Design of evolutionary algorithms-A statistical perspective

摘要: This paper describes a statistical method that helps to find good parameter settings for evolutionary algorithms. The builds functional relationship between the algorithm's performance and its values. relationship-a model-can be identified thanks simulation data. Estimation test procedures are used evaluate effect of variation. In addition, can investigated with reduced number experiments. Problem labeling also considered as model variable enables identifying classes problems which algorithm behaves similarly. Defining such increases quality estimations without increasing computational cost.