Efficient sum-to-one subsets algorithm for logic optimization

Abstract

An optimization algorithm, RENO, was proposed by K.C. Chen et al. (1991), in which a given network was minimized by optimally resynthesizing each gate in the network. It is shown that the resynthesis problem in RENO can be transformed into a minimum-cost sum-to-one subset problem based on a given cost function, which is an important problem that often occurs in logic optimization algorithms. Efficient procedures for solving both sum-to-one subsets and minimum-cost sum-to-one subset problems are presented and applied to multilevel network optimization algorithms. Both the efficiency and quality of these algorithms are greatly improved. The application of these techniques to multinode minimization using Boolean relations is also discussed.<>