Genetic annealing search for index assignment in vector quantization

Abstract

A Genetic Annealing Algorithm (GAA) has been applied to an Index Assignment (IA) problem in Vector Quantization (VQ). The proposed combination of annealing, selection and recombination mechanisms is new and possesses an advantage with regard to an implementation on a single processor machine for the reason of satisfactory performance when using a reasonably small population. The assessment of the proposed method by using a vector quantization of Line Spectrum Frequencies (LSF) of a speech coder, the experiments with the vector quantizers trained on the Gaussian data as well as the tests on the Quadratic Assignment Problem (QAP) benchmarks are presented.