« first day (1478 days earlier)      last day (3012 days later) » 

Jason B
Jason B
1:01 PM
Just out of curiosity here,

http://mathematica.stackexchange.com/questions/105623/color-a-single-contour-from-a-listcontourplot3d-using-a-values-from-another-list

Can anyone who voted to close this post give me the reasoning? The OP clearly did include enough information to answer the question. The question is valid, and doesn't arise from a simple mistake. The actual data being plotted is irrelevant to answering the question.
 
 
2 hours later…
Jason B
Jason B
3:11 PM
meh, no big deal, not really important
 
Szabolcs
Szabolcs
@JasonB Try pinging them ...
I have a mesh region. What is the quickest way to identify the boundary points by cell index? (Not coordinates, but cell index.)
 
Szabolcs
Szabolcs
3:34 PM
Right now I am using a very naive and slow First@First@Position[MeshCoordinates[dreg], #] & /@ MeshCoordinates@RegionBoundary[dreg]; where dreg is my mesh region.
 
Jason B
Jason B
I don't know if this is any better, but MeshCoordinates@
RegionBoundary[dreg] /. (MapIndexed[Rule, MeshCoordinates[dreg]] //
Dispatch) // Flatten would give the same I think
 
Szabolcs
Szabolcs
Another question: I need a GraphDistanceMatrix for a subset of nodes only. My graph has tens of thousands of nodes but I need the matrix between only a thousand or so, and I need it fast. Does anyone have experience with this?
What would make sense: use GraphDistance[graph, #]& /@ vertices and throw away what I don't need. But GraphDistance is surprisingly slow and I haven't gotten to the bottom of it yet ...
Maybe it's time to update IGraph/M!
 
 
2 hours later…
Searke
Searke
5:35 PM
@Szabolcs I don't know much about graph theory algorithms, but I thought the logic behind a distance matrix was that asking for the distances between every vertex offered a speed up compared to asking for the minimum distance between each pair of vertices individually
That is, to getting the min distance from a vertex to all other vertex is just a o(n log n) depth first search
If you're only interested in a subset, you don't get that, so it's asymptotically equal to just asking for the min distance between every pair of vertices you're interested in.
Unless you plan on doing something tricky.
Or just plan on making a faster implementation
 
 
1 hour later…
Szabolcs
Szabolcs
6:43 PM
@Searke Yes, that's right. But that means that in an 1000 vertex graph it would still make sense to compute only a 10*1000 distance matrix instead of an 1000*1000 one. But you are right that 10*10 woulnd't be faster than 10*1000. What bothers me is this timing: 
In[5]:= g = GridGraph[{1, 1, 1} 20];

In[6]:= GraphDistanceMatrix[g]; // AbsoluteTiming

Out[6]= {1.34153, Null}

In[9]:= GraphDistance[g, 1]; // AbsoluteTiming

Out[9]= {0.647454, Null}
Why is GraphDistance so slow relative to GraphDistanceMatrix?
One to all distance is only 2x faster than all-to-all distance.
 
 
2 hours later…
Emad
Emad
8:19 PM
hello All
I wanna solve a non-linear Overdetermined system of equations(7 equations with six unknown).
would you please somebody help me how can I do it in Mathematica.
I have tried:
NMinimize[{eqn11, eqn2, eqn3, eqn4, eqn5, eqn6, eqn1}, {a[2], c[2],
a[3], b[3], c[3], d}]

NMinimize::nnum: The function value False is not a number at {d,Subscript[a, 2],Subscript[a, 3],Subscript[b, 3],Subscript[c, 2],Subscript[c, 3]} = {0.990791,0.281773,0.300512,0.930797,-0.345651,0.281501}. >>

NMinimize::nnum: The function value False is not a number at {d,Subscript[a, 2],Subscript[a, 3],Subscript[b, 3],Subscript[c, 2],Subscript[c, 3]} = {0.990791,0.281773,0.300512,0.930797,-0.345651,0.281501}. >>
But it didn't work out.
 

« first day (1478 days earlier)      last day (3012 days later) »