Applications, Examples and Libraries

Share your work here

I experienced a significant obstacle while trying to generate independent random samples with Statistics:-Sample on different nodes of a Grid multi-processing environment. After many hours of trial-and-error, I discovered an astonishing workaround, and I achieved excellent time and memory performance. Since this seems like a generally useful computation, I thought that it was worthy of a Post.

This Post is also worth reading to learn how to use Grid when you need to initialize a substantial environment on each node before using Grid:-Map or Grid:-Seq.

All remaining details are in the following worksheet.
 

How to use Statistics:-Sample in the `Grid` environment

Author: Carl Love <carl.j.love@gmail.com> 1 August 2019

 

I experienced a significant obstacle while trying to generate indenpendent random samples with Statistics:-Sample on the nodes of a multi-processor Grid (on a single computer). After several hours of trial-and-error, I discovered that two things are necessary to do this:

1. 

The random number generator needs to be seeded differently in each node. (The reason for this is easy to understand.)

2. 

The random variables generated by Statistics:-RandomVariable need to have different names in each node. This one is mind-boggling to me. Afterall, each node has its own kernel and so its own memory It's as if the names of random variables are stored in a disk file which all kernels access. And also the generator has been seeded differently in each node.

 

Once these things were done, the time and memory performance of the computation were excellent.

restart
:

Digits:= 15
:

#Specify the size of the computation:
(n1,n2,n3):= (100, 100, 1000):
# n1 = size of each random sample;
# n2 = number of samples in a batch;
# n3 = number of batches.

#
#Procedure to initialize needed globals on each node:
Init:= proc(n::posint)
local node:= Grid:-MyNode();
   #This is wrapped in parse so that it'll act globally. Otherwise, an environment
   #variable would be reset when this procedure ends.
   parse("Digits:= 15;", 'statement');

   randomize(randomize()+node); #Initialize independent RNG for this node.
   #If repeatability of results is desired, remove the inner randomize().

   (:-X,:-Y):= Array(1..n, 'datatype'= 'hfloat') $ 2;

   #Perhaps due to some oversight in the design of Statistics, it seems necessary that
   #r.v.s in different nodes **need different names** in order to be independent:
   N||node:= Statistics:-RandomVariable('Normal'(0,1));
   :-TRS:= (X::rtable)-> Statistics:-Sample(N||node, X);
   #To verify that different names are needed, change N||node to N in both lines.
   #Doing so, each node will generate identical samples!

   #Perform some computation. For the pedagogical purpose of this worksheet, all that
   #matters is that it's some numeric computation on some Arrays of random Samples.
   :-GG:= (X::Array, Y::Array)->
      evalhf(
         proc(X::Array, Y::Array, n::posint)
         local s, k, S:= 0, p:= 2*Pi;
            for k to n do
               s:= sin(p*X[k]);  
               S:= S + X[k]^2*cos(p*Y[k])/sqrt(2-sin(s)) + Y[k]^2*s
            od
         end proc
         (X, Y, n)
      )      
   ;
   #Perform a batch of the above computations, and somehow numerically consolidate the
   #results. Once again, pedagogically it doesn't matter how they're consolidated.  
   :-TRX1:= (n::posint)-> add(GG(TRS(X), TRS(Y)), 1..n);
   
   #It doesn't matter much what's returned. Returning `node` lets us verify that we're
   #actually running this on a grid.
   return node
end proc
:

The procedure Init above uses the :- syntax to set variables globally for each node. The variables set are X, Y, N||node, TRS, GG, and TRX1. Names constructed by concatenation, such as N||node, are always global, so :- isn't needed for those.

#
#Time the initialization:
st:= time[real]():
   #Send Init to each node, but don't run it yet:
   Grid:-Set(Init)
   ;
   #Run Init on each node:
   Nodes:= Grid:-Run(Init, [n1], 'wait');
time__init_Grid:= time[real]() - st;

Array(%id = 18446745861500764518)

1.109

The only purpose of array Nodes is that it lets us count the nodes, and it lets us verify that Grid:-MyNode() returned a different value on each node.

num_nodes:= numelems(Nodes);

8

#Time the actual execution:
st:= time[real]():
   R1:= [Grid:-Seq['tasksize'= iquo(n3, num_nodes)](TRX1(k), k= [n2 $ n3])]:
time__run_Grid:= time[real]() - st

4.440

#Just for comparison, run it sequentially:
st:= time[real]():
   Init(n1):
time__init_noGrid:= time[real]() - st;

st:= time[real]():
   R2:= [seq(TRX1(k), k= [n2 $ n3])]:
time__run_noGrid:= time[real]() - st;

0.16e-1

24.483

R1 and R2 will be different because different random numbers were used, but they should have similar histograms.

plots:-display(
   Statistics:-Histogram~(
      <R1 | R2>, #side-by-side plots
      'title'=~ <<"With Grid\n"> | <"Without Grid\n">>,
      'gridlines'= false
   )
);

(Plot output deleted because MaplePrimes cannot handle side-by-side plots!)

They look similar enough to me!

 

Let's try to quantify the benefit of using Grid:

speedup_factor:= time__run_noGrid / time__run_Grid;

5.36319824753560

Express that as a fraction of the theoretical maximum speedup:

efficiency:= speedup_factor / num_nodes;

.670399780941950

I think that that's really good!

 

The memory usage of this code is insignificant, which can be verified from an external memory monitor such as Winodws Task Manager. It's just a little bit more than that needed to start a kernel on each node. It's also possible to measure the memory usage programmatically. Doing so for a Grid:-Seq computation is a little bit beyond the scope of this worksheet.

 


 

Download GridRandSample.mw

Here are the histograms:

In this post, the Numbrix Puzzle is solved by the branch and bound method (see the details of this puzzle in  https://www.mapleprimes.com/posts/210643-Solving-A-Numbrix-Puzzle-With-Logic). The main difference from the solution using the  Logic  package is that here we get not one but all possible solutions. In the case of a unique solution, the  NumbrixPuzzle procedure is faster than the  Numbrix  one (for convenience, I inserted the code for Numbrix procedure into the worksheet below). In the case of many solutions, the  Numbrix  procedure is usually faster (see all the examples below).

 

restart;

NumbrixPuzzle:=proc(A::Matrix)
local A1, L, N, S, MS, OneStepLeft, OneStepRight, F1, F2, m, L1, p, q, a, b, T, k, s1, s, H, n, L2, i, j, i1, j1, R;
uses ListTools;
S:=upperbound(A); N:=nops(op(A)[3]); MS:=`*`(S);
A1:=convert(A, listlist);
for i from 1 to S[1] do
for j from 1 to S[2] do
for i1 from i to S[1] do
for j1 from 1 to S[2] do
if A1[i,j]<>0 and A1[i1,j1]<>0 and abs(A1[i,j]-A1[i1,j1])<abs(i-i1)+abs(j-j1) then return `no solutions` fi;
od; od; od; od;
L:=sort(select(e->e<>0, Flatten(A1)));
L1:=[`if`(L[1]>1,seq(L[1]-k, k=0..L[1]-2),NULL)];
L2:=[seq(seq(`if`(L[i+1]-L[i]>1,L[i]+k,NULL),k=0..L[i+1]-L[i]-2), i=1..nops(L)-1), `if`(L[-1]<MS,seq(L[-1]+k,k=0..MS-L[-1]-1),NULL)];
  

OneStepLeft:=proc(A1::listlist)
local s, M, m, k, T;
uses ListTools;
s:=Search(a, Matrix(A1));   
M:=[[s[1]-1,s[2]],[s[1]+1,s[2]],[s[1],s[2]-1],[s[1],s[2]+1]];
T:=table(); k:=0;
for m in M do
if m[1]>=1 and m[1]<=S[1] and m[2]>=1 and m[2]<=S[2] and A1[op(m)]=0 then k:=k+1; T[k]:=subsop(m=a-1,A1);
fi;
od;
convert(T, list);
end proc;

 
OneStepRight:=proc(A1::listlist)
local s, M, m, k, T, s1;
uses ListTools;
s:=Search(a, Matrix(A1));  s1:=Search(a+2, Matrix(A1));  
M:=[[s[1]-1,s[2]],[s[1]+1,s[2]],[s[1],s[2]-1],[s[1],s[2]+1]];
T:=table(); k:=0;
for m in M do
if m[1]>=1 and m[1]<=S[1] and m[2]>=1 and m[2]<=S[2] and A1[op(m)]=0 and `if`(a+2 in L, `if`(is(abs(s1[1]-m[1])+abs(s1[2]-m[2])>1),false,true),true) then k:=k+1; T[k]:=subsop(m=a+1,A1);
fi;
od;
convert(T, list);   
end proc;

F1:=LM->ListTools:-FlattenOnce(map(OneStepLeft, LM));
F2:=LM->ListTools:-FlattenOnce(map(OneStepRight, LM));

T:=[A1];
for a in L1 do
T:=F1(T);
od;

for a in L2 do
T:=F2(T);
od;

R:=map(t->convert(t,Matrix), T);
if nops(R)=0 then return `no solutions` else R[] fi;

end proc:

Numbrix := proc( M :: ~Matrix, { inline :: truefalse := false } )

local S, adjacent, eq, i, initial, j, k, kk, m, n, one, single, sol, unique, val, var, x;

    (m,n) := upperbound(M);

    initial := &and(seq(seq(ifelse(M[i,j] = 0
                                   , NULL
                                   , x[i,j,M[i,j]]
                                  )
                            , i = 1..m)
                        , j = 1..n));

    adjacent := &and(seq(seq(seq(x[i,j,k] &implies &or(NULL
                                                       , ifelse(i>1, x[i-1, j, k+1], NULL)
                                                       , ifelse(i<m, x[i+1, j, k+1], NULL)
                                                       , ifelse(j>1, x[i, j-1, k+1], NULL)
                                                       , ifelse(j<n, x[i, j+1, k+1], NULL)
                                                      )
                                 , i = 1..m)
                             , j = 1..n)
                         , k = 1 .. m*n-1));

    one := &or(seq(seq(x[i,j,1], i=1..m), j=1..n));   


    single := &not(&or(seq(seq(seq(seq(x[i,j,k] &and x[i,j,kk], kk = k+1..m*n), k = 1..m*n-1)
                                , i = 1..m), j = 1..n)));

    sol := Logic:-Satisfy(&and(initial, adjacent, one, single));
    
    if sol = NULL then
        error "no solution";
    end if;
if inline then
        S := M;
     else
        S := Matrix(m,n);
    end if;

    for eq in sol do
        (var, val) := op(eq);
        if val then
            S[op(1..2, var)] := op(3,var);
        end if;
    end do;
    S;
end proc:

           Two simple examples

A:=<0,0,5; 0,0,0; 0,0,9>;
# The unique solution
NumbrixPuzzle(A);

A:=<0,0,5; 0,0,0; 0,8,0>;
# 4 solutions
NumbrixPuzzle(A);

Matrix(3, 3, {(1, 1) = 0, (1, 2) = 0, (1, 3) = 5, (2, 1) = 0, (2, 2) = 0, (2, 3) = 0, (3, 1) = 0, (3, 2) = 0, (3, 3) = 9})

 

Matrix(3, 3, {(1, 1) = 3, (1, 2) = 4, (1, 3) = 5, (2, 1) = 2, (2, 2) = 7, (2, 3) = 6, (3, 1) = 1, (3, 2) = 8, (3, 3) = 9})

 

Matrix(3, 3, {(1, 1) = 0, (1, 2) = 0, (1, 3) = 5, (2, 1) = 0, (2, 2) = 0, (2, 3) = 0, (3, 1) = 0, (3, 2) = 8, (3, 3) = 0})

 

Matrix(%id = 18446746210121682686), Matrix(%id = 18446746210121682806), Matrix(%id = 18446746210121674750), Matrix(%id = 18446746210121674870)

(1)


Comparison with Numbrix procedure. The example is taken from
http://rosettacode.org/wiki/Solve_a_Numbrix_puzzle 

 A:=<0, 0, 0, 0, 0, 0, 0, 0, 0;
 0, 0, 46, 45, 0, 55, 74, 0, 0;
 0, 38, 0, 0, 43, 0, 0, 78, 0;
 0, 35, 0, 0, 0, 0, 0, 71, 0;
 0, 0, 33, 0, 0, 0, 59, 0, 0;
 0, 17, 0, 0, 0, 0, 0, 67, 0;
 0, 18, 0, 0, 11, 0, 0, 64, 0;
 0, 0, 24, 21, 0, 1, 2, 0, 0;
 0, 0, 0, 0, 0, 0, 0, 0, 0>;
CodeTools:-Usage(NumbrixPuzzle(A));
CodeTools:-Usage(Numbrix(A));

Matrix(9, 9, {(1, 1) = 0, (1, 2) = 0, (1, 3) = 0, (1, 4) = 0, (1, 5) = 0, (1, 6) = 0, (1, 7) = 0, (1, 8) = 0, (1, 9) = 0, (2, 1) = 0, (2, 2) = 0, (2, 3) = 46, (2, 4) = 45, (2, 5) = 0, (2, 6) = 55, (2, 7) = 74, (2, 8) = 0, (2, 9) = 0, (3, 1) = 0, (3, 2) = 38, (3, 3) = 0, (3, 4) = 0, (3, 5) = 43, (3, 6) = 0, (3, 7) = 0, (3, 8) = 78, (3, 9) = 0, (4, 1) = 0, (4, 2) = 35, (4, 3) = 0, (4, 4) = 0, (4, 5) = 0, (4, 6) = 0, (4, 7) = 0, (4, 8) = 71, (4, 9) = 0, (5, 1) = 0, (5, 2) = 0, (5, 3) = 33, (5, 4) = 0, (5, 5) = 0, (5, 6) = 0, (5, 7) = 59, (5, 8) = 0, (5, 9) = 0, (6, 1) = 0, (6, 2) = 17, (6, 3) = 0, (6, 4) = 0, (6, 5) = 0, (6, 6) = 0, (6, 7) = 0, (6, 8) = 67, (6, 9) = 0, (7, 1) = 0, (7, 2) = 18, (7, 3) = 0, (7, 4) = 0, (7, 5) = 11, (7, 6) = 0, (7, 7) = 0, (7, 8) = 64, (7, 9) = 0, (8, 1) = 0, (8, 2) = 0, (8, 3) = 24, (8, 4) = 21, (8, 5) = 0, (8, 6) = 1, (8, 7) = 2, (8, 8) = 0, (8, 9) = 0, (9, 1) = 0, (9, 2) = 0, (9, 3) = 0, (9, 4) = 0, (9, 5) = 0, (9, 6) = 0, (9, 7) = 0, (9, 8) = 0, (9, 9) = 0})

 

memory used=7.85MiB, alloc change=-3.01MiB, cpu time=172.00ms, real time=212.00ms, gc time=93.75ms

 

Matrix(9, 9, {(1, 1) = 49, (1, 2) = 50, (1, 3) = 51, (1, 4) = 52, (1, 5) = 53, (1, 6) = 54, (1, 7) = 75, (1, 8) = 76, (1, 9) = 81, (2, 1) = 48, (2, 2) = 47, (2, 3) = 46, (2, 4) = 45, (2, 5) = 44, (2, 6) = 55, (2, 7) = 74, (2, 8) = 77, (2, 9) = 80, (3, 1) = 37, (3, 2) = 38, (3, 3) = 39, (3, 4) = 40, (3, 5) = 43, (3, 6) = 56, (3, 7) = 73, (3, 8) = 78, (3, 9) = 79, (4, 1) = 36, (4, 2) = 35, (4, 3) = 34, (4, 4) = 41, (4, 5) = 42, (4, 6) = 57, (4, 7) = 72, (4, 8) = 71, (4, 9) = 70, (5, 1) = 31, (5, 2) = 32, (5, 3) = 33, (5, 4) = 14, (5, 5) = 13, (5, 6) = 58, (5, 7) = 59, (5, 8) = 68, (5, 9) = 69, (6, 1) = 30, (6, 2) = 17, (6, 3) = 16, (6, 4) = 15, (6, 5) = 12, (6, 6) = 61, (6, 7) = 60, (6, 8) = 67, (6, 9) = 66, (7, 1) = 29, (7, 2) = 18, (7, 3) = 19, (7, 4) = 20, (7, 5) = 11, (7, 6) = 62, (7, 7) = 63, (7, 8) = 64, (7, 9) = 65, (8, 1) = 28, (8, 2) = 25, (8, 3) = 24, (8, 4) = 21, (8, 5) = 10, (8, 6) = 1, (8, 7) = 2, (8, 8) = 3, (8, 9) = 4, (9, 1) = 27, (9, 2) = 26, (9, 3) = 23, (9, 4) = 22, (9, 5) = 9, (9, 6) = 8, (9, 7) = 7, (9, 8) = 6, (9, 9) = 5})

 

memory used=1.21GiB, alloc change=307.02MiB, cpu time=37.00s, real time=31.88s, gc time=9.30s

 

Matrix(%id = 18446746210094669942)

(2)


In the example below, which has 104 solutions, the  Numbrix  procedure is faster.

C:=Matrix(5,{(1,1)=1,(5,5)=25});
CodeTools:-Usage(NumbrixPuzzle(C)):
nops([%]);
CodeTools:-Usage(Numbrix(C)):

Matrix(5, 5, {(1, 1) = 1, (1, 2) = 0, (1, 3) = 0, (1, 4) = 0, (1, 5) = 0, (2, 1) = 0, (2, 2) = 0, (2, 3) = 0, (2, 4) = 0, (2, 5) = 0, (3, 1) = 0, (3, 2) = 0, (3, 3) = 0, (3, 4) = 0, (3, 5) = 0, (4, 1) = 0, (4, 2) = 0, (4, 3) = 0, (4, 4) = 0, (4, 5) = 0, (5, 1) = 0, (5, 2) = 0, (5, 3) = 0, (5, 4) = 0, (5, 5) = 25})

 

memory used=0.94GiB, alloc change=-22.96MiB, cpu time=12.72s, real time=11.42s, gc time=2.28s

 

104

 

memory used=34.74MiB, alloc change=0 bytes, cpu time=781.00ms, real time=783.00ms, gc time=0ns

 

 


 

Download NumbrixPuzzle.mw

For no particular reason at all, these are parametric equations that print "Maplesoft" in handwritten cursive script when plotted

restart:
X := -2.05*sin(-2.70 + 2.45*t) - 3.36*sin(1.12 + 1.43*t) - 4.82*sin(-2.19 + 2.03*t) - 2.02*sin(1.36 + 2.31*t) - 2.41*sin(1.08 + 2.59*t) - 14.2*sin(1.51 + 0.185*t) - 5.25*sin(-2.04 + 1.85*t) - 2.81*sin(0.984 + 2.36*t) - 3.01*sin(-2.04 + 1.80*t) - 1.80*sin(-2.61 + 2.73*t) - 0.712*sin(-3.94 + 1.89*t) - 6.90*sin(-1.90 + 1.52*t) - 0.600*sin(-3.39 + 2.26*t) - 0.631*sin(-4.65 + 2.68*t) - 3.10*sin(-2.22 + 2.17*t) - 2.95*sin(1.38 + 1.25*t) - 1.43*sin(0.383 + 2.40*t) - 8.25*sin(-1.66 + 0.323*t) - 1.39*sin(-3.08 + 2.63*t) - 0.743*sin(-2.43 + 0.647*t) - 6.25*sin(-1.73 + 0.832*t) - 273.*sin(-1.58 + 0.0462*t) - 4.58*sin(-2.00 + 1.48*t) - 5.70*sin(-1.80 + 1.20*t) - 2.30*sin(1.42 + 0.462*t) - 3.24*sin(1.51 + 0.277*t) - 16.0*sin(-1.64 + 0.231*t) - 1.58*sin(0.779 + 1.71*t) - 0.571*sin(-2.08 + 0.970*t) - 8.85*sin(-1.88 + 1.34*t) - 1.10*sin(-2.24 + 2.08*t) - 1.49*sin(-2.27 + 1.02*t) - 2.19*sin(-1.70 + 1.94*t) - 4.47*sin(-2.06 + 1.57*t) - 2.08*sin(-2.02 + 1.06*t) - 5.70*sin(-1.86 + 1.62*t) - 2.26*sin(-1.66 + 1.16*t) - 3.95*sin(-1.98 + 1.29*t) - 0.928*sin(-2.08 + 1.76*t) - 2.98*sin(1.36 + 1.11*t) - 0.390*sin(-2.33 + 2.22*t) - 3.81*sin(1.01 + 2.54*t) - 0.613*sin(-1.43 + 1.66*t) - 19.7*sin(-1.60 + 0.138*t) - 0.524*sin(-2.87 + 0.414*t) - 2.15*sin(-4.63 + 0.694*t) - 0.782*sin(-1.56 + 2.49*t) - 5.27*sin(-1.81 + 1.38*t) - 5.18*sin(1.51 + 0.0923*t) - 6.83*sin(1.37 + 0.923*t) - 0.814*sin(-1.72 + 0.600*t) - 2.98*sin(-1.82 + 0.738*t) - 5.49*sin(1.44 + 0.509*t) - 3.90*sin(-1.76 + 0.785*t) - 0.546*sin(-2.18 + 0.876*t) - 1.92*sin(0.755 + 1.98*t) - 8.16*sin(1.38 + 0.553*t) - 0.504*sin(-1.56 + 0.371*t) - 3.43*sin(1.14 + 2.12*t):
Y := -1.05*sin(-3.81 + 2.68*t) - 7.72*sin(-4.59 + 0.231*t) - 6.38*sin(1.37 + 1.11*t) - 4.24*sin(-2.36 + 2.31*t) - 7.06*sin(1.18 + 1.80*t) - 4.60*sin(1.28 + 2.03*t) - 0.626*sin(-0.285 + 2.45*t) - 0.738*sin(-1.89 + 2.26*t) - 1.45*sin(-1.73 + 1.57*t) - 2.30*sin(-4.51 + 2.59*t) - 9.58*sin(-2.07 + 1.71*t) - 0.792*sin(-0.578 + 0.647*t) - 4.55*sin(1.49 + 1.25*t) - 14.0*sin(-2.13 + 1.62*t) - 1.02*sin(0.410 + 0.277*t) - 19.2*sin(-1.54 + 0.0462*t) - 17.3*sin(-1.86 + 1.20*t) - 1.96*sin(-0.845 + 2.63*t) - 0.754*sin(-0.0904 + 2.73*t) - 4.74*sin(1.11 + 1.48*t) - 1.79*sin(0.860 + 2.17*t) - 25.2*sin(-1.77 + 0.832*t) - 3.88*sin(1.30 + 0.462*t) - 20.8*sin(-1.66 + 0.323*t) - 17.6*sin(1.20 + 1.29*t) - 4.83*sin(0.169 + 2.36*t) - 10.8*sin(-2.01 + 1.85*t) - 8.69*sin(-2.17 + 2.22*t) - 5.48*sin(-1.69 + 1.34*t) - 18.1*sin(1.18 + 1.43*t) - 4.71*sin(0.728 + 2.08*t) - 1.15*sin(-3.44 + 1.52*t) - 2.53*sin(-2.61 + 2.54*t) - 5.48*sin(-2.02 + 1.94*t) - 4.67*sin(1.30 + 1.66*t) - 9.10*sin(1.37 + 0.970*t) - 6.45*sin(1.31 + 1.02*t) - 5.18*sin(-2.09 + 1.76*t) - 18.3*sin(-1.77 + 1.06*t) - 27.3*sin(1.31 + 1.16*t) - 2.83*sin(-3.01 + 2.40*t) - 2.93*sin(-1.70 + 0.138*t) - 4.17*sin(-2.06 + 2.12*t) - 1.60*sin(-4.25 + 1.38*t) - 2.69*sin(-1.89 + 0.371*t) - 7.92*sin(-1.78 + 0.600*t) - 19.6*sin(-1.79 + 0.738*t) - 22.6*sin(1.48 + 0.509*t) - 13.5*sin(1.21 + 0.923*t) - 5.53*sin(-1.64 + 0.0923*t) - 1.20*sin(0.145 + 2.49*t) - 3.15*sin(-1.57 + 0.414*t) - 1.74*sin(0.655 + 1.98*t) - 3.98*sin(-2.14 + 0.876*t) - 11.3*sin(-1.82 + 0.694*t) - 10.4*sin(0.987 + 1.89*t) - 8.39*sin(-1.53 + 0.185*t) - 27.8*sin(-1.76 + 0.785*t) - 9.39*sin(1.38 + 0.553*t):
plot([X, Y, t = 0 .. 68], scaling = constrained, axes = boxed);

Hare in the forest

The rocket flies

  

Быльнов_raketa_letit.mws

 

Plotting the function of a complex variable

Plotting_the_function_of_a_complex_variable.mws

 

Animated 3-D cascade of dolls

 

3d_matryoshkas_en.mws

 

With this application developed entirely in Maple using native syntax and embedded components for science and engineering students. Just replace your data and you're done.

Pearson_Coeficient.mw

Lenin Araujo Castillo

Ambassador of Maple

 

Foucault’s Pendulum Exploration Using MAPLE18

https://www.ias.ac.in/describe/article/reso/024/06/0653-0659

In this article, we develop the traditional differential equation for Foucault’s pendulum from physical situation and solve it from
standard form. The sublimation of boundary condition eliminates the constants and choice of the local parameters (latitude, pendulum specifications) offers an equation that can be used for a plot followed by animation using MAPLE. The fundamental conceptual components involved in preparing differential equation viz; (i) rotating coordinate system, (ii) rotation of the plane of oscillation and its dependence on the latitude, (iii) effective gravity with latitude, etc., are discussed in detail. The accurate calculations offer quantities up to the sixth decimal point which are used for plotting and animation. This study offers a hands-on experience. Present article offers a know-how to devise a Foucault’s pendulum just by plugging in the latitude of reader’s choice. Students can develop a miniature working model/project of the pendulum.

Exercises solved online with Maple exclusively in space. I attach the explanation links on my YouTube channel.

Part # 01

https://www.youtube.com/watch?v=8Aa2xzU8LwQ

Part # 02

https://www.youtube.com/watch?v=qyGT28CeSz4

Part # 03

https://www.youtube.com/watch?v=yf8rjSPbv5g

Part # 04

https://www.youtube.com/watch?v=FwHPW7ncZTg

Part # 05

https://www.youtube.com/watch?v=bm3frpukb0I

Link for download the file:

Vector_Exercises-Force_in_space.mw

Lenin AC

Ambassador of Maple

 

 

 


 

Solving a Numbrix Puzzle with Logic

Background

 

 

Parade magazine, a filler in the local Sunday newspaper, contains a Numbrix puzzle, the object of which is to find a serpentine path of consecutive integers, 1 through 81, through a nine by nine grid.  The puzzle typically specifies the content of every other border cell.

 

The Maple Logic  package has a procedure, Satisfy , that can be used to solve this puzzle.  Satisfy is a SAT-solver; given a boolean expression it attempts to find a set of equations of the form {x__1 = b__1, x__2 = b__2, () .. ()}, where x__i are the boolean variables in the given expression and b__i are boolean values (true or false) that satisfy the expression (cause it to evaluate true).

 

A general technique to solve this and other puzzles with Satisfy is to select boolean-values variables that encode the state of the puzzle (a trial solution, whether valid or not), generate a boolean-expression of these variables that is satisfied (true) if and only if the variables are given values that correspond to a solution, pass this expression to Satisfy, then translate the returned set of boolean values (if any) to the puzzle solution.

Setup

 

Assign a matrix that defines the grid and the initial position.  Use zeros to indicate the cells that need values. To make it easy to inspect the expressions, a small 2 x 3 matrix is used for this demo---a full size example is given at the end.

M := Matrix(2,3, {(1,1) = 1, (1,3) = 5});

Matrix(2, 3, {(1, 1) = 1, (1, 2) = 0, (1, 3) = 5, (2, 1) = 0, (2, 2) = 0, (2, 3) = 0})

(2.1)

 

Extract the dimensions of the Matrix

(m,n) := upperbound(M);

2, 3

(2.2)

Boolean Variables

 

Let the boolean variable x[i,j,k] mean that cell (i,j) has value k. For example, x[2,3,6] is true when cell (2,3) contains 6, otherwise it is false. There are (m*n)^2 boolean variables.

Initial Position

 

The initial position can be expressed as the following and-clause.

initial := &and(seq(seq(ifelse(M[i,j] = 0, NULL, x[i,j,M[i,j]]), i = 1..m), j = 1..n));

`&and`(x[1, 1, 1], x[1, 3, 5])

(4.1)

Adjacent Cells

 

The requirement that an interior cell with value k is adjacent to the cell with value k+1 can be expressed as the implication
   

   x[i,j,k] &implies &or(x[i-1,j,k+1], x[i+1,j,k+1], x[i,j-1,k+1], x[i,j+1,k+1])

 

Extending this to handle all cells results in the following boolean expression.

adjacent := &and(seq(seq(seq(
         x[i,j,k] &implies &or(NULL
                               , ifelse(1<i, x[i-1, j, k+1], NULL)
                               , ifelse(i<m, x[i+1, j, k+1], NULL)
                               , ifelse(1<j, x[i, j-1, k+1], NULL)
                               , ifelse(j<n, x[i, j+1, k+1], NULL)
                               )
                            , i = 1..m)
                        , j = 1..n)
                    , k = 1 .. m*n-1));

`&and`(`&implies`(x[1, 1, 1], `&or`(x[2, 1, 2], x[1, 2, 2])), `&implies`(x[2, 1, 1], `&or`(x[1, 1, 2], x[2, 2, 2])), `&implies`(x[1, 2, 1], `&or`(x[2, 2, 2], x[1, 1, 2], x[1, 3, 2])), `&implies`(x[2, 2, 1], `&or`(x[1, 2, 2], x[2, 1, 2], x[2, 3, 2])), `&implies`(x[1, 3, 1], `&or`(x[2, 3, 2], x[1, 2, 2])), `&implies`(x[2, 3, 1], `&or`(x[1, 3, 2], x[2, 2, 2])), `&implies`(x[1, 1, 2], `&or`(x[2, 1, 3], x[1, 2, 3])), `&implies`(x[2, 1, 2], `&or`(x[1, 1, 3], x[2, 2, 3])), `&implies`(x[1, 2, 2], `&or`(x[2, 2, 3], x[1, 1, 3], x[1, 3, 3])), `&implies`(x[2, 2, 2], `&or`(x[1, 2, 3], x[2, 1, 3], x[2, 3, 3])), `&implies`(x[1, 3, 2], `&or`(x[2, 3, 3], x[1, 2, 3])), `&implies`(x[2, 3, 2], `&or`(x[1, 3, 3], x[2, 2, 3])), `&implies`(x[1, 1, 3], `&or`(x[2, 1, 4], x[1, 2, 4])), `&implies`(x[2, 1, 3], `&or`(x[1, 1, 4], x[2, 2, 4])), `&implies`(x[1, 2, 3], `&or`(x[2, 2, 4], x[1, 1, 4], x[1, 3, 4])), `&implies`(x[2, 2, 3], `&or`(x[1, 2, 4], x[2, 1, 4], x[2, 3, 4])), `&implies`(x[1, 3, 3], `&or`(x[2, 3, 4], x[1, 2, 4])), `&implies`(x[2, 3, 3], `&or`(x[1, 3, 4], x[2, 2, 4])), `&implies`(x[1, 1, 4], `&or`(x[2, 1, 5], x[1, 2, 5])), `&implies`(x[2, 1, 4], `&or`(x[1, 1, 5], x[2, 2, 5])), `&implies`(x[1, 2, 4], `&or`(x[2, 2, 5], x[1, 1, 5], x[1, 3, 5])), `&implies`(x[2, 2, 4], `&or`(x[1, 2, 5], x[2, 1, 5], x[2, 3, 5])), `&implies`(x[1, 3, 4], `&or`(x[2, 3, 5], x[1, 2, 5])), `&implies`(x[2, 3, 4], `&or`(x[1, 3, 5], x[2, 2, 5])), `&implies`(x[1, 1, 5], `&or`(x[2, 1, 6], x[1, 2, 6])), `&implies`(x[2, 1, 5], `&or`(x[1, 1, 6], x[2, 2, 6])), `&implies`(x[1, 2, 5], `&or`(x[2, 2, 6], x[1, 1, 6], x[1, 3, 6])), `&implies`(x[2, 2, 5], `&or`(x[1, 2, 6], x[2, 1, 6], x[2, 3, 6])), `&implies`(x[1, 3, 5], `&or`(x[2, 3, 6], x[1, 2, 6])), `&implies`(x[2, 3, 5], `&or`(x[1, 3, 6], x[2, 2, 6])))

(5.1)

 

All Values Used

 

The following expression is true when each integer k, from 1 to m*n, is assigned to one or more cells.

allvals := &and(seq(seq(&or(seq(x[i,j,k], k=1..m*n)), i=1..m), j=1..n));

`&and`(`&or`(x[1, 1, 1], x[1, 1, 2], x[1, 1, 3], x[1, 1, 4], x[1, 1, 5], x[1, 1, 6]), `&or`(x[2, 1, 1], x[2, 1, 2], x[2, 1, 3], x[2, 1, 4], x[2, 1, 5], x[2, 1, 6]), `&or`(x[1, 2, 1], x[1, 2, 2], x[1, 2, 3], x[1, 2, 4], x[1, 2, 5], x[1, 2, 6]), `&or`(x[2, 2, 1], x[2, 2, 2], x[2, 2, 3], x[2, 2, 4], x[2, 2, 5], x[2, 2, 6]), `&or`(x[1, 3, 1], x[1, 3, 2], x[1, 3, 3], x[1, 3, 4], x[1, 3, 5], x[1, 3, 6]), `&or`(x[2, 3, 1], x[2, 3, 2], x[2, 3, 3], x[2, 3, 4], x[2, 3, 5], x[2, 3, 6]))

(6.1)

Single Value

 

The following expression is satisfied when each cell has no more than one value.

 single := &not &or(seq(seq(seq(seq(x[i,j,k] &and x[i,j,kk], kk = k+1..m*n), k = 1..m*n-1), i = 1..m), j = 1..n));

`&not`(`&or`(`&and`(x[1, 1, 1], x[1, 1, 2]), `&and`(x[1, 1, 1], x[1, 1, 3]), `&and`(x[1, 1, 1], x[1, 1, 4]), `&and`(x[1, 1, 1], x[1, 1, 5]), `&and`(x[1, 1, 1], x[1, 1, 6]), `&and`(x[1, 1, 2], x[1, 1, 3]), `&and`(x[1, 1, 2], x[1, 1, 4]), `&and`(x[1, 1, 2], x[1, 1, 5]), `&and`(x[1, 1, 2], x[1, 1, 6]), `&and`(x[1, 1, 3], x[1, 1, 4]), `&and`(x[1, 1, 3], x[1, 1, 5]), `&and`(x[1, 1, 3], x[1, 1, 6]), `&and`(x[1, 1, 4], x[1, 1, 5]), `&and`(x[1, 1, 4], x[1, 1, 6]), `&and`(x[1, 1, 5], x[1, 1, 6]), `&and`(x[2, 1, 1], x[2, 1, 2]), `&and`(x[2, 1, 1], x[2, 1, 3]), `&and`(x[2, 1, 1], x[2, 1, 4]), `&and`(x[2, 1, 1], x[2, 1, 5]), `&and`(x[2, 1, 1], x[2, 1, 6]), `&and`(x[2, 1, 2], x[2, 1, 3]), `&and`(x[2, 1, 2], x[2, 1, 4]), `&and`(x[2, 1, 2], x[2, 1, 5]), `&and`(x[2, 1, 2], x[2, 1, 6]), `&and`(x[2, 1, 3], x[2, 1, 4]), `&and`(x[2, 1, 3], x[2, 1, 5]), `&and`(x[2, 1, 3], x[2, 1, 6]), `&and`(x[2, 1, 4], x[2, 1, 5]), `&and`(x[2, 1, 4], x[2, 1, 6]), `&and`(x[2, 1, 5], x[2, 1, 6]), `&and`(x[1, 2, 1], x[1, 2, 2]), `&and`(x[1, 2, 1], x[1, 2, 3]), `&and`(x[1, 2, 1], x[1, 2, 4]), `&and`(x[1, 2, 1], x[1, 2, 5]), `&and`(x[1, 2, 1], x[1, 2, 6]), `&and`(x[1, 2, 2], x[1, 2, 3]), `&and`(x[1, 2, 2], x[1, 2, 4]), `&and`(x[1, 2, 2], x[1, 2, 5]), `&and`(x[1, 2, 2], x[1, 2, 6]), `&and`(x[1, 2, 3], x[1, 2, 4]), `&and`(x[1, 2, 3], x[1, 2, 5]), `&and`(x[1, 2, 3], x[1, 2, 6]), `&and`(x[1, 2, 4], x[1, 2, 5]), `&and`(x[1, 2, 4], x[1, 2, 6]), `&and`(x[1, 2, 5], x[1, 2, 6]), `&and`(x[2, 2, 1], x[2, 2, 2]), `&and`(x[2, 2, 1], x[2, 2, 3]), `&and`(x[2, 2, 1], x[2, 2, 4]), `&and`(x[2, 2, 1], x[2, 2, 5]), `&and`(x[2, 2, 1], x[2, 2, 6]), `&and`(x[2, 2, 2], x[2, 2, 3]), `&and`(x[2, 2, 2], x[2, 2, 4]), `&and`(x[2, 2, 2], x[2, 2, 5]), `&and`(x[2, 2, 2], x[2, 2, 6]), `&and`(x[2, 2, 3], x[2, 2, 4]), `&and`(x[2, 2, 3], x[2, 2, 5]), `&and`(x[2, 2, 3], x[2, 2, 6]), `&and`(x[2, 2, 4], x[2, 2, 5]), `&and`(x[2, 2, 4], x[2, 2, 6]), `&and`(x[2, 2, 5], x[2, 2, 6]), `&and`(x[1, 3, 1], x[1, 3, 2]), `&and`(x[1, 3, 1], x[1, 3, 3]), `&and`(x[1, 3, 1], x[1, 3, 4]), `&and`(x[1, 3, 1], x[1, 3, 5]), `&and`(x[1, 3, 1], x[1, 3, 6]), `&and`(x[1, 3, 2], x[1, 3, 3]), `&and`(x[1, 3, 2], x[1, 3, 4]), `&and`(x[1, 3, 2], x[1, 3, 5]), `&and`(x[1, 3, 2], x[1, 3, 6]), `&and`(x[1, 3, 3], x[1, 3, 4]), `&and`(x[1, 3, 3], x[1, 3, 5]), `&and`(x[1, 3, 3], x[1, 3, 6]), `&and`(x[1, 3, 4], x[1, 3, 5]), `&and`(x[1, 3, 4], x[1, 3, 6]), `&and`(x[1, 3, 5], x[1, 3, 6]), `&and`(x[2, 3, 1], x[2, 3, 2]), `&and`(x[2, 3, 1], x[2, 3, 3]), `&and`(x[2, 3, 1], x[2, 3, 4]), `&and`(x[2, 3, 1], x[2, 3, 5]), `&and`(x[2, 3, 1], x[2, 3, 6]), `&and`(x[2, 3, 2], x[2, 3, 3]), `&and`(x[2, 3, 2], x[2, 3, 4]), `&and`(x[2, 3, 2], x[2, 3, 5]), `&and`(x[2, 3, 2], x[2, 3, 6]), `&and`(x[2, 3, 3], x[2, 3, 4]), `&and`(x[2, 3, 3], x[2, 3, 5]), `&and`(x[2, 3, 3], x[2, 3, 6]), `&and`(x[2, 3, 4], x[2, 3, 5]), `&and`(x[2, 3, 4], x[2, 3, 6]), `&and`(x[2, 3, 5], x[2, 3, 6])))

(7.1)

Solution

 

Combine the boolean expressions into a a single and-clause and pass it to Satisfy.

sol := Logic:-Satisfy(&and(initial, adjacent, allvals, single));

{x[1, 1, 1] = true, x[1, 1, 2] = false, x[1, 1, 3] = false, x[1, 1, 4] = false, x[1, 1, 5] = false, x[1, 1, 6] = false, x[1, 2, 1] = false, x[1, 2, 2] = false, x[1, 2, 3] = false, x[1, 2, 4] = false, x[1, 2, 5] = false, x[1, 2, 6] = true, x[1, 3, 1] = false, x[1, 3, 2] = false, x[1, 3, 3] = false, x[1, 3, 4] = false, x[1, 3, 5] = true, x[1, 3, 6] = false, x[2, 1, 1] = false, x[2, 1, 2] = true, x[2, 1, 3] = false, x[2, 1, 4] = false, x[2, 1, 5] = false, x[2, 1, 6] = false, x[2, 2, 1] = false, x[2, 2, 2] = false, x[2, 2, 3] = true, x[2, 2, 4] = false, x[2, 2, 5] = false, x[2, 2, 6] = false, x[2, 3, 1] = false, x[2, 3, 2] = false, x[2, 3, 3] = false, x[2, 3, 4] = true, x[2, 3, 5] = false, x[2, 3, 6] = false}

(8.1)

Select the equations whose right size is true

sol := select(rhs, sol);

{x[1, 1, 1] = true, x[1, 2, 6] = true, x[1, 3, 5] = true, x[2, 1, 2] = true, x[2, 2, 3] = true, x[2, 3, 4] = true}

(8.2)

Extract the lhs of the true equations

vars := map(lhs, sol);

{x[1, 1, 1], x[1, 2, 6], x[1, 3, 5], x[2, 1, 2], x[2, 2, 3], x[2, 3, 4]}

(8.3)

Extract the result from the indices of the vars and assign to a new Matrix

S := Matrix(m,n):

for v in vars do S[op(1..2,v)] := op(3,v); end do:

S;

Matrix(2, 3, {(1, 1) = 1, (1, 2) = 6, (1, 3) = 5, (2, 1) = 2, (2, 2) = 3, (2, 3) = 4})

(8.4)

Procedure

 

We can now combine the manual steps into a procedure that takes an initialized Matrix and fills in a solution.

Numbrix := proc( M :: ~Matrix, { inline :: truefalse := false } )

Example

 

Create the initial position for a 9 x 9 Numbrix and solve it.

P := Matrix(9, {(1,1)=11, (1,3)=7, (1,5)=3, (1,7)=81, (1,9)=77, (3,9)=75, (5,9)=65, (7,9)=55, (9,9)=53
               , (9,7)=47, (9,5)=41, (9,3)=39, (9,1)=37, (7,1)=21, (5,1)=17, (3,1)=13});

Matrix(9, 9, {(1, 1) = 11, (1, 2) = 0, (1, 3) = 7, (1, 4) = 0, (1, 5) = 3, (1, 6) = 0, (1, 7) = 81, (1, 8) = 0, (1, 9) = 77, (2, 1) = 0, (2, 2) = 0, (2, 3) = 0, (2, 4) = 0, (2, 5) = 0, (2, 6) = 0, (2, 7) = 0, (2, 8) = 0, (2, 9) = 0, (3, 1) = 13, (3, 2) = 0, (3, 3) = 0, (3, 4) = 0, (3, 5) = 0, (3, 6) = 0, (3, 7) = 0, (3, 8) = 0, (3, 9) = 75, (4, 1) = 0, (4, 2) = 0, (4, 3) = 0, (4, 4) = 0, (4, 5) = 0, (4, 6) = 0, (4, 7) = 0, (4, 8) = 0, (4, 9) = 0, (5, 1) = 17, (5, 2) = 0, (5, 3) = 0, (5, 4) = 0, (5, 5) = 0, (5, 6) = 0, (5, 7) = 0, (5, 8) = 0, (5, 9) = 65, (6, 1) = 0, (6, 2) = 0, (6, 3) = 0, (6, 4) = 0, (6, 5) = 0, (6, 6) = 0, (6, 7) = 0, (6, 8) = 0, (6, 9) = 0, (7, 1) = 21, (7, 2) = 0, (7, 3) = 0, (7, 4) = 0, (7, 5) = 0, (7, 6) = 0, (7, 7) = 0, (7, 8) = 0, (7, 9) = 55, (8, 1) = 0, (8, 2) = 0, (8, 3) = 0, (8, 4) = 0, (8, 5) = 0, (8, 6) = 0, (8, 7) = 0, (8, 8) = 0, (8, 9) = 0, (9, 1) = 37, (9, 2) = 0, (9, 3) = 39, (9, 4) = 0, (9, 5) = 41, (9, 6) = 0, (9, 7) = 47, (9, 8) = 0, (9, 9) = 53})

(10.1)

CodeTools:-Usage(Numbrix(P));

memory used=0.77GiB, alloc change=220.03MiB, cpu time=15.55s, real time=12.78s, gc time=3.85s

 

Matrix(9, 9, {(1, 1) = 11, (1, 2) = 10, (1, 3) = 7, (1, 4) = 81, (1, 5) = 3, (1, 6) = 4, (1, 7) = 81, (1, 8) = 78, (1, 9) = 77, (2, 1) = 12, (2, 2) = 9, (2, 3) = 8, (2, 4) = 7, (2, 5) = 6, (2, 6) = 5, (2, 7) = 80, (2, 8) = 79, (2, 9) = 76, (3, 1) = 13, (3, 2) = 14, (3, 3) = 27, (3, 4) = 28, (3, 5) = 71, (3, 6) = 72, (3, 7) = 73, (3, 8) = 74, (3, 9) = 75, (4, 1) = 16, (4, 2) = 15, (4, 3) = 26, (4, 4) = 29, (4, 5) = 70, (4, 6) = 69, (4, 7) = 68, (4, 8) = 67, (4, 9) = 66, (5, 1) = 17, (5, 2) = 18, (5, 3) = 25, (5, 4) = 30, (5, 5) = 61, (5, 6) = 62, (5, 7) = 63, (5, 8) = 64, (5, 9) = 65, (6, 1) = 20, (6, 2) = 19, (6, 3) = 24, (6, 4) = 31, (6, 5) = 60, (6, 6) = 59, (6, 7) = 58, (6, 8) = 57, (6, 9) = 56, (7, 1) = 21, (7, 2) = 22, (7, 3) = 23, (7, 4) = 32, (7, 5) = 43, (7, 6) = 44, (7, 7) = 49, (7, 8) = 50, (7, 9) = 55, (8, 1) = 36, (8, 2) = 35, (8, 3) = 34, (8, 4) = 33, (8, 5) = 42, (8, 6) = 45, (8, 7) = 48, (8, 8) = 51, (8, 9) = 54, (9, 1) = 37, (9, 2) = 38, (9, 3) = 39, (9, 4) = 40, (9, 5) = 41, (9, 6) = 46, (9, 7) = 47, (9, 8) = 52, (9, 9) = 53})

(10.2)

 

numbrix.mw

I describe here a finite difference scheme for solving the boundary value
problem for the heat equation

"(&PartialD; u)/(&PartialD; t)= ((&PartialD;)^)/((&PartialD;)^( )x^)(c(x)(&PartialD; u)/(&PartialD; x)) + f(x,t)   a<x<b,   t>0"

for the unknown temperature u(x, t)subject to the boundary conditions

u(a, t) = alpha(t), u(b, t) = beta(t), t > 0

and the initial condition

"u(x,0)=`u__0`(x),    a < x < b."

 

This finite difference scheme is designed expressly with the goal of avoiding

differentiating the conductivity c(x), therefore c(x) is allowed to be

nonsmooth or even discontinuous. A discontinuous c(x) is particularly
important in applications where the heat conduction takes place through layers
of distinct types of materials.

 

The animation below, extracted from the worksheet, demonstrates a solution 

corresponding to a discontinuous c(x).  The limit of that solution as time goes to

infinity, which may be calculated independently and exactly, is shown as a gray
line.

Download worksheet: heat-finite-difference.mw

 

 

 

 

Maple 2019 has a new add-on package Maple Quantum Chemistry Toolbox from RDMChem for computing the energies and properties of molecules.  As a member of the team at RDMChem that developed the package, I would like to tell the story of its origins and provide a brief demonstration of the package.  

 

Thinking about Quantum Chemistry at Harvard

 

The story of the Maple Quantum Chemistry Toolbox begins with my graduate studies in Chemical Physics at Harvard University in the late 1990s.  Even in 1998 programs for computing the energies and properties of molecules were extremely complicated and nonintuitive.  Many of the existing programs had begun in the 1970s on computers whose programs would be recorded on punchcards.

Fig. 1: Used Punchcard by Pete Birkinshaw from Manchester, UK CC BY 2.0

 

Even today some of these programs have remnants of their early versions such as input files that must start on the second column to account for the margin of the now non-existent punchcards.  As a student, I made a bound copy of one of these manuals at a local Kinkos photocopy shop and later found myself in Harvard Yard, thinking that there must be a better way to present quantum chemistry computations.  The idea for a Maple-like package for quantum chemistry was born in that moment.

 

At the same time I was learning about something called the two-electron reduced density matrix (2-RDM).  The basic variable in quantum chemistry is the wave function which is the probability amplitude for finding each of the electrons in a molecule.  Because electrons are indistinguishable with pairwise interactions, the wave function contains much more information than is needed for computing the energies and electronic properties of molecules.  The energies and properties of any molecule with any number of electrons can be expressed as a function of a 2 electron matrix, the 2-RDM [1-3].  A quantum chemistry based on the 2-RDM, it was known, would have potentially significant advantages over wave function calculations in terms of accuracy and computational cost, especially for molecules far from the mean-field limit.  A 2-RDM approach to quantum chemistry became the focus of my Ph.D. thesis.

 

Representing Many Electrons with Only Two Electrons

 

The idea of using the 2-RDM in quantum chemistry can be attributed to four scientists: two physicists Kodi Husimi and Joseph Mayer, a chemist Per-Olov Lowdin, and a mathematician John Coleman [1-3].  In the early 1940s Husimi first published the idea in a Japanese physics journal, but in the midst of World War II the paper was not widely disseminated in the West.  In the summer of 1951 John Coleman, which attending a physics conference at Chalk River, realized that the ground-state energy of any atom or molecule could be expressed as functional of the 2-RDM, and similar ideas later occurred to Per-Olov Lowdin and Joseph Mayer who published their ideas in Physical Review in 1955.  It was soon recognized that computing the ground-state energy of an atom or molecule with the 2-RDM was potentially difficult because not every two-electron density matrix corresponds to an N-electron density matrix or wave function.  The search for the appropriate constraints on the 2-RDM, known as N-representability conditions, became known as the N-representability problem [1-3].  

 

Beginning in the late 1990s and early 2000s, Carmela Valdemoro and Diego Alcoba at the Consejo Superior de Investigaciones Científicas (Madrid, Spain), Hiroshi Nakatsuji, Koji Yasuda, and Maho Nakata at Kyoto University (Kyoto, Japan), Jerome Percus and Bastiaan Braams at the Courant Institute (New York, USA), John Coleman and Robert Erdahl at Queens University (Kingston, Canada), and my research group and I at The University of Chicago (Chicago, USA) began to make significant progress in the computation of the 2-RDM without computing the many-electron wave function [1-3].  Further contributions were made by Eric Cances and Claude Le Bris at CERMICS, Ecole Nationale des Ponts et Chaussées (Marne-la-Vallée, France), Paul Ayers at McMaster University (Hamilton, Canada), and Dimitri Van Neck at the University of Ghent (Ghent, Belgium) and their research groups.  By 2014 several powerful 2-RDM methods had emerged for the computation of molecules.  The Army Research Office (ARO) issued a proposal call for a company to develop a modern, built-from-scratch package for quantum chemistry that would contain two newly developed 2-RDM-based methods from our group: the parametric 2-RDM method [1] and the variational 2-RDM method with a fast algorithm for solving the semidefinite program [4,5,6].   The company RDMChem LLC was founded to work with the ARO to develop such a package built around RDMs, and hence, the name of the company RDMChem was selected as a hybrid of the RDM abbreviation for Reduced Density Matrices and the Chem colloquialism for Chemistry.  To achieve a really new design for an electronic structure package with access to numeric and symbolic computations as well as advanced visualizations, the team at RDMChem and I developed a partnership with Maplesoft to build something new that became the Maple Quantum Chemistry Package (or Toolbox), which was released with Maple 2019 on Pi Day.

 

Maple Quantum Chemistry Toolbox

 The Maple Quantum Chemistry Toolbox provides a powerful, parallel platform for quantum chemistry calculations that is directly integrated into the Maple 2019 environment.  It is optimized for both cutting-edge research as well as chemistry education.  The Toolbox can be used from the worksheet, document, or command-line interfaces.  Plus there is a Maplet interface for rapid exploration of molecules and their properties.  Figure 2 shows the Maplet interface being applied to compute the ground-state energy of 1,3-dibromobenzene by density functional theory (DFT) in a 6-31g basis set.           

Fig. 2: Maplet interface to the Quantum Chemistry Toolbox 2019, showing a density functional theory (DFT) calculation         

After entering a name into the text box labeled Name, the user can click on: (1) the button Web to import the geometry from an online database containing more than 96 million molecules,  (2) the button File to read the geometry from a standard XYZ file, or (3) the button Input to enter the geometry.  As soon the geometry is entered, the Maplet displays a 3D picture of the molecule in the window on the right of the options.  Dropdown menus allow the user to select the basis set, the electronic structure method, and a boolean for geometry optimization.  The user can click on the Compute button to perform the computation.  When the quantum computation completes, the total energy appears in the box labeled Total Energy.  The dropdown menu Analyze contains a list of data tables, plots, and animations that can be selected and then displayed by clicking the Analyze button.  The Maplet interface contains nearly all of the options available in the worksheet interface.   The Help Pages of the Toolbox include extensive curricula and lessons that can be used in undergraduate, graduate, and even high school chemistry courses.  Next we look at some sample calculations in the worksheet interface.     

 

Reproducing an Early 2-RDM Calculation

 

One of the earliest variational calculations of the 2-RDM was performed in 1975 by Garrod, Mihailović,  and  Rosina [1-3].  They minimized the electronic ground state of the 4-electron atom beryllium as a functional of only two electrons, the 2-RDM.  They imposed semidefinite constraints on the particle-particle (D), hole-hole (Q), and particle-hole (G) metric matrices.  They solved the resulting optimization problem of minimizing the energy as a linear function of the 2-RDM subject to the semidefinite constraints, known as a semidefinite program, by a cutting-plane algorithm.  Due to limitations of the cutting-plane algorithm and computers circa 1975, the calculation was a difficult one, likely taking a significant amount of computer time and memory.

 

With the Quantum Chemistry Toolbox we can use the command Variational2RDM to reproduce the calculation on a Windows laptop.  First, in a Maple 2019 worksheet we load the commands of the Add-on Quantum Chemistry Toolbox:

with(QuantumChemistry);

[AOLabels, ActiveSpaceCI, ActiveSpaceSCF, AtomicData, BondAngles, BondDistances, Charges, ChargesPlot, CorrelationEnergy, CoupledCluster, DensityFunctional, DensityPlot3D, Dipole, DipolePlot, Energy, FullCI, GeometryOptimization, HartreeFock, Interactive, Isotopes, MOCoefficients, MODiagram, MOEnergies, MOIntegrals, MOOccupations, MOOccupationsPlot, MOSymmetries, MP2, MolecularData, MolecularGeometry, NuclearEnergy, NuclearGradient, Parametric2RDM, PlotMolecule, Populations, RDM1, RDM2, ReadXYZ, SaveXYZ, SearchBasisSets, SearchFunctionals, SkeletalStructure, Thermodynamics, Variational2RDM, VibrationalModeAnimation, VibrationalModes, Video]

(1.1)

Then we define the atom (or molecule) using a Maple list of lists that we assign to the variable atom:

atom := [["Be",0,0,0]];

[["Be", 0, 0, 0]]

(1.2)

 

We can then perform the variational 2-RDM method with the Variational2RDM command to compute the ground-state energy and properties of beryllium in a minimal basis set like the one used by Rosina and his collaborators.  By default the method uses the D, Q, and G N-representability conditions and the minimal "sto-3g" basis set.  The calculation, which completes in seconds, contains a wealth of information in the form of a convenient Maple table that we assign to the variable data.

data := Variational2RDM(atom);

table(%id = 18446744313704784158)

(1.3)

 

The table contains the total ground-state energy of the beryllium atom in the atomic unit of energy (hartrees)

data[e_tot];

HFloat(-14.40370016681039)

(1.4)

 

We also have the atomic orbitals (AOs) employed in the calculation

data[aolabels];

Vector(5, {(1) = "0 Be 1s", (2) = "0 Be 2s", (3) = "0 Be 2px", (4) = "0 Be 2py", (5) = "0 Be 2pz"})

(1.5)

 

as well as the Mulliken populations of these orbitals

data[populations];

Vector(5, {(1) = 1.9995807710723152, (2) = 1.7913484714571852, (3) = 0.6969023822632789e-1, (4) = 0.6969026475511847e-1, (5) = 0.6969029119010149e-1})

(1.6)

 

We see that 2 electrons are located in the 1s orbital, 1.8 electrons in the 2s orbital, and about 0.2 electrons in the 2p orbitals.  By default the calculation also returns the 1-RDM

data[rdm1];

Matrix(5, 5, {(1, 1) = 1.9999258249189755, (1, 2) = -0.37784860208539793e-2, (1, 3) = 0., (1, 4) = 0., (1, 5) = 0., (2, 1) = -0.37784860208539793e-2, (2, 2) = 1.7910034176105256, (2, 3) = 0., (2, 4) = 0., (2, 5) = 0., (3, 1) = 0., (3, 2) = 0., (3, 3) = 0.6969023822632789e-1, (3, 4) = 0., (3, 5) = 0., (4, 1) = 0., (4, 2) = 0., (4, 3) = 0., (4, 4) = 0.6969026475511847e-1, (4, 5) = 0., (5, 1) = 0., (5, 2) = 0., (5, 3) = 0., (5, 4) = 0., (5, 5) = 0.6969029119010149e-1})

(1.7)

 

The eigenvalues of the 1-RDM are the natural orbital occupations

LinearAlgebra:-Eigenvalues(data[rdm1]);

Vector(5, {(1) = 1.9999941387490443+0.*I, (2) = 1.7909351037804568+0.*I, (3) = 0.6969023822632789e-1+0.*I, (4) = 0.6969026475511847e-1+0.*I, (5) = 0.6969029119010149e-1+0.*I})

(1.8)

 

We can display the density of the 2s-like 2nd natural orbital using the DensityPlot3D command providing the atom, the data, and the orbitalindex keyword

DensityPlot3D(atom,data,orbitalindex=2);

 

 

Similarly,  using the DensityPlot3D command, we can readily display the 2p-like 3rd natural orbital

DensityPlot3D(atom,data,orbitalindex=3);

 

 

By using Maple keyword arguments in the Variational2RDM command, we can readily change the basis set, use point-group symmetry, add active orbitals with or without self-consistent-field, change the N-representability conditions, as well as explore many other options.  Having reenacted one of the first variational 2-RDM calculations ever, let's examine a more complicated molecule.

 

Explosive TNT

 

We consider the molecule TNT that is used as an explosive. Using the command MolecularGeometry, we can import the experimental geometry of TNT from the online PubChem database.

mol := MolecularGeometry("TNT");

[["O", .5454, -3.514, 0.12e-2], ["O", .5495, 3.5137, 0.8e-3], ["O", 2.4677, -2.4539, -0.5e-3], ["O", 2.4705, 2.4513, 0.3e-3], ["O", -3.5931, -1.0959, 0.4e-3], ["O", -3.5922, 1.0993, 0.6e-3], ["N", 1.2142, -2.454, 0.2e-3], ["N", 1.217, 2.4527, 0], ["N", -2.9846, 0.15e-2, 0.1e-3], ["C", 1.2253, -0.6e-3, -0.9e-3], ["C", .5271, -1.2082, -0.8e-3], ["C", .5284, 1.2078, -0.8e-3], ["C", -1.5646, 0.8e-3, -0.4e-3], ["C", -.8678, -1.2074, -0.6e-3], ["C", -.8666, 1.2084, -0.6e-3], ["C", 2.7239, -0.16e-2, 0.11e-2], ["H", -1.4159, -2.1468, -0.3e-3], ["H", -1.4137, 2.1483, -0.3e-3], ["H", 3.1226, .2418, -.9891], ["H", 3.0863, .6934, .7662], ["H", 3.3154, -.8111, .4109]]

(1.9)

 

The command PlotMolecule generates a 3D ball-and-stick plot of the molecule

PlotMolecule(mol);

 

 

We perform a variational calculation of the 2-RDM of TNT in an active space of 10 electrons and 10 orbitals by setting the keyword active to the list [10,10].  The keyword casscf is set to true to optimize the active orbitals during the calculation.  The keyword basis is used to set the basis set to a minimal basis set sto-3g for illustration.   

data := Variational2RDM(mol, active=[10,10], casscf=true, basis="sto-3g");

table(%id = 18446744493271367454)

(1.10)

 

The ground-state energy of TNT in hartrees is

data[e_tot];

HFloat(-868.8629631593426)

(1.11)

 

Unlike beryllium, the electric dipole moment of TNT in debyes is nonzero

data[dipole];

Vector(3, {(1) = .5158925019252739, (2) = -0.5985274393363119e-1, (3) = .1277528280025474})

(1.12)

 

We can easily visualize the dipole moment relative to the molecule's ball-and-stick model with the DipolePlot command

DipolePlot(mol,method=Variational2RDM, active=[10,10], casscf=true, basis="sto-3g");

 

 

The 1-RDM is returned by default

data[rdm1];

_rtable[18446744313709602566]

(1.13)

 

The natural molecular-orbital (MO) occupations are the eigenvalues of the 1-RDM

data[mo_occ];

_rtable[18446744313709600150]

(1.14)

 

All of the occupations can be viewed at once by converting the Vector to a list

convert(data[mo_occ], list);

[HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(2.0), HFloat(1.9110133620349001), HFloat(1.8984139688344246), HFloat(1.6231436866358906), HFloat(1.6158489471020905), HFloat(1.6145310163161273), HFloat(0.38920731792133734), HFloat(0.387039366894289), HFloat(0.37786347287813526), HFloat(0.09734187094597906), HFloat(0.08559699476985069), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0), HFloat(0.0)]

(1.15)

 

We can visualize these occupations with the MOOccupationsPlot command

MOOccupationsPlot(mol,method=Variational2RDM, active=[10,10], casscf=true, basis="sto-3g");

 

 

The occupations, we observe, show significant deviations from 0 and 2, indicating that the electrons have substantial correlation beyond the mean-field (Hartree-Fock) limit.  The blue lines indicate the first N/2 spatial orbitals where N is the total number of electrons while the red lines indicate the remaining spatial orbitals.  We can visualize the highest "occupied" molecular orbital (58) with the DensityPlot3D command

DensityPlot3D(mol,data, orbitalindex=58);

 

 

Similarly, we can visualize the lowest "unoccupied" molecular orbital (59) with the DensityPlot3D command

DensityPlot3D(mol,data, orbitalindex=59);

 

 

Comparison of orbitals 58 and 59 reveals an increase in the number of nodes (changes in the phase of the orbitals denoted by green and purple), which reflects an increase in the energy of the orbital.

 

Looking Ahead

 

The Maple Quantum Chemistry Toolbox 2019, an new Add-on for Maple 2019 from RDMChem, provides a easy-to-use, research-grade environment for the computation of the energies and properties of atoms and molecules.  In this blog we discussed its origins in graduate research at Harvard, its reproduction of an early 2-RDM calculation of beryllium, and its application to the explosive molecule TNT.  We have illustrated only some of the many features and electronic structure methods of the Maple Quantum Chemistry package.  There is much more chemistry and physics to explore.  Enjoy!    

 

Selected References

 

[1] D. A. Mazziotti, Chem. Rev. 112, 244 (2012). "Two-electron Reduced Density Matrix as the Basic Variable in Many-Electron Quantum Chemistry and Physics"

[2]  Reduced-Density-Matrix Mechanics: With Application to Many-Electron Atoms and Molecules (Adv. Chem. Phys.) ; D. A. Mazziotti, Ed.; Wiley: New York, 2007; Vol. 134.

[3] A. J. Coleman and V. I. Yukalov, Reduced Density Matrices: Coulson’s Challenge (Springer-Verlag,  New York, 2000).

[4] D. A. Mazziotti, Phys. Rev. Lett. 106, 083001 (2011). "Large-scale Semidefinite Programming for Many-electron Quantum Mechanics"

[5] A. W. Schlimgen, C. W. Heaps, and D. A. Mazziotti, J. Phys. Chem. Lett. 7, 627-631 (2016). "Entangled Electrons Foil Synthesis of Elusive Low-Valent Vanadium Oxo Complex"

[6] J. M. Montgomery and D. A. Mazziotti, J. Phys. Chem. A 122, 4988-4996 (2018). "Strong Electron Correlation in Nitrogenase Cofactor, FeMoco"

 

Download QCT2019_PrimesV17_05.05.19.mw

In this Post I derive the differential equations of motion of a homogeneous elliptic lamina of mass m and the major and minor axes of lengths of a and b which rolls without slipping along the horizontal x axis within the vertical xy plane.

If the initial angular velocity is large enough, the ellipse will roll forever and go to ±∞ in the x direction, otherwise it will just rock.

I have attached two files:

 rolling-ellipse.mw
        Worksheet to solve the differential equations and animate the motion

rolling-ellipse.pdf
         Documentation containing the derivation of the differential equations

And here are two animations extracted from the worksheet.

First 13 14 15 16 17 18 19 Last Page 15 of 68