Created by W.Langdon from gp-bibliography.bib Revision:1.7630
directed genetic programming (DGP), monte Carlo, 'straightforward GP...leads to poor results' p8451 DGP adds form of solution? Fset={+,-<*,/} pop=100 G<=250. Ekart well best(?) -1.5576eV, most within 0.5 percent. Even better with Bessel function. Also tried Gaussian. NB 'proper choice of the grid is important' p852. Asymptotic region dominates tunnelling. 'we believe that...more readily find the solution that has the simplest functional form' p8542 (ie the lowest energy eigenstate).
Excited states. Harmonic oscillator creation operator. problems with second excited state? Hartree approximation (p8544), separate x and y dimensions, use same bell curve for both x and y. x,y back together? Seeded run? Fset now also includes exp First excited state E=2.534.",
Genetic Programming entries for Dmitrii E Makarov Horia Metiu