Dear ALPS users and developers, I solved the problem which I sent to this mailing list on February 1st by myself, but I got another problem about Kondo lattice model. I am studying one dimensional Kondo lattice model with ALPS DMRG application, and in a small system(L=4),I compared the ground state energy which I got by ALPS DMRG with the one which I got by exact diagoanlization in the case t=0,J=1 and I fixed the number of conduction electrons 4(i.e. half-filled).Then I got E0=-1.6160254... by DMRG and E0 =-3.00000...by exact diagonalization. The latter seems to be reasonable, but I don't have any idea about where DMRG is wrong. I used ALPS lattice library and "2 band open chain lattice", and the model library to which I added a quantum number "N" and a constraint in order to fix the number of conduction electrons. I have attached the model library xml file to this email. And I used the following parameter file: 'LATTICE' : "2 band open chain lattice", 'MODEL_LIBRARY' :"kondo_lattice_ model.xml", 'MODEL' : "Kondo lattice ", 'CONSERVED_QUANTUMNUMBERS' :"N", 'N_total' :4, 'local_spin' :0.5, 't' :0, 'J' :1, 'SWEEPS' : 6, 'NUMBER_EIGENVALUES' : 2, 'L' : 4, 'MAXSTATES' : 150 Any advice will help me. Best regards, Riku Masui