Gravitational search algorithm based design of optimal wideband IIR digital differentiator

This paper aims to design IIR digital differentiator by adopting a new heuristic algorithm called a gravitational search algorithm (GSA). In GSA, agents are treated as the masses and the force of attraction between agents is calculated based on Newton´s gravitation law to find optimal solutions to the problem. The agent with greater mass is treated as the near optimal solution and their position is a reflection of the solution. Simulations to find optimal filter coefficients of second, third and fourth order differentiators using GSA have been performed. Total absolute magnitude error and maximum phase error are the performance measures that are used to evaluate performance of the proposed differentiators. The result of the proposed GSA based approach has been compared with the standard existing algorithms like PSO, GA, SA, pole zero (PZ) and segment rule. Simulations show that GSA gives superior results when compared with the optimization abilities of other standard algorithms.