A Metaheuristic Approach to Two Dimensional Recursive Digital Filter Design - Advances in Heuristic Signal Processing and Applications - page 165

Chapter 8
A Metaheuristic Approach to Two Dimensional
Recursive Digital Filter Design

Abhronil Sengupta, Tathagata Chakraborti, and Amit Konar

Abstract The two dimensional IIR digital filter design problem has received in-
creased attention over the past few years. Recently, several metaheuristic algorithms
have been employed in this domain and have produced promising results. Invasive
Weed Optimization is one of the latest population-based metaheuristic algorithms
that mimics the colonizing action of weeds. In this chapter, an improvement to the
classical weed optimization algorithm has been proposed by introducing a constric-
tion factor in the seed dispersal phase. Temporal Difference Q-Learning has been
employed to adapt this parameter for different population members through the
successive generations. Such hybridization falls under a special class of adaptive
Memetic Algorithms. The proposed memetic realization, called Intelligent Invasive
Weed Optimization (IIWO), has been applied to the two-dimensional recursive digi-
tal filter design problem and it has outperformed several competitive algorithms that
have been applied in this research field in the past.

8.1 Introduction

Digital filters can be classified into two broad categories, namely, finite impulse
response (FIR) filters and infinite impulse response (IIR) filters. FIR filters are easier
to implement as they are non-recursive and are always stable. However, IIR filters
are much more efficient in comparison to FIR filters as they are capable of producing
very sharp and accurate frequency responses [ 4 ]. Since the design of IIR filters is
more challenging than its FIR counterpart, we have selected 2D IIR filter design as
the fundamental problem of this chapter.

B

A. Sengupta ( A. Konar

Dept. of Electronics and Telecommunication Eng., Jadavpur University, Kolkata, India
e-mail: [e-mail]
T. Chakraborti
e-mail: [e-mail]
A. Konar
e-mail: [e-mail]

) · T. Chakraborti ·