## speeding up code (with fsolve)...

Hi

I have to numerically solve this equation many (tens of thousands) of times for theta_n (as I vary different parameters, especially n):

tan(theta_n) - C_a*Z_0*(-v^2*(Pi*n-theta_n)^2/x_l^2+omega_a^2)*x_l/(v*(Pi*n-theta_n)) = 0

theta_n should be in [-pi/2, pi/2). It seems like solving this is the slowest single component of the chain of calculations (that follow this).

Currently I do it with this function:

get_theta_n_array:=proc(max_n::integer, omega_a::float, v::float, x_l::float, C_a::float, Z_0::float)
local theta_n_array:=Array(
select(x->Re(x)<evalf(Pi/2) and Re(x)>=-evalf(Pi/2), [seq(
#NOTE: careful with fsolve - in some cases returns unevaluated equation
fsolve(tan(theta_n) - C_a*Z_0*(-v^2*(Pi*n-theta_n)^2/x_l^2+omega_a^2)*x_l/(v*(Pi*n-theta_n)) = 0, theta_n) , n=1..max_n)]
, real)
,datatype=float[8]);
if ArrayNumElems(theta_n_array) <> max_n then
printf("Bad Array Dimensions! Got too many or not enough solutions.");
theta_n_array:="CHECK: get_theta_n_array()": #dirrrrty hack that will ring an alarm bell if array is not the right size
end;
theta_n_array;
end;

And call it like so (for say n=1000)

st:=time();
result:=get_theta_n_array(1000, 100e9, 1e8, 0.3, 20e-14, 50.0);
time()-st;

This will take say 3.5s on my PC.

Does anyone have any ideas how to speed this up? I would hope this to take at least an order of magnitude less time. I played with DirectSearch lib but that was not faster.

Also, I should note that this is the only portion of my code that is not thread safe (because of the fsolve call), which leads to "extra" slowdowns because I have to use Grid:-Map, instead of Thread:-Map when parallelizing, and more importantly because I can't compile the rest of the code (Grid:-Map is not compatible with compiled functions).

Let me know if you have any ideas...

thanks!

## problem in using fsolve when determine a range for...

i want to solve an equation by fsolve but i cant assign a value as an input for next step!

s := fsolve(G), x = -1 .. 1

s := .1449607418, x = -1 .. 1

a:=s+1

Error, invalid input: subs received .1449607418, which is not valid for its 1st argument

## I have problem to solve equation...

I want to solve system of equation but it has unknow parameter.

Then I test system of equation. It hasn't unknowparameter.

eq1 := x^2+y^2 = 4

eq2 := y-x^2 = 0

fsolve({eq1, eq2}, {x, y})

{x = -1.249621068, y = 1.561552813}

So I get answer by using fsolve.

Then I try to put unknow parameter in system of equation.

eq3 := x^2+ky^2 = 4

eq4 := ay-hx^2 = 0

fsolve({eq3, eq4}, {x, y})

Error, (in fsolve) {ay, hx, ky} are in the equation, and are not solved for

## Fsolve and Isolate Missing Solutions...

To motivate some ideas in my research, I've been looking at the expected number of real roots of random polynomials (and their derivatives).  In doing so I have noticed an issue/bug with fsolve and RootFinding[Isolate].  One of the polynomials I came upon was

f(x) = -32829/50000-(9277/50000)*x-(37251/20000)*x^2-(6101/6250)*x^3-(47777/20000)*x^4+(291213/50000)*x^5.

We know that f(x) has at least 1 real root and, in fact, graphing shows that f(x) has exactly 1 real root (~1.018).  However, fsolve(f) and Isolate(f) both return no real roots.  On the other hand, Isolate(f,method=RC) correctly returns the root near 1.018.  I know that fsolve's details page says "It may not return all roots for exceptionally ill-conditioned polynomials", though this system does not seem especially ill-conditioned.  Moreover, Isolate's help page says confidently "All significant digits returned by the program are correct, and unlike purely numerical methods no roots are ever lost, although repeated roots are discarded" which is clearly not the case here.  It also seems interesting that the RealSolving package used by Isolate(f,method=RS) (default method) misses the root while the RegularChains package used by Isolate(f,method=RC) correctly finds the root.

All-in-all, I am not sure what to make of this.  Is this an issue which has been fixed in more recent incarnations of fsolve or Isolate?  Is this a persistent problem?  Is there a theoretical reason why the root is being missed, particularly for Isolate?

Any help or insight would be greatly appreciated.

## Solving a system of equations...

Hi! I'm trying to solve a system of four non-linear equations in Maple 17 but it doesn't work.

Equations are: F, Fw, Ft, Fk and varibles are T,w,k,ki.

Parametars are Mn=10, Ms=2, alfa=0.2 and Tf=(k*T^(alfa)/Mn)^(1/alfa).

Solutions must be positive.

This is maple script:

restart;
> assume (w>0):
> assume (T>0):
> assume (k>0):
> assume (ki>0):
> assume (Mn>0):
> assume (Ms>0):
> assume (Tf>0):
> assume (alfa>0):

> Gp:=1/exp(sqrt(w*I));

> Cf:=((T*w*I+1)/(Tf*w*I+1));

> Cpi:=(k*w*I+ki)/(w*I);

> L:=Cf*Cpi*Gp;
> L:=evalc(L):
> F:=subs(Ms=2,Tf=(k*T^(alfa)/Mn)^(1/alfa),evalc(abs(1+L)^2-1/Ms^2));
> F:=subs(Mn=10,alfa=0.2,evalc(F));

> Fw:=diff(F,w):
> Fk:=diff(F,k):
> Ft:=diff(F,T):
> fsolve({F,Fw,Fk,Ft},{w,k,ki,T});

Thanks in advance for any help. Dragoslav

## problem with solving equation, I couldn't find any...

i := I;

assume(a>0);

assume(b>0);

assume(r>0);

assume(k>0);

assume(w>0);

Hz := k*(z^2-2*r*a+r^2)/((z-1)*(z^2-2*b*z+1));

Hzw := eval(Hz, z = exp(i*w)); assume(a > 0);

Habs := simplify(abs(Hzw)^2);

p1 := eval(Habs, w = Pi) = (10^((-3.3018)*(1/20)))^2;

p2 := eval(Habs, w = (1/2)*Pi) = (10^((-.1758)*(1/20)))^2;

p3 := eval(Habs, w = (1/4)*Pi) = (10^(6.425*(1/20)))^2;

p4 := eval(Habs, w = (1/8)*Pi) = (10^(54.5699*(1/20)))^2;

solve({p1, p2, p3, p4})

## Problems in very high! CPU time when I use DirectS...

Dear Friends

I have a problem in CPU time in MAPLE.

I write the codes in maple related to the nonlinear heat conduction problem in one dimension by Collocation method, but after 30 minutes no solution has been observed!!!

My codes are for N=4!, i.e., I have 25 equations with 25 unknowns!!!

If MAPLE can not solve this simple system, How can I solve 3 dimensional pdes by N=9,

In this case, I have 1000 equations with 1000 unknowns!!!

I should remark that my problem is stated in this paper

http://www.sciencedirect.com/science/article/pii/S1018364713000025

If there exist any other suitable method, I will be happy to receive any support.

With kind regards,

Emran Tohidi.

> restart;
> Digits := 20; N := 4; st := time(); u := sum(sum(a[m, n]*x^m*t^n, m = 0 .. N), n = 0 .. N); u := unapply(u, x, t); ut := diff(u(x, t), `\$`(t, 1)); ut := unapply(ut, x, t); ku := simplify(1+u(x, t)^2); ku := unapply(ku, x, t); ux := diff(u(x, t), `\$`(x, 1)); ux := unapply(ux, x, t); K := ku(x, t)*ux(x, t); K := unapply(K, x, t); Kx := diff(K(x, t), `\$`(x, 1)); Kx := unapply(Kx, x, t); f := proc (x, t) options operator, arrow; x*exp(t)*(1-2*exp(2*t)) end proc;
print(`output redirected...`); # input placeholder
> S1 := {seq(u(i/N, 0)-i/N = 0, i = 0 .. N)}; S2 := {seq(u(0, j/N) = 0, j = 1 .. N)}; S3 := {seq(u(1, j/N)+ux(1, j/N)-2*exp(j/N) = 0, j = 1 .. N)}; S4 := {seq(seq(Kx(i/N, j/N)+f(i/N, j/N)-ut(i/N, j/N) = 0, i = 1 .. N-1), j = 1 .. N)}; S := `union`(`union`(`union`(S1, S2), S3), S4); sol := DirectSearch:-SolveEquations([op(S)], tolerances = 10^(-4), evaluationlimit = 1000000);
print(`output redirected...`); # input placeholder
> assign(sol);
%;
> u(x, t);
> CPUTIME := time()-st;
plot3d(u(x, t) - x exp(t), x = 0 .. 1, t = 0 .. 1)

## how to solve this?...

Hy all.

I want to solve this equation, with„dd” as numerical result. What do I do wrong? Thanks. Nico

restart;
TTot := 70;
TC := 17;
GM := .26;
QMax := 870;
V := 3600*GM*QMax*TTot;
eq := V = int(QMax*exp((-t+TC)/dd)*(1+(t-TC)/TC)^(TC/dd), t = 0 .. TTot);
fsolve(eq, dd);

## fsolve cannot solve this equations and return them...

hello

this is my program and fsolve for low intensity solve the equations but for high intensity cannot solve why?

this is my code:

ep0 := 1/(4*3.14);

el := 8.54*10^(-2);

hbar := 1;

vf := 1/300;

kb := 1;

tem := 2.586*10^(-2);

ci := 1;

p := 1.458*10^16;

beta := 2;

ai := 7.1*10^(-4);

bi := ai/sqrt(3);

enph := .196;

d := enph/(kb*tem);

n0 := 1/(exp(enph/(kb*tem))-1);

gama := hbar*vf;

intensity:=9000000

w := 7.28;

impurity := 7.2*10^3;

g := hbar*beta/(bi^2*sqrt(2*p*enph));

aa := g^2*(n0+1)/(2*Pi*hbar*gama^2);

bb := g^2*n0/(2*Pi*hbar*gama^2);

cc := 2/(Pi*gama^2);

l := (1*hbar)*w/(2*kb*tem);

u := el^2*intensity/(32*w*hbar^2);

[fsolve({op([((enph*ln(1+exp(c+enph/(kb*tem)))/(kb*tem)-polylog(2, -exp(c))+polylog(2, -exp(c+enph/(kb*tem))))*enph*(kb*tem)^2-(enph^2*ln(1+exp(c+enph/(kb*tem)))/(kb^2*tem^2)+2*enph*polylog(2, -exp(c+enph/(kb*tem)))/(kb*tem)+2*polylog(3, -exp(c))-2*polylog(3, -exp(c+enph/(kb*tem))))*(kb*tem)^3+(-exp(b)*enph*ln(1+exp(c+enph/(kb*tem)))+exp(c+d)*enph*ln(1+exp(b-d+enph/(kb*tem)))+exp(b)*kb*tem*polylog(2, -exp(c))-exp(c+d)*kb*tem*polylog(2, -exp(b-d))-exp(b)*kb*tem*polylog(2, -exp(c+enph/(kb*tem)))+exp(c+d)*kb*tem*polylog(2, -exp(b-d+enph/(kb*tem))))*enph*(kb*tem)^2/((exp(b)-exp(c+d))*kb*tem)+(exp(b)*enph^2*ln(1+exp(c+enph/(kb*tem)))-exp(c+d)*enph^2*ln(1+exp(b-d+enph/(kb*tem)))+2*exp(b)*enph*kb*tem*polylog(2, -exp(c+enph/(kb*tem)))-2*exp(c+d)*enph*kb*tem*polylog(2, -exp(b-d+enph/(kb*tem)))+2*exp(b)*kb^2*tem^2*polylog(3, -exp(c))-2*exp(c+d)*kb^2*tem^2*polylog(3, -exp(b-d))-2*exp(b)*kb^2*tem^2*polylog(3, -exp(c+enph/(kb*tem)))+2*exp(c+d)*kb^2*tem^2*polylog(3, -exp(b-d+enph/(kb*tem))))*(kb*tem)^3/((exp(b)-exp(c+d))*kb^2*tem^2))*bb+u*(1/(1+exp(-l-c))-1/((1+exp(-l-c))*(1+exp(l-b))))-(((1*enph)*(enph-2*kb*tem*ln(1+exp(-b+enph/(kb*tem))))/(2*kb^2*tem^2)+2*kb^2*tem^2*(-polylog(2, -exp(-b+enph/(kb*tem)))+polylog(2, -cosh(b)+sinh(b))))*enph*(kb*tem)^2-(enph^2*(enph-3*kb*tem*ln(1+exp(-b+enph/(kb*tem))))-6*kb^2*tem^2*(enph*polylog(2, -exp(-b+enph/(kb*tem)))+kb*tem*(-polylog(3, -exp(-b+enph/(kb*tem)))+polylog(3, -cosh(b)+sinh(b)))))*(kb*tem)^3/(3*kb^3*tem^3)-(-exp(b)*enph^2+exp(c+d)*enph^2-2*exp(c+d)*enph*kb*tem*ln(1+exp(-b+enph/(kb*tem)))+2*exp(b)*enph*kb*tem*ln(1+exp(-c-d+enph/(kb*tem)))+2*exp(c+d)*kb^2*tem^2*polylog(2, -exp(-b))-2*exp(b)*kb^2*tem^2*polylog(2, -exp(-c-d))-2*exp(c+d)*kb^2*tem^2*polylog(2, -exp(-b+enph/(kb*tem)))+2*exp(b)*kb^2*tem^2*polylog(2, -exp(-c-d+enph/(kb*tem))))*enph*(kb*tem)^2/((2*(-exp(b)+exp(c+d)))*kb^2*tem^2)-(exp(b)*enph^3-exp(c+d)*enph^3+3*exp(c+d)*enph^2*kb*tem*ln(1+exp(-b+enph/(kb*tem)))-3*exp(b)*enph^2*kb*tem*ln(1+exp(-c-d+enph/(kb*tem)))+6*exp(c+d)*enph*kb^2*tem^2*polylog(2, -exp(-b+enph/(kb*tem)))-6*exp(b)*enph*kb^2*tem^2*polylog(2, -exp(-c-d+enph/(kb*tem)))+6*exp(c+d)*kb^3*tem^3*polylog(3, -exp(-b))-6*exp(b)*kb^3*tem^3*polylog(3, -exp(-c-d))-6*exp(c+d)*kb^3*tem^3*polylog(3, -exp(-b+enph/(kb*tem)))+6*exp(b)*kb^3*tem^3*polylog(3, -exp(-c-d+enph/(kb*tem))))*(kb*tem)^3/((3*(-exp(b)+exp(c+d)))*kb^3*tem^3))*aa-u*(1/(1+exp(l-b))-1/((1+exp(-l-c))*(1+exp(l-b)))) = 0, -cc*polylog(2, -exp(b))+cc*polylog(2, -exp(-c))-impurity = 0])}, {op([b, c])})];

thank you.

## A problem has been occured in fsolve command...

Dear All

I have a question about applying fsolve in MAPLE.

These codes come from an applicable model from heat transfer.

I previously solve such systems, but the fsolve takes long time, for example for N=2, I have 12 equations with 12 unknowns. But the fsolve does not work!

I try by N=3 or N=4, but a similar result has been obtained.

If kindly is possible, please give me a help for using fsolve more efficiently that solve my problem.

I have MAPLE 13 in my PC.

With kind regards,

Emran Tohidi.

> restart;
> with(orthopoly);
print(`output redirected...`); # input placeholder
> Digits := 20;
> N := 3; f := proc (x) options operator, arrow; cos(Pi*(x-1/2)) end proc; h1 := proc (t) options operator, arrow; 0 end proc; h2 := proc (t) options operator, arrow; 0 end proc; E := proc (t) options operator, arrow; cos(.2*Pi)*exp(-t^2) end proc;
print(`output redirected...`); # input placeholder
> u := sum(sum(b[m, n]*P(m, 2*x-1)*P(n, 2*t-1), m = 0 .. N), n = 0 .. N);
print(`output redirected...`); # input placeholder
> u := unapply(u, x, t);
print(`output redirected...`); # input placeholder
> ut := diff(u(x, t), `\$`(t, 1)); ut := unapply(ut, x, t);
print(`output redirected...`); # input placeholder
> uxx := diff(u(x, t), `\$`(x, 2)); uxx := unapply(uxx, x, t);
print(`output redirected...`); # input placeholder
> a := sum(c[m]*P(m, 2*t-1), m = 0 .. N); a := unapply(a, t);
print(`output redirected...`); # input placeholder
> A := fsolve(P(N+1, 2*x-1) = 0);
print(`output redirected...`); # input placeholder
> S1 := {seq(seq(ut(A[i], A[j])-a(A[j])*uxx(A[i], A[j]) = 0, i = 2 .. N), j = 2 .. N+1)}; S2 := {seq(u(A[i], 0)-f(A[i]) = 0, i = 2 .. N)}; S3 := {seq(u(0, A[j])-h1(A[j]) = 0, j = 1 .. N+1)}; S4 := {seq(u(1, A[j])-h2(A[j]) = 0, j = 1 .. N+1)}; S5 := {seq(u(.3, A[j])-E(A[j]) = 0, j = 1 .. N+1)}; S := `union`(`union`(`union`(`union`(S1, S2), S3), S4), S5);
print(`output redirected...`); # input placeholder
> sol := fsolve(S);

## Problem in using Fsolve command: It takes a long t...

Dear Users

I have a problem for solving a system of linear equations that arise from collocation method for getting approximate solution of a coupled PDE and ODE in Food engineering problems.

When it reach to the fsolve command it takes long time!!!

I used maple 13.

With kind regards,

Emran Tohidi.

> Restart;
print(`output redirected...`); # input placeholder
> h := 50; hm := 0.1e-3; rhodp := 1500; Y := 0.5e-1; T0 := 20; rhoair := 1.2041; Dair := 0.2e-8; DD := 0.85e-9; C := 3240; L := 0.4e-1; X0 := 1.5; V := .2; delta := 0.2e-2; Yair := 0.5e-1; nu := .2; Tair := 60; Hnu := 2400; rho := 1359; tt := 3;
%;
> N := 5; Digits := 20;
> X := sum(sum(a[m, n]*z^m*t^n, m = 0 .. N), n = 0 .. N); X := unapply(X, z, t); Xt := diff(X(z, t), `\$`(t, 1)); Xt := unapply(Xt, z, t); Xz := diff(X(z, t), `\$`(z, 1)); Xz := unapply(Xz, z, t); Xzz := diff(X(z, t), `\$`(z, 2)); Xzz := unapply(Xzz, z, t); T := sum(b[n]*t^n, n = 0 .. (N+1)^2-1); T := unapply(T, t); Tt := diff(T(t), `\$`(t, 1)); Tt := unapply(Tt, t); aw := exp(.914)*X(z, t)^.5639-.5*exp(1.828)*X(z, t)^(2*.5639); aw := unapply(aw, z, t); TT := 8.3036+3816.44*(1+T(t)/(46.13)+T(t)^2/46.13^2)/(46.13); TT := unapply(TT, t); pwv := 133.3*(1+TT(t)+(1/2)*TT(t)^2); pwv := unapply(pwv, t); Yi := .622*pwv(t)*aw(z, t)*(1+pwv(t)*aw(z, t)/rho+(pwv(t)*aw(z, t)/rho)^2)/rho; Yi := unapply(Yi, z, t);
%;
> S1 := {seq(seq(Xt(delta*i/N, tt*j/N)-DD*Xzz(delta*i/N, tt*j/N) = 0, i = 1 .. N-1), j = 1 .. N)};
> S2 := {seq(DD*rhodp*Xz(delta, tt*j/N)+hm*rhoair*Yi(delta, tt*j/N) = 0, j = 0 .. N)};
> S3 := {seq(Xz(0, tt*j/N) = 0, j = 0 .. N)};
> S4 := {seq(X(delta*i/N, 0) = 0, i = 1 .. N-1)};
> S5 := {seq(seq(rho*delta*C*Tt(tt*j/N)-h*(Tair-T(tt*j/N))+hm*Hnu*rhoair*(Yair-Yi(delta*i/N, tt*j/N)) = 0, j = 1 .. N), i = 0 .. N)};
print(`output redirected...`); # input placeholder
> S6 := {seq(rho*delta*C*Tt(0)-h*(Tair-T0)+hm*Hnu*rhoair*(Yair-Yi(delta*i/N, 0)) = 0, i = 0 .. N)};
%;
> SS := `union`(`union`(`union`(`union`(`union`(S1, S2), S3), S4), S5), S6);
> sol := fsolve(SS);

## I want to solve a system of equations by fsolve...

Dear All

the fsolve command doesnt work!

> restart;
> with(LinearAlgebra); Digits := 20;
print(`output redirected...`); # input placeholder
> Z := Matrix(2); N := 3; y11 := sum(a[n]*x^n, n = 0 .. N); y12 := sum(b[n]*x^n, n = 0 .. N); y21 := sum(c[n]*x^n, n = 0 .. N); y22 := sum(d[n]*x^n, n = 0 .. N); y11 := unapply(y11, x); y12 := unapply(y12, x); y21 := unapply(y21, x); y22 := unapply(y22, x);
print(`output redirected...`); # input placeholder
> A := linalg[matrix](2, 2, [1, -1, 1, exp(x)]); B := linalg[matrix](2, 2, [-3*exp(-x)-1, 2-2*exp(-x), -3*exp(-x)-2, 1-2*cosh(x)]);
print(`output redirected...`); # input placeholder
> C := eval(B, [x = 1]);
print(`output redirected...`); # input placeholder
> Y := linalg[matrix](2, 2, [y11(x), y12(x), y21(x), y22(x)]);
print(`output redirected...`); # input placeholder
> yy11 := diff(y11(x), x); yy12 := diff(y12(x), x); yy21 := diff(y21(x), x); yy22 := diff(y22(x), x); yy11 := unapply(yy11, x); yy12 := unapply(yy12, x); yy21 := unapply(yy21, x); yy22 := unapply(yy22, x);
> YY := linalg[matrix](2, 2, [yy11(x), yy12(x), yy21(x), yy22(x)]);
print(`output redirected...`); # input placeholder
> S := {seq(eval(YY, [x = n/N])-(eval(A, [x = n/N]))*(eval(Y, [x = n/N]))-(eval(B, [x = n/N])) = Matrix(2), n = 1 .. N)};
print(`output redirected...`); # input placeholder
> S1 := {eval(Y, [x = 0]) = linalg[matrix](2, 2, [3, 0, 1, 1])};
print(??); # input placeholder
> SS := `union`(S, S1);
print(??); # input placeholder
> sol := fsolve(SS);

## solve with assumptions - is this a bug?...

I am trying to solve these equations:

eq1:=cot(theta[n])=(omega[a]^2 - omega[n]^2)/omega[n];

eq2:=cot(theta[n]+ omega[n])=-(omega[b]^2 - omega[n]^2)/omega[n];

for some values of omega[a] and omega[b]. I will in the end create two 3d plots for omega[n] and theta[n] as a funciton of ometga[a] and omega[b]

So for example for some particular values have:

eqs:=subs(omega[a]=0.2, omega[b]=1, [eq1, eq2]);

Then solving,

solve(eqs, [omega[n],theta[n]]);

gives:

I'm however only interested in solutions where omega[n]>=0, but doing this:

solve(eqs, [omega[n],theta[n]], UseAssumptions) assuming omega[n]>=0.0;

returns an empty list. Adding the condition omega[n]>=0.0 to the list of equations also does not work.

Is this a bug or am I missing something? Note, I realize that i can manually go through the entries myself and pick the right solutions, I am just asking how to force maple to do this automagically.

thanks

Here is the full code:

restart:

eq1:=cot(theta[n])=(omega[a]^2 - omega[n]^2)/omega[n];

eq2:=cot(theta[n]+ omega[n])=-(omega[b]^2 - omega[n]^2)/omega[n];

eqs:=subs(omega[a]=0.2, omega[b]=1, [eq1, eq2]);

solve(eqs, [omega[n],theta[n]]);

solve(eqs, [omega[n],theta[n]], UseAssumptions) assuming omega[n]>=0.0;

## concerning the identify command / master theorem...

Greetings to all.

At the following Math.Stackexchange Discussion a certain constant was computed in relation to a Master Theorem Type recurrence being solved. This prompted me to try to identify it by the use of the eponymous command. What follows is the content of the Maple session. You may want to read the post in order to get an understanding of what the constant means and how its exact value is calculated.

```> fsolve(2/2^a+1/4^a=1, a);
1.271553303

> identify(%);
1.271553303

> identify(%,all);
1/2    1/2
2 2      3
arcsech(------ + ---- - 1/6 Zeta(5))
7       6

> evalf(log[2](1+sqrt(2)));
1.271553303
```

My questions/observations are:

• Why does the algorithm fail to spot as simple a constant as the one above or am I just not invoking it correctly?
• If that last formula were true the author of this code would certainly win a prize for calculating a closed form expression for an odd integer zeta function value! (To be fair here I did notice that the spurious identification disappears when the number of working digits is increased and I do understand that the identification depends critically on the number of digits.)

In concluding I would like to say, why the complicated formula and not the simple one? Let me congratulate you just the same on providing this very useful command. I have worked on pandigital approximations which are slightly related and I understand that adding an operation like the logarithm to an integer base up to some max base value can dramatically increase the search space and may not always be feasible.

Best regards,

Marko Riedel

 5 6 7 8 9 10 11 Page 7 of 13
﻿