State encoding for low power in FSM using non-oscillating self-adaptive particle swarm optimization (NOS-SAPSO)
D. Godwin Raj, A. B. Kalpana, Manoj Kumar Singh · Journal of Information and Optimization Sciences · 2020
In this paper, states encoding for the low power design in the Finite State Machine has been presented. Power reduction is considered in the dynamic portion of power consumption where switching activity between states decides the majority factor of power dissipation. Switching activity can be minimized up to a great extent with the optimal assignment of state encoding. Hence, a non-oscillatory variant of discrete self-adaptive particle swarm optimization has proposed to obtain the optimal encoding of states. The proposed method provides the diversity-based self-adaptive parametrization of inertia weight, cognition and social constant along with the best solution evolved ever to direct the progress for the population. The proposed method has delivered the non-oscillatory faster convergence with a high rate of success in delivering a global solution. It is found that the proposed method has surpassed various variants of the existing form of particle swarm optimization over the benchmark problem of the finite state machine.