Elias at 22:28, 17 January 2012

2012-01-17T22:28:20Z

← Older revision		Revision as of 17:28, 17 January 2012
Line 45:		Line 45:
	Open the file for viewing, and execute mcpQG in Matlab.		Open the file for viewing, and execute mcpQG in Matlab.

	<~~syntaxhighlight~~ lang="matlab">		<source lang="matlab">
	% mcpQG.m		% mcpQG.m
	%		%
Line 175:		Line 175:
	%		%
	% I.e, we have the complementarity conditions satisfied.		% I.e, we have the complementarity conditions satisfied.
	</~~syntaxhighlight~~>		</source>

Elias at 06:42, 11 August 2011

2011-08-11T06:42:07Z

← Older revision		Revision as of 01:42, 11 August 2011
Line 1:		Line 1:
			{{Part Of Manual\|title=the Quickguide Manual\|link=[[Quickguide\|Quickguide]]}}
	TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem). It is recommended that the user supply as many analytical functions as possible.		TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem). It is recommended that the user supply as many analytical functions as possible.

Elias: Created page with "TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem..."

2011-08-11T06:41:55Z

Created page with "TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem..."

New page

TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem). It is recommended that the user supply as many analytical functions as possible.

The constrained nonlinear complementarity problem is defined as:

<math>
\begin{array}{ll}
\min\limits_{x} & f(x) \\
& \\
s/t & \begin{array}{lcccl}
x_{L} & \leq & x & \leq & x_{U} \\
b_{L} & \leq & A x & \leq & b_{U} \\
c_{L} & \leq & c(x) & \leq & c_{U} \\
\end{array}
\end{array}
</math>

The complementarity conditions can be any combination of decision variables and linear/nonlinear constraints:

*<math>x(i) \perp x(j)</math>
*<math>x(i) \perp A(j,:) x</math>
*<math>x(i) \perp c(j)</math>
*<math>A(i,:) x \perp A(j,:) x</math>
*<math>A(i,:) x \perp c(j)</math>
*<math>c(i) \perp c(j)</math>

The following files define the problem in TOMLAB.

'''File: '''<span class="roman">tomlab/quickguide/mcpQG_f.m, mcpQG_g.m, mcpQG_H.m, mcpQG_c.m, mcpQG_dc.m, mcpQG_d2c.m</span>

<pre>
f: Function value
g: Gradient vector
H: Hessian matrix
c: Nonlinear constraint vector
dc: Nonlinear constraint gradient matrix
d2c: The second part of the Hessian to the Lagrangian function for the nonlinear constraints.
</pre>

The following file illustrates how to solve a MCP problem in TOMLAB. Also view the m-files specified above for more information.

'''File: '''<span class="roman">tomlab/quickguide/mcpQG.m</span>

Open the file for viewing, and execute mcpQG in Matlab.

<syntaxhighlight lang="matlab">
% mcpQG.m
%
% Demonstrates how to setup and solve a nonlinear problem with equilibrium
% (complementary) constraints.
%
% TOMLAB /KNITRO is capable of solving general complementarity problems
%
% The problem is the following:
%
% minimize f = (x(1)-5)^2 + (2*x(2)+1)^2
%
% subject to
%
% 2*(x(2)-1) - 1.5*x(1) + x(3) - 0.5*x(4) + x(5) ==0 % c(1)
% 3*x(1) - x(2) - 3 >= 0 % c(2)
% -x(1) + 0.5*x(2) + 4 >= 0 % c(3)
% -x(1) - x(2) + 7 >= 0 % c(4)
%
% x(1:5) >= 0
%
% The following complementarity demands are also imposed:
%
% 0 <= c(2) _|_ x(3) >= 0
% 0 <= c(3) _|_ x(4) >= 0
% 0 <= c(4) _|_ x(5) >= 0
%
% All four constraints are in fact linear, but are modelled as nonlinear
% constraints with the constants moved to the LHS.
%
% The same problem is solved as a explicitly quadratic complementarity
% problem in qcpQG.m

% There are no (explicitly) linear constraints.
A = [];
b_L = [];
b_U = [];

% Provide a pattern for the nonlinear constraints
ConsPattern = [
1 1 1 1 1
1 1 1 0 0
1 1 0 0 0
1 1 0 0 0
];

HessPattern = [];

% First constraint is equality ==0, the remaining are >= 0
c_L = [0,0,0,0]';
c_U = [0,inf,inf,inf]';

x_L = zeros(5,1);
x_U = inf*ones(5,1);
x_0 = x_L;

Name = 'JF Bard 1998 MPEC';

% Functions for calculating the nonlinear function and derivative values.
f = 'mcpQG_f';
g = 'mcpQG_g';
H = 'mcpQG_H';
c = 'mcpQG_c';
dc = 'mcpQG_dc';
d2c = 'mcpQG_d2c';

% Each row of mpec is one pair. The first says x(3) _|_ c(2) should be
% complementary. Exactly two nonzeros per row is allowed.

mpec = [ ...
3,0, 0,0, 2,0; ...
4,0, 0,0, 3,0; ...
5,0, 0,0, 4,0; ...
];

Prob = mcpAssign(f, g, H, HessPattern, x_L, x_U, Name, x_0, mpec, ...
[], ...
A, b_L, b_U, c, dc, d2c, ConsPattern, c_L, c_U);

Prob.PriLevOpt = 2;

% KNITRO options. Algorithm 3 works good on this problem.
opts = [];
opts.ALG = 3;
Prob.KNITRO.options = opts;

% Now solve this problem with KNITRO:
Result = tomRun('knitro',Prob,2);

% The original problem is available as Prob.orgProb.

N0 = Prob.orgProb.N;

x = Result.x_k;
s = x(N0+1:end) % The slacks added by BuildMPEC
x = x(1:N0) % The "original" variables

% Constraint values for the modified problem. If not infeasible, these
% should all be zero because BuildMPEC has changed the constraints in the
% mpec pairs to include slack variables.

c = Result.c_k

% The permutation matrix for the slacks in the nonlinear constraints is
% available:

MP = full(Prob.MPEC.MP);

% The values of the constraints + their respective slacks:

C = c + MP*s

x(3:5), C(2:4)

% This should give something like:
%
% ans =
%
% 3.5000
% 0
% 0
%
%
% ans =
%
% 0
% 3
% 6
%
% I.e, we have the complementarity conditions satisfied.
</syntaxhighlight>

Quickguide MCP Problem - Revision history

Elias at 22:28, 17 January 2012

Elias at 06:42, 11 August 2011

Elias: Created page with "TOMLAB requires that general nonlinear complementarity problems are defined in Matlab m-files. The function to be optimized does not need to be supplied (pure equilibrium problem..."