CON2VERT - convert a convex set of constraint inequalities into the set of vertices at the intersections of those inequalities;i.e., solve the "vertex enumeration" problem.
V = con2vert(A,b)
Converts the polytope (convex polygon, polyhedron, etc.) defined by the system of inequalities A*x <= b into a list of vertices V. Each ROW of V is a vertex. For n variables:
A = m x n matrix, where m >= n (m constraints, n variables)
b = m x 1 vector (m constraints)
V = p x n matrix (p vertices, n variables)
(1) This program emplyes a primal-dual polytope method.
(2) In dimensions higher than 2, duplicate vertices can appear using this method. This program detects duplicates at up to twelve digits of precision, then returns the unique vertices.
(3) Non-bounding constraints give erroneous results; therefore, the program detects non-bounding constraints and returns an error. You may wish to implement large "box" constraints on your variables if you need to induce bounding. For example, if x is a person's height in feet, the box constraint
-1 <= x <= 1000 would be a reasonable choice to induce boundedness, since no possible solution for x would be prohibited by the bounding box.
(4) This program requires that the feasible region have some finite extent in all dimensions. For example, the feasible region cannot be a line segment in 2-D space, or a plane in 3-D space.
(5) At least two dimensions are required.
(6) See companion function VERT2CON.
(7) Numerous examples are provided.
(8) ver 1.0: initial version, June 2005
(9) ver 1.1: enhanced redundancy checks, July 2005
(10) Written by Michael Kleder
Michael Kleder (2021). CON2VERT - constraints to vertices (https://www.mathworks.com/matlabcentral/fileexchange/7894-con2vert-constraints-to-vertices), MATLAB Central File Exchange. Retrieved .
I've made an updated branch of this code at https://www.mathworks.com/matlabcentral/fileexchange/91595-con2vert-constraints-to-vertices-updated which should fix some of the issues. In particular, this new version (1) has a better method to find a strictly feasible point, now using linear programming, and (2) it can now handle equality constraints (via a change-of-variables trick).
@William Warriner, I don't think that fix is right. You don't want the absolute value there, since then it may be Ax>b whereas we want Ax<b. @MattJ, this new version fixes the issues you had with your example: it can now find an interior point, and it won't return those NaNs (those come from points that were not really vertices and should have been eliminated).
Is it possible to provide a reference to the corresponding algorithm?
This code has some errors.
There appears to be an error in checking for singularities on line 153. It should probably read: if ~all(abs(A*c - b) < tol) for some reasonable tolerance. Changing this line as described fixed an issue with https://www.mathworks.com/matlabcentral/fileexchange/50772-polytope-bounded-voronoi-diagram-in-2d-and-3d. I have also contacted that author.
Explanation or any sort of documentation of the algorithm used would be very helpful
Hello Michael Kleder,
as Reza Jafari before I would like to use your code for my thesis. Would you please send me sufficient information for citation? My mail-address is "email@example.com". Thanks in advance!
Thanks for the useful code which your wrote. I used your code for the part of my dissemination and I need to cite your work. Would you please send me the information that I can cite your code? You can send your email to firstname.lastname@example.org
I want to believe in this tool, but problems with it continue to surface. I will change my rating if we at least get a reference explaining the algorithm that it uses.
As the latest difficulty, CON2VERT fails with this set of data,
A = [0 0 0 -1 2 -1; -1 2 -1 0 0 0;eye(6);-eye(6)];
which looks like it should be fine (bounded and solid in R^6), but convhulln chokes on it in line 158.
Part of the problem is again that it cannot find an interior point c in line 151, but it doesn't end there. Even if I override the code and start at Line 158 with my own interior point
c=[ 2.5000 -5.0000 2.5000 1.4286 -4.2857 1.4286].'
I still get some infinite/nan vertices as output,
NaN -Inf 10.0000 -10.0000 -10.0000 -10.0000
I suspect part of the problem is that Line 167 should be skipped if F is not full rank, but without a reference to the algorithm used, I can't be sure...
Can you give the reference for the algorithm used
With a little more investigation into the problem, I've discovered that with the A,b data in my previous comment, the call to FMINSEARCH fails to find an initial point c that lies inside the polytope.
if ef ~= 1...
fails to catch this. A better test would probably be
I don't really know what a good fix would be. From what I've read, finding an initial point inside a polytope is a sophisticated problem.
Not sure if this file is being supported anymore, but there appears to be a problem. The point x=[2.5,3.5,1.8]'
is easily shown to satisfy
A*x<=b for the A,b data below. However, CON2VERT returns vertices that clearly do not enclose this point.
I'm wondering if the issue might be originating at this line
D = A ./ repmat(b,[1 size(A,2)]);
If any of the b(i) are negative, dividing through by b(i) will flip the direction of the corresponding inequality.
A=[1 0 0
0 1 0
0 0 1
-1 0 0
0 -1 0
0 0 -1
-0.387696185212551 -0.295276274520099 0.703237014010810
5.73326513862449 -0.422448628998633 -0.113292931747631
-3.79255233000727 -0.360645958414035 -0.673892379691217
-1.27188203250127 0.544750863219514 -0.791324376598499
-2.30540199426269 -2.50522755054822 -0.291771858420232
0.616578245027929 -0.654271120193835 0.574406987898971
-1.49508806389885 -2.68877782397954 -0.989225453798092
5.34556895341194 -0.717724903518731 0.589944082263178
-3.40485614479472 -0.0653696838939366 -1.37712939370203
-0.884185847288714 0.840027137739613 -1.49456139060931
-1.91770580905014 -2.20995127602812 -0.995008872431042
0.228882059815378 -0.949547394713933 1.27764400190978
-1.10739187868630 -2.39350154945944 -1.69246246780890
1.94071280861722 -0.783094587412668 -0.787185311438849
4.46138310612323 0.122302234220881 -0.904617308346131
3.42786314436180 -2.92767617954685 -0.405064790167863
5.11668689359657 0.231822491195202 -0.687699919646602
4.23817707472564 -3.11122645297817 -1.10251838554572
-2.52067029750601 -0.905396821633549 0.117431996907282
1.48715033574458 -2.14458159213419 0.382120521270985
-3.17597408497934 -1.01491707860787 -0.0994853917922465
2.29746426610842 -2.32813186556550 -0.315333074106875
-1.03351996176143 -3.04997841376774 0.499552518178267
-0.655303787473336 -0.109520256974320 -0.216917388699529
-0.223206031397585 -3.23352868719905 -0.197901077199593
-1.68882374923476 -3.15949867074206 0.282635129478739
-0.810313930363841 0.183550273431316 0.697453595377860
-0.878509818870922 -3.34304894417337 -0.414818465899121];
Hi, is there a way to resolve the error:
"??? Error using ==> con2vert at 161
Non-bounding constraints detected. (Consider box constraints on variables.)" ?
I wish the program could determinate which vertices belong to the same facet.
Great! It helped me to do my asssignment!
That's what I've been looking for.
works for me. No problems so far. good job! and thanks.
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!