A novel accelerated artificial bee colony algorithm for optimal design of two dimensional FIR filter

This paper presents a novel approach for the design of two-dimensional (2D) Finite Impulse Response (FIR) filters. The design of FIR filters is generally non-differentiable, multimodal and higher dimensional; especially for 2D filters. A large number of filter coefficients are optimized either using constrained or unconstrained optimization approach. Due to the large number of constraints, traditional design methods cannot produce optimal filters required for some crucial applications. This makes meta-heuristic algorithms as good alternatives for addressing such constraints more efficiently. In order to improve the performance of 2D filters, we propose an Accelerated Artificial Bee Colony algorithm, termed as AABC algorithm. The earlier reported ABC based methods perform the modification of a single parameter of the solution in each cycle. But in this proposed AABC algorithm, we have adopted multiple parameters change of search equation at each step. This in turn improves the convergence speed of the algorithm by three times than the classical ABC algorithm and two times with respect to recently developed CABC method. In order to achieve better exploration behaviour of abandoned bees, we have also introduced a change during the initialization strategy of scout bees in the proposed AABC algorithm. The efficiency and robustness of the proposed algorithm are demonstrated by comparing its performance with classical Genetic Algorithm (GA), Particle Swarm Optimization , ABC and CABC methods.