Meta-optimization based on self-organizing map and genetic algorithm
Anatoly Karpenko, Z. O. Svianadze · Optical Memory and Neural Networks · 2011
The paper considers the parametric optimization of search optimization algorithms (metaoptimization). The meta-optimization method enables to find the best strategy for an algorithm during the execution of the program based on this algorithm. The method uses the clusterization of a set of problems of a particular class with the help of Kohenen self-organizing maps and tackles the metaoptimization problem proper with the aid of the continuous genetic algorithm.