. Our platform contains reusable implementations of such paradigms. A rst application to set covering problems is presented. Experimental results for this application demonstrate the quality of such an approach.