An efficient method for decomposition of multiple-output Boolean functions and assigned sequential machines

Artificial Intelligence in Logic Design

“…The parallel search approaches that work with populations of solutions and consider a number of solutions in parallel (e.g. genetic algorithms (Beasley et al 1993a, 1993bDavis 1991Goldberg 1975;Holland 1992Holland , 1995JOZwiak et al 1998;Jozwiak and Postula 2002;Louis and Rawlins 1992) or double-beam search (Jozwiak and Kolsteren 1991;Jozwiak 1992a;Jozwiak and Volf 1992;Jozwiak 1992b; Vol 1995», have a good chance of performing this task adequately. They not only enable modification of solutions, but also interplay among various solutions, by combination of their parts, their comparison and selection.…”

“…Typically, sub-problems in digital circuit and system design have very large and "irregular" discrete solution spaces. They can be modelled in the form of multi-objective constrained optimisation problems (MOCOP) (Jozwiak and Kolsteren 1991;Jozwiak and Volf 1992;Jozwiak and Volf 1995). A MOCOP can be characterized by four sets of components:…”

“…The double-beam search scheme was implemented in numerous circuit synthesis methods and tools solving different kinds of synthesis problems, and when tested on benchmarks, it efficiently produced very good results for all considered applications (Jozwiak and Kolsteren 1991;Jozwiak 1992a;Jozwiak and Volf 1992;Jozwiak 1992b;Jozwiak and Volf 1995;Jozwiak and Chojnacki 2001).…”

“…With such a generic partitioning-based solution model its corresponding MOCOP model can be associated, by defining a set of some problem-specific objectives, constraints, and trade-off information. So modelled decomposition problems that involve hard constraints can be solved by special multi-dimensional and multi-objective packing approaches (Jozwiak and Kolsteren 1991;Jozwiak and Volf 1992;Jozwiak and Volf 1995). Problems without hard constraints can be solved by special multi-objective clustering (Jozwiak 1992;).…”

“…So modelled decomposition problems that involve hard constraints can be solved by special multi-dimensional and multi-objective packing approaches (Jozwiak and Kolsteren 1991;Jozwiak and Volf 1992;Jozwiak and Volf 1995). Problems without hard constraints can be solved by special multi-objective clustering (Jozwiak 1992;). The partitioning process, which produces partitions on the set of atomic computations is preceded by analysis of characteristic features of an original function/machine related to the characteristics of building blocks.…”