A method and system for optimizing state assignments for a finite state machine. The method generates a random initial state assignment for each of a plurality of states of the finite state machine, determines an initial cost associated with the random initial state assignments, identifies a code swap to explore as a function of a code swap probability. Further, the method calculates a cost for the code swap when one or more criteria is satisfied, updates the code swap probability as a function of the cost of the code swap and a best cost, performs the code swap when the cost of the swap is smaller than the best cost and/or a current cost to optimize the state assignments, and outputs optimized state assignments.