ABSTRACT
The objective of this paper is to present a novel MA for VQ design. In the algorithm, instead of using the generational GA for global search, a
1 INTRODUCTION
Genetic Algorithm (GA) (A.E. Eiben, & J.D. Smith, 2003, T.K. Lin et al. 2008, M. Srinivas & L.M. Patnaik, 1994) is a general-purpose search algorithm for solving optimisation problems by simulating natural evolution over populations of candidate solutions. Inspired by biological evolution, the GA consists of a set of genetic strings, which are evaluated by a fitness function. The fittest strings are then regenerated at the expense of the others. Moreover, crossover and mutation are employed to obtained better strings. The mutation operator changes individual elements of a string, and the crossover operation interchanges parts between strings. In the generational GA, the combination of these operations is called a generation. The evolution of genetic strings may be continued for several generations to attain a near global optimal solution.
