MUSE: a multilevel symbolic encoding algorithm for state assignment

A novel state assignment algorithm, called MUSE (multilevel symbolic encoding), for the encoding of FSMs (finite state machines) targeted for multilevel implementation is presented. Novel methods are discussed for the computation of state pair costs that are based onmultilevel algebraic structures derived from the one hot encoded state machine by purely algebraic techniques. Both Boolean (distance-1 MERGE and consensus) and algebraic (SUB-EXPRESSION EXTRACTION and CO-KERNEL EXTRACTION) operations are used to calculate the encoding affinity of state pairs, and account for face embedding constraints as well. Both heuristic and simulated annealing encoding techniques are used. >