| Here arre some equations that need Solving - Questions are |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Can we Goal seek ? |
|
The puzzle was built for Excel -
with 24 variables. |
|
|
How to maintain INTEGER SOLUTION ? |
|
| Use natural
Letter column Heads for the side lengths |
|
Solution is 38 - side length of
Square N in my notation. |
|
|
| Text of
Equations. Essentially the idea is to
set the lowest few values and let the rest build up.. |
|
Assume I,J |
|
|
|
J or X are smallest |
|
Logical order |
|
| First
guess at some equations - see right for thoughts developing... |
|
G = I + J |
|
| H = G + J |
|
B = D + E + F |
|
C = O + P |
|
P = W + V |
|
H = G + J |
D + G + I |
|
| E = G + H |
|
A = B + D |
|
S = R + X |
|
Q = R + S |
|
E = G + H |
|
| D + G + I |
|
N = K + L |
|
R = T + X |
|
F = E + H |
|
| G = I + J |
|
M = K + N |
|
W = U + V |
|
1) Set X, then manually solve U |
|
B = D + E + F |
| F = E + H |
|
= |
|
A = B + D |
|
|
2) reverse solve the top left ABMN
rectangle. |
|
| A |
B |
C |
D |
E |
F |
G |
H |
I |
J |
K |
L |
M |
N |
O |
P |
Q |
R |
S |
T |
U |
V |
W |
X |
Y |
Z |
Now need more properties of the ABMN rectangle |
|
|
unknowns |
|
L = F + E - I |
|
| 55 |
39 |
81 |
16 |
9 |
14 |
4 |
5 |
3 |
1 |
18 |
20 |
56 |
38 |
30 |
51 |
64 |
31 |
33 |
29 |
35 |
8 |
43 |
2 |
|
|
Now we have sim eqn's |
| K = A + D + I - M |
|
Which defines it now in terms of
knowns except I |
|
But is circular on M |
|
|
And L is now N - K |
|
But Better L is F + E - I |
|
A+B = M + N |
|
| O = B + F +
L + N - C |
|
BUT O and C defined in terms of each
other. |
|
N = K + L |
|
|
what about AMQ-TUC |
|
PV-T ? |
|
Ok |
|
M = K + N |
|
| T = M + N +
O - Q - R |
|
Can
perhaps make this into |
|
SU - RV |
|
UX-V ? |
|
| U = R + T + V - S |
|
subs |
|
|
A+B = K + 2 N |
|
| And we have 4 equations for Z |
|
A+ B = K + 2K + 2L |
|
| Surely
we just need to keep inserting all possible triplets for I,J,X until it
solves. BUT CANNOT FIND FORMULAES FOR
UVK either |
|
3K = A + B - 2 L |
|
|
| Seems
that subtractive formulaes are like saying you have unknowns… |
|
Aha and now need to find I, J that
make K integer |
|
|
Try a different tack |
|
See
Right scroll |
|
2,1 |
|
| So we sum
of squares is Z square where Z is |
|
5,1 |
|
| AMQ |
|
QSUW |
|
WPC |
|
ABC |
|
8,1 |
|
| 175 |
|
175 |
|
175 |
|
175 |
|
7,2 |
|
| 30625 |
|
30625 |
|
30625 |
|
30625 |
|
10,2 |
|
|
QSUW |
|
is directly adjustable |
on U |
|
to get the others closer together |
|
etc… |
|
|
Have to make th Z match on bottom
and Right |
|
| Started with
eyeballing the Diagram in Observer |
|
|
J is smallest |
|
Now if we examine Lower corner
correctly |
|
|
then X, I G |
|
Assume
X and U are unkown |
|
V looks about G + H = E |
|
S = U-X |
|
|
U is huge… |
|
R=S-X |
|
|
T=R-X |
|
| 3025 |
1521 |
6561 |
256 |
81 |
196 |
16 |
25 |
9 |
1 |
324 |
400 |
3136 |
1444 |
900 |
2601 |
4096 |
961 |
1089 |
841 |
1225 |
64 |
1849 |
4 |
|
V= U+X-T |
|
|
and we solved V |
|
| 30625 |
|
|
Curiously Guess X = 2
and play with U till QSUW = WPC finds U is 35 |
|
But X=1 then U doesn't
solve, |
16-17 |
|
|
Now maybe just need to tweak I, J
until AM and AB get big enough |
|
And X=
3 |
it is 52-53 |
|
|
|
But
at sheet 5 attempt we see that with J=1 I=8 AMQ gets too big |
|
X=4 => U= 70 |
and a pattern |
|
|
|
|
Looks like getting near a method -
for increasing X - find U then look for I,J to complete the size of top left. |
|
|
|
Can we set up equations that Back
solve the top rectangle ? - Yes - gaol seek it. |
|
|
|
A+M = C + O = Z - Q |
|
Z=length of outer side. |
|
|
A + B = Z - C |
|
|
M+N = Q + R + T - O |
|
Well yes we have it down to one
unknown hence goal seek will find the others |
|
| Set A to
satisfy… |
|
|
z = |
175 |
|
c= |
81 |
|
q= |
64 |
|
| 55 |
39 |
81 |
16 |
9 |
14 |
4 |
5 |
3 |
1 |
18 |
20 |
56 |
38 |
|
64 |
|
175 |
|
|
B=Z-C-A |
|
1 |
|
|
0 |
|
M=Z-Q-A |
|
|
D=A-B |
|
N=A+B-M |
|
|
K=M-n |
|
|
L=n-k |
|
|
f=A+M-b-n-l |
|
|
e=b-d-f |
|
|
h=f-e |
|
Goalseek
to zero |
|
|
g=e-h |
|
j=L-f-h |
|
By changing A |
|
|
Hmm also h-g |
|
|
i=g-j |
|
Fix some errors - look for
consistency in fwd and Back |
|
|
| Obviously
the solution is scaleable… |
|
X increasing in multiples of 2 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|