EngineeringSystems BiologyProcess ControlControl SystemsControl TheoryCrossover FormationCrossover OperatorsFeedback ControlOptimal Control
This paper introduces a simple, general, representation neutral method to implement endogenous crossover control in genetic algorithms. A series of experiments are offered to motivate the various components of this mechanism and to demonstrate its effectiveness. Convergence is presented as an endemic problem with Holland's prototypical GA and many derivative algorithms. Simple modifications to exogenous fitness functions are proposed and demonstrated to be effective solutions. 1 Introduction Whence sprang life? How was the molecular machinery that accomplishes replication, transcription and expression of DNA originally constructed? Although numerous authors have addressed the mystery of the genesis of the replicator [2, 1] it remains unresolved. This paper will not resolve it either, but may offer some insight into this abstruse question by investigating related, but much simpler questions. This paper is a small beginning towards a thorough investigation of endogenous control of genet...