MPRM expressions minimization based on simulated annealing genetic algorithm

摘要

A new simulated annealing genetic algorithm (SAGA) is presented to solve a NP-hard combinatorial optimization problem called Mixed-Polarity Reed-Muller (MPRM) expressions minimization. Since genetic algorithm (GA) is easily trapped to a local optimum and simulated annealing algorithm (SAA) has the limitation of poor convergence, a method incorporating annealing process into genetic operations is presented so that SAGA can converge to the global optimal solutions rapidly. Our experiments over several large scale benchmark circuits show that SAGA obtains much better performance compared to SAA and GA applied alone.