ABSTRACT
The cost function that estimates the optimality of an encoding depends on the target implementation: two-level or multilevel. Two-level implementations opti mize the number of product terms or the area of a programmable logic array (PLA). Multilevel implementations optimize the number of literals of a technolo gy-independent representation of the logic. Optimality may be based on more complex cost functions that take into account other criteria, like testability. It may even be the case that area becomes a secondary optimization objective as in the case of state assignment for asynchronous sequential circuits, where the main concern is the correctness of the behavior of the encoded circuit.
