GAP: A genetic algorithm approach to optimize two-bit decoder PLAs

PLAs with two-bit decoders at the inputs require a smaller area compared with standard two-level PLAs (Sasao 1984). The number of product rows required for such PLAs is a function of the assignment of pairs of variables to the decoders. This paper describes a minimization procedure that uses a genetic algorithm approach to reduce the size of two-bit decoder PLAs. Results are compared with those obtained by other approaches such as the Tomczuk and Miller heuristic (TMA) (1992) and the simulated annealing technique (Abd-El-Barr and Choy 1993). For large randomly generated test cases and bench-marks, our results are optimal or very near optimal.