Finite State Machine State Assignment for Multi-objective Optimization

Relay-race algorithm has been shown to be an effective method to achieve fast convergence.By utilizing the concept of the algorithm,the paper proposes rough search,focusing search,guide operator and relay operator to obtain good finite state machine state assignment with less power and area consumption in an efficient way.Mirror rotation and non-neighboring swap methods are proposed in the rough search and neighboring swap method is proposed in the focusing search to gain an effective convergence and escape local optimum.Guide operator uses statistical calculation to direct the optimization.By combining the best solution and current solution,relay operator is designed to achieve convergence in an effective way.In comparison with published results,the proposed algorithm can save more power dissipation and has advantage on area improvement but with less CPU time.