Cultural algorithm with double populations for FIR digital filter design

A new cultural algorithm (CA) with double populations is proposed for designing Finite Impulse Response (FIR) digital filters. Two populations evolve independently according to different evolutionary schemes. One population employs variable inertia weight particle swarm optimization (PSO) algorithm, the other uses PSO with constriction factor. Belief space of CA plays the role of knowledge link in mutual cooperation and promotion between populations. This algorithm provides a new way for the co-evolution of multi-population. The computer simulations of FIR filter design indicate that the proposed algorithm is practicable and superior in terms of convergence speed and optimization effect compared with other algorithms.