The LEM3 System for Multitype Evolutionary Optimization

Abstract

LEM3 is the newest version of the learnable evolution model (LEM), a non-Darwinian evolutionary computation methodology that employs machine learning to guide evolutionary processes. Due to the deep integration of different modes of operation, several novel elements in its algorithm, and the use of the advanced machine learning system AQ21, the LEM3 system is a highly efficient and effective implementation of the methodology. LEM3 is particularly attractive for multitype optimization because it supports, and treats accordingly, different attribute types for describing candidate solutions in the population. These attribute types are nominal, ordinal, structured, cyclic, interval, and ratio. Application to optimization of parameters of a complex system illustrates multitype optimization problem.