A lattice based evolution scheme with applications to design optimization

Design optimization is a multi-modal and complex problem which can not be treated successfully in scalar optimizer, such as evolutionary techniques. For typical genetic algorithms, most generalizations and variations do not spread to vector valued fitness functions. In this paper, we first propose a lattice based evolution technique with vector valued fitness function. Then we present a evolutionary searching paradigm for the design optimization problem. We present a simple Markov analysis of the proposed genetic algorithms.