| |||||||||

Solution support sets Description Nonlinear example demonstrating the type of solutions returned by solvers Further explanation of this example:
'Xpress NonLinear Reference Manual'
Source Files xnlp_supportset.mos ! XNLP example demonstrating the type of solutions returned by solvers ! ! This examples solves simple optimzation problems with different solvers, ! demonstrating the fundamental properites of the solutions returned. ! ! This example demonstrates a particular non-linear optimization concept as related ! to Xpress NonLinear. ! The version of the example is for Xpress 7.5. ! ! (c) 2013 Fair Isaac Corporation ! author: Zsolt Csizmadia model mmxnlp_nlp_duals uses "mmxnlp"; ! Scaling parameter for problem size. Please note, there is a quadratic response ! in terms of complexity parameters N = 10 end-parameters declarations R = 1..N x : array(R) of mpvar end-declarations !setparam("xnlp_verbose",true) sum( i in R) (x(i)) >= N ! Minimize using the simplex method minimize(XPRS_PRI, sum( i in R) (x(i))) writeln("A simplex solution to an LP:") forall(i in R) write(getsol(x(i)), " ") writeln ! And compare to a barrier solution without crossover setparam("XPRS_CROSSOVER",0) setparam("XPRS_PRESOLVE",0) minimize(XPRS_BAR, sum( i in R) (x(i))) writeln("The same with the barrier:") forall(i in R) write(getsol(x(i)), " ") writeln writeln writeln ! The very same phenomena with quadratic problems: writeln("The same experiment over a convex quadratic optimization problem:") writeln ! Minimize using the simplex method minimize(XPRS_PRI, sum( i in 1..ceil(N/2)) (x(i)) + sum( i in ceil(N/2)..N) (x(i)^2) ) writeln("A simplex solution to an QP:") forall(i in R) write(getsol(x(i)), " ") writeln ! And compare to a barrier solution without crossover setparam("XPRS_CROSSOVER",0) setparam("XPRS_PRESOLVE",0) minimize(XPRS_BAR, sum( i in 1..ceil(N/2)) (x(i)) + sum( i in ceil(N/2)..N) (x(i)^2) ) writeln("The same with the barrier:") forall(i in R) write(getsol(x(i)), " ") writeln ! Notice that for the quadratic part, the quadratic objetcive itself has served ! as an equalizer end-model | |||||||||

Copyright 2017 Fair Isaac Corporation. |