Constrained optimization with normed moving truncations penalty-functions (Q1134630)

The normed moving truncations penalty-function method differs from the common exterior moving truncations penalty function method in that the problem functions are normalized after every completed unconstrained optimization, by multiplying with the reciprocal values attained by the functions. Computational results on 12 problems of varying complexity are reported, and compared with the results achieved by two other penalty-function methods.