Logic Reduction/Factorization Berkeley’s Espresso (Cont.)
3) At this point, the solution is usually pretty good, but under certain conditions it can still be improved. There might be another cover with fewer terms or the same number of terms but fewer literals. To try to find a better cover, Espresso first shrinks the prime implicants to the smallest size that still covers the logic function. This process is called reduce.
4) Since reduction yields a cover that is typically no longer prime, the sequence of steps reduce, expand, and irredundant cover are repeated in such a fashion that alternative prime implicants are derived. Espresso will continue repeating these steps as long as it generates a cover that improves on the last solution found.