A tentative research on complexity of Automatic Programming
Kang Li-shan, Yan Li, Chen Yu-ping · Wuhan University Journal of Natural Sciences · 2001
In this paper, based on the following theoretical framework: Evolutionary Algorithms + Program Structures = Automatic Programming, some results on complexity of automatic programming for function modeling is given, which show that the complexity of automatic programming is an exponential function of the problem dimension N , the size of operator set | F | and the height of the program parse tree H . Following this results, the difficulties of automatic programming are discussed. Some function models discovered automatically from database by evolutionary modeling method are given too.