Home > matpower5.1 > miqps_glpk.m

miqps_glpk

PURPOSE ^

MIQPS_GLPK Mixed Integer Linear Program Solver based on GLPK - GNU Linear Programming Kit.

SYNOPSIS ^

function [x, f, eflag, output, lambda] = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)

DESCRIPTION ^

MIQPS_GLPK  Mixed Integer Linear Program Solver based on GLPK - GNU Linear Programming Kit.
   [X, F, EXITFLAG, OUTPUT, LAMBDA] = ...
       MIQPS_GLPK(H, C, A, L, U, XMIN, XMAX, X0, OPT)
   A wrapper function providing a MATPOWER standardized interface for using
   GLKP to solve the following LP (linear programming) problem:

       min C'*X
        X

   subject to

       L <= A*X <= U       (linear constraints)
       XMIN <= X <= XMAX   (variable bounds)

   Inputs (all optional except H, C, A and L):
       H : dummy matrix (possibly sparse) of quadratic cost coefficients
           for QP problems, which GLPK does not handle
       C : vector of linear cost coefficients
       A, L, U : define the optional linear constraints. Default
           values for the elements of L and U are -Inf and Inf,
           respectively.
       XMIN, XMAX : optional lower and upper bounds on the
           X variables, defaults are -Inf and Inf, respectively.
       X0 : optional starting value of optimization vector X (NOT USED)
       VTYPE : character string of length NX (number of elements in X),
               or 1 (value applies to all variables in x),
               allowed values are 'C' (continuous), 'B' (binary) or
               'I' (integer).
       OPT : optional options structure with the following fields,
           all of which are also optional (default values shown in
           parentheses)
           verbose (0) - controls level of progress output displayed
               0 = no progress output
               1 = some progress output
               2 = verbose progress output
           skip_prices (0) - flag that specifies whether or not to
               skip the price computation stage, in which the problem
               is re-solved for only the continuous variables, with all
               others being constrained to their solved values
           glpk_opt - options struct for GLPK, value in
               verbose overrides these options
       PROBLEM : The inputs can alternatively be supplied in a single
           PROBLEM struct with fields corresponding to the input arguments
           described above: H, c, A, l, u, xmin, xmax, x0, vtype, opt

   Outputs:
       X : solution vector
       F : final objective function value
       EXITFLAG : exit flag, 1 - optimal, <= 0 - infeasible, unbounded or other
       OUTPUT : struct with errnum and status fields from GLPK output args
       LAMBDA : struct containing the Langrange and Kuhn-Tucker
           multipliers on the constraints, with fields:
           mu_l - lower (left-hand) limit on linear constraints
           mu_u - upper (right-hand) limit on linear constraints
           lower - lower bound on optimization variables
           upper - upper bound on optimization variables

   Note the calling syntax is almost identical to that of GLPK. The main
   difference is that the linear constraints are specified with A, L, U
   instead of A, B, Aeq, Beq.

   Calling syntax options:
       [x, f, exitflag, output, lambda] = ...
           miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)

       x = miqps_glpk(H, c, A, l, u)
       x = miqps_glpk(H, c, A, l, u, xmin, xmax)
       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0)
       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype)
       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)
       x = miqps_glpk(problem), where problem is a struct with fields:
                       H, c, A, l, u, xmin, xmax, x0, vtype, opt
                       all fields except 'c', 'A' and 'l' or 'u' are optional
       x = miqps_glpk(...)
       [x, f] = miqps_glpk(...)
       [x, f, exitflag] = miqps_glpk(...)
       [x, f, exitflag, output] = miqps_glpk(...)
       [x, f, exitflag, output, lambda] = miqps_glpk(...)


   Example: (problem from from http://www.jmu.edu/docs/sasdoc/sashtml/iml/chap8/sect12.htm)
       H = [   1003.1  4.3     6.3     5.9;
               4.3     2.2     2.1     3.9;
               6.3     2.1     3.5     4.8;
               5.9     3.9     4.8     10  ];
       c = zeros(4,1);
       A = [   1       1       1       1;
               0.17    0.11    0.10    0.18    ];
       l = [1; 0.10];
       u = [1; Inf];
       xmin = zeros(4,1);
       x0 = [1; 0; 0; 1];
       opt = struct('verbose', 2);
       [x, f, s, out, lambda] = miqps_glpk(H, c, A, l, u, xmin, [], x0, vtype, opt);

   See also GLPK.

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:

SOURCE CODE ^

0001 function [x, f, eflag, output, lambda] = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)
0002 %MIQPS_GLPK  Mixed Integer Linear Program Solver based on GLPK - GNU Linear Programming Kit.
0003 %   [X, F, EXITFLAG, OUTPUT, LAMBDA] = ...
0004 %       MIQPS_GLPK(H, C, A, L, U, XMIN, XMAX, X0, OPT)
0005 %   A wrapper function providing a MATPOWER standardized interface for using
0006 %   GLKP to solve the following LP (linear programming) problem:
0007 %
0008 %       min C'*X
0009 %        X
0010 %
0011 %   subject to
0012 %
0013 %       L <= A*X <= U       (linear constraints)
0014 %       XMIN <= X <= XMAX   (variable bounds)
0015 %
0016 %   Inputs (all optional except H, C, A and L):
0017 %       H : dummy matrix (possibly sparse) of quadratic cost coefficients
0018 %           for QP problems, which GLPK does not handle
0019 %       C : vector of linear cost coefficients
0020 %       A, L, U : define the optional linear constraints. Default
0021 %           values for the elements of L and U are -Inf and Inf,
0022 %           respectively.
0023 %       XMIN, XMAX : optional lower and upper bounds on the
0024 %           X variables, defaults are -Inf and Inf, respectively.
0025 %       X0 : optional starting value of optimization vector X (NOT USED)
0026 %       VTYPE : character string of length NX (number of elements in X),
0027 %               or 1 (value applies to all variables in x),
0028 %               allowed values are 'C' (continuous), 'B' (binary) or
0029 %               'I' (integer).
0030 %       OPT : optional options structure with the following fields,
0031 %           all of which are also optional (default values shown in
0032 %           parentheses)
0033 %           verbose (0) - controls level of progress output displayed
0034 %               0 = no progress output
0035 %               1 = some progress output
0036 %               2 = verbose progress output
0037 %           skip_prices (0) - flag that specifies whether or not to
0038 %               skip the price computation stage, in which the problem
0039 %               is re-solved for only the continuous variables, with all
0040 %               others being constrained to their solved values
0041 %           glpk_opt - options struct for GLPK, value in
0042 %               verbose overrides these options
0043 %       PROBLEM : The inputs can alternatively be supplied in a single
0044 %           PROBLEM struct with fields corresponding to the input arguments
0045 %           described above: H, c, A, l, u, xmin, xmax, x0, vtype, opt
0046 %
0047 %   Outputs:
0048 %       X : solution vector
0049 %       F : final objective function value
0050 %       EXITFLAG : exit flag, 1 - optimal, <= 0 - infeasible, unbounded or other
0051 %       OUTPUT : struct with errnum and status fields from GLPK output args
0052 %       LAMBDA : struct containing the Langrange and Kuhn-Tucker
0053 %           multipliers on the constraints, with fields:
0054 %           mu_l - lower (left-hand) limit on linear constraints
0055 %           mu_u - upper (right-hand) limit on linear constraints
0056 %           lower - lower bound on optimization variables
0057 %           upper - upper bound on optimization variables
0058 %
0059 %   Note the calling syntax is almost identical to that of GLPK. The main
0060 %   difference is that the linear constraints are specified with A, L, U
0061 %   instead of A, B, Aeq, Beq.
0062 %
0063 %   Calling syntax options:
0064 %       [x, f, exitflag, output, lambda] = ...
0065 %           miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)
0066 %
0067 %       x = miqps_glpk(H, c, A, l, u)
0068 %       x = miqps_glpk(H, c, A, l, u, xmin, xmax)
0069 %       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0)
0070 %       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype)
0071 %       x = miqps_glpk(H, c, A, l, u, xmin, xmax, x0, vtype, opt)
0072 %       x = miqps_glpk(problem), where problem is a struct with fields:
0073 %                       H, c, A, l, u, xmin, xmax, x0, vtype, opt
0074 %                       all fields except 'c', 'A' and 'l' or 'u' are optional
0075 %       x = miqps_glpk(...)
0076 %       [x, f] = miqps_glpk(...)
0077 %       [x, f, exitflag] = miqps_glpk(...)
0078 %       [x, f, exitflag, output] = miqps_glpk(...)
0079 %       [x, f, exitflag, output, lambda] = miqps_glpk(...)
0080 %
0081 %
0082 %   Example: (problem from from http://www.jmu.edu/docs/sasdoc/sashtml/iml/chap8/sect12.htm)
0083 %       H = [   1003.1  4.3     6.3     5.9;
0084 %               4.3     2.2     2.1     3.9;
0085 %               6.3     2.1     3.5     4.8;
0086 %               5.9     3.9     4.8     10  ];
0087 %       c = zeros(4,1);
0088 %       A = [   1       1       1       1;
0089 %               0.17    0.11    0.10    0.18    ];
0090 %       l = [1; 0.10];
0091 %       u = [1; Inf];
0092 %       xmin = zeros(4,1);
0093 %       x0 = [1; 0; 0; 1];
0094 %       opt = struct('verbose', 2);
0095 %       [x, f, s, out, lambda] = miqps_glpk(H, c, A, l, u, xmin, [], x0, vtype, opt);
0096 %
0097 %   See also GLPK.
0098 
0099 %   MATPOWER
0100 %   Copyright (c) 2010-2015 by Power System Engineering Research Center (PSERC)
0101 %   by Ray Zimmerman, PSERC Cornell
0102 %
0103 %   $Id: miqps_glpk.m 2661 2015-03-20 17:02:46Z ray $
0104 %
0105 %   This file is part of MATPOWER.
0106 %   Covered by the 3-clause BSD License (see LICENSE file for details).
0107 %   See http://www.pserc.cornell.edu/matpower/ for more info.
0108 
0109 %% check for Optimization Toolbox
0110 % if ~have_fcn('quadprog')
0111 %     error('miqps_glpk: requires the MEX interface to GLPK');
0112 % end
0113 
0114 %%----- input argument handling  -----
0115 %% gather inputs
0116 if nargin == 1 && isstruct(H)       %% problem struct
0117     p = H;
0118     if isfield(p, 'opt'),   opt = p.opt;    else,   opt = [];   end
0119     if isfield(p, 'vtype'), vtype = p.vtype;else,   vtype = []; end
0120     if isfield(p, 'x0'),    x0 = p.x0;      else,   x0 = [];    end
0121     if isfield(p, 'xmax'),  xmax = p.xmax;  else,   xmax = [];  end
0122     if isfield(p, 'xmin'),  xmin = p.xmin;  else,   xmin = [];  end
0123     if isfield(p, 'u'),     u = p.u;        else,   u = [];     end
0124     if isfield(p, 'l'),     l = p.l;        else,   l = [];     end
0125     if isfield(p, 'A'),     A = p.A;        else,   A = [];     end
0126     if isfield(p, 'c'),     c = p.c;        else,   c = [];     end
0127     if isfield(p, 'H'),     H = p.H;        else,   H = [];     end
0128 else                                %% individual args
0129     if nargin < 10
0130         opt = [];
0131         if nargin < 9
0132             vtype = [];
0133             if nargin < 8
0134                 x0 = [];
0135                 if nargin < 7
0136                     xmax = [];
0137                     if nargin < 6
0138                         xmin = [];
0139                     end
0140                 end
0141             end
0142         end
0143     end
0144 end
0145 
0146 %% define nx, set default values for missing optional inputs
0147 if isempty(H) || ~any(any(H))
0148     if isempty(A) && isempty(xmin) && isempty(xmax)
0149         error('miqps_glpk: LP problem must include constraints or variable bounds');
0150     else
0151         if ~isempty(A)
0152             nx = size(A, 2);
0153         elseif ~isempty(xmin)
0154             nx = length(xmin);
0155         else    % if ~isempty(xmax)
0156             nx = length(xmax);
0157         end
0158     end
0159 else
0160     error('miqps_glpk: GLPK handles only LP problems, not QP problems');
0161     nx = size(H, 1);
0162 end
0163 if isempty(c)
0164     c = zeros(nx, 1);
0165 end
0166 if isempty(A) || (~isempty(A) && (isempty(l) || all(l == -Inf)) && ...
0167                                  (isempty(u) || all(u == Inf)))
0168     A = sparse(0,nx);           %% no limits => no linear constraints
0169 end
0170 nA = size(A, 1);                %% number of original linear constraints
0171 if isempty(u)                   %% By default, linear inequalities are ...
0172     u = Inf(nA, 1);             %% ... unbounded above and ...
0173 end
0174 if isempty(l)
0175     l = -Inf(nA, 1);            %% ... unbounded below.
0176 end
0177 if isempty(xmin)                %% By default, optimization variables are ...
0178     xmin = -Inf(nx, 1);         %% ... unbounded below and ...
0179 end
0180 if isempty(xmax)
0181     xmax = Inf(nx, 1);          %% ... unbounded above.
0182 end
0183 if isempty(x0)
0184     x0 = zeros(nx, 1);
0185 end
0186 
0187 %% default options
0188 if ~isempty(opt) && isfield(opt, 'verbose') && ~isempty(opt.verbose)
0189     verbose = opt.verbose;
0190 else
0191     verbose = 0;
0192 end
0193 
0194 %% split up linear constraints
0195 ieq = find( abs(u-l) <= eps );          %% equality
0196 igt = find( u >=  1e10 & l > -1e10 );   %% greater than, unbounded above
0197 ilt = find( l <= -1e10 & u <  1e10 );   %% less than, unbounded below
0198 ibx = find( (abs(u-l) > eps) & (u < 1e10) & (l > -1e10) );
0199 AA = [ A(ieq, :); A(ilt, :); -A(igt, :); A(ibx, :); -A(ibx, :) ];
0200 bb = [ u(ieq);    u(ilt);    -l(igt);    u(ibx);    -l(ibx)];
0201 
0202 %% grab some dimensions
0203 nlt = length(ilt);      %% number of upper bounded linear inequalities
0204 ngt = length(igt);      %% number of lower bounded linear inequalities
0205 nbx = length(ibx);      %% number of doubly bounded linear inequalities
0206 neq = length(ieq);      %% number of equalities
0207 nie = nlt+ngt+2*nbx;    %% number of inequalities
0208 
0209 ctype = [repmat('S', neq, 1); repmat('U', nlt+ngt+2*nbx, 1)];
0210 
0211 if isempty(vtype) || isempty(find(vtype == 'B' | vtype == 'I'))
0212     mi = 0;
0213     vtype = repmat('C', nx, 1);
0214 else
0215     mi = 1;
0216     %% expand vtype to nx elements if necessary
0217     if length(vtype) == 1 && nx > 1
0218         vtype = char(vtype * ones(nx, 1));
0219     elseif size(vtype, 2) > 1   %% make sure it's a col vector
0220         vtype = vtype';
0221     end
0222 end
0223 
0224 %% set options struct for GLPK
0225 if ~isempty(opt) && isfield(opt, 'glpk_opt') && ~isempty(opt.glpk_opt)
0226     glpk_opt = glpk_options(opt.glpk_opt);
0227 else
0228     glpk_opt = glpk_options;
0229 end
0230 glpk_opt.msglev = verbose;
0231 
0232 %% call the solver
0233 [x, f, errnum, extra] = ...
0234     glpk(c, AA, bb, xmin, xmax, ctype, vtype, 1, glpk_opt);
0235 
0236 %% set exit flag
0237 if isfield(extra, 'status')             %% status found in extra.status
0238     output.errnum = errnum;
0239     output.status = extra.status;
0240     eflag = -errnum;
0241     if eflag == 0 && extra.status == 5
0242         eflag = 1;
0243     end
0244 else                                    %% status found in errnum
0245     output.errnum = [];
0246     output.status = errnum;
0247     if have_fcn('octave')
0248         if errnum == 180 || errnum == 151 || errnum == 171
0249             eflag = 1;
0250         else
0251             eflag = 0;
0252         end
0253     else
0254         if errnum == 5
0255             eflag = 1;
0256         else
0257             eflag = 0;
0258         end
0259     end
0260 end
0261 
0262 %% repackage lambdas
0263 if isempty(extra) || ~isfield(extra, 'lambda') || isempty(extra.lambda)
0264     lambda = struct( ...
0265         'mu_l', zeros(nA, 1), ...
0266         'mu_u', zeros(nA, 1), ...
0267         'lower', zeros(nx, 1), ...
0268         'upper', zeros(nx, 1) ...
0269     );
0270 else
0271     lam.eqlin = extra.lambda(1:neq);
0272     lam.ineqlin = extra.lambda(neq+(1:nie));
0273     lam.lower = extra.redcosts;
0274     lam.upper = -extra.redcosts;
0275     lam.lower(lam.lower < 0) = 0;
0276     lam.upper(lam.upper < 0) = 0;
0277     kl = find(lam.eqlin > 0);   %% lower bound binding
0278     ku = find(lam.eqlin < 0);   %% upper bound binding
0279 
0280     mu_l = zeros(nA, 1);
0281     mu_l(ieq(kl)) = lam.eqlin(kl);
0282     mu_l(igt) = -lam.ineqlin(nlt+(1:ngt));
0283     mu_l(ibx) = -lam.ineqlin(nlt+ngt+nbx+(1:nbx));
0284 
0285     mu_u = zeros(nA, 1);
0286     mu_u(ieq(ku)) = -lam.eqlin(ku);
0287     mu_u(ilt) = -lam.ineqlin(1:nlt);
0288     mu_u(ibx) = -lam.ineqlin(nlt+ngt+(1:nbx));
0289 
0290     lambda = struct( ...
0291         'mu_l', mu_l, ...
0292         'mu_u', mu_u, ...
0293         'lower', lam.lower(1:nx), ...
0294         'upper', lam.upper(1:nx) ...
0295     );
0296 end
0297 
0298 if mi && eflag == 1 && (~isfield(opt, 'skip_prices') || ~opt.skip_prices)
0299     if verbose
0300         fprintf('--- Integer stage complete, starting price computation stage ---\n');
0301     end
0302     tol = 1e-7;
0303     k = find(vtype == 'I' | vtype == 'B');
0304     x(k) = round(x(k));
0305     xmin(k) = x(k);
0306     xmax(k) = x(k);
0307     x0 = x;
0308     opt.glpk_opt.lpsolver = 1;      %% simplex
0309     opt.glpk_opt.dual = 0;          %% primal simplex
0310     if have_fcn('octave') && have_fcn('octave', 'vnum') >= 3.007
0311         opt.glpk_opt.dual = 1;      %% primal simplex
0312     end
0313     
0314     [x_, f_, eflag_, output_, lambda] = qps_glpk(H, c, A, l, u, xmin, xmax, x0, opt);
0315     if eflag ~= eflag_
0316         error('miqps_glpk: EXITFLAG from price computation stage = %d', eflag_);
0317     end
0318     if abs(f - f_)/max(abs(f), 1) > tol
0319         warning('miqps_glpk: relative mismatch in objective function value from price computation stage = %g', abs(f - f_)/max(abs(f), 1));
0320     end
0321     xn = x;
0322     xn(abs(xn)<1) = 1;
0323     [mx, k] = max(abs(x - x_) ./ xn);
0324     if mx > tol
0325         warning('miqps_glpk: max relative mismatch in x from price computation stage = %g (%g)', mx, x(k));
0326     end
0327     output.price_stage = output_;
0328 end

Generated on Fri 20-Mar-2015 18:23:34 by m2html © 2005