A Novel Diversity-Controlled Genetic Algorithm for Optimization of BIBO Stable Digital IF Filters Over CSD Multiplier Coefficient Space

This paper presents a novel diversity-controlled (DC) genetic algorithm (GA) for the optimization of digital Intermediate Frequency (IF) filters over the (finiteprecision) canonical signed-digit (CSD) multiplier coefficient space. This optimization exploits the bilinear-losslessdiscrete-integrator (bilinear-LDI) lattice digital filter design approach for the realization of the required infinite-precision seed digital IF filter chromosome. A look-up table (LUT) approach is proposed to ensure that the finite-precision CSD digital IF filter chromosomes generated in the course of DCGA optimization are guaranteed to be bounded-input bounded-output (BIBO) stable. The salient feature of DCGA optimization is that it permits external control over the population diversity (i.e. the parent selection pressure) to achieve a high convergence speed. This feature is illustrated through the application of the proposed DCGA optimization to the design of a pair of practical digital IF filters satisfying different design specifications. It is observed that, for both digital IF filter designs, the DCGA optimization results in around an order of magnitude improvement in the convergence speed as compared to a conventional GA optimization.