# Mathematica:Having problem with my calculation(numerical)

1. Aug 14, 2010

### shafieza_garl

What wrong with my calculation.the problem is from the attachment.
It keeps saying

NMinimize::nnum:The function value .........is not a number at {x1,x2}=.....

given x0=0 and x3=5

#### Attached Files:

• ###### Untitled-2.jpg
File size:
21.7 KB
Views:
165
Last edited: Aug 14, 2010
2. Aug 14, 2010

### phyzguy

Did you tell it what Y (capital Y) is? If so, you didn't tell us. NMinimize needs to know all of the values, it can't deal with variables.

3. Aug 14, 2010

### Bill Simpson

In[1]:=x0=0;x3=5;
F[x_,y_]:=0.25(4044.44444444445-66.66666666666667x+E^(-0.015x+0.015y)(-4044.44444444445+66.66666666666667y))^2;
H[x_,y_]:=Integrate[0.25(E^(-0.015t)(E^(0.015t)(4044.44444444445-66.66666666666667t)+E^(0.015y)(-4044.44444444445+66.66666666666667y)))^2,{t,x,y}];
NMinimize[{27+3*(F[x0,x1]+F[x1,x2]+F[x2,x3])+0.2*(H[x0,x1]+H[x1,x2]+H[x2,x3]),x0≤x1&&x1≤x2&&x2≤x3},{x1,x2}]

Out[4]={4.457054464608864*10^8, {x1 -> 0.06690213249753463, x2 -> 1.1629357790698616}}

If I did not make an error typing this all back in again then I cannot explain what your problem is. But I believe you have larger problems than this. Quit kernel and restart

In[1]:=a=1/4;b=4044+4/9;c=66+2/3;d=15/1000;x0=0;x3=5;
F[x_,y_]:=a*(b-c*x+E^(-d*x+d*y)*(-b+c*y))^2;
H[x_,y_]:=Integrate[a*(E^(-d*t)*(E^(d*t)*(b-c*t)+ E^(a*y)*(-b+c*y)))^2,{t,x,y}];
NMinimize[{27+3*(F[x0,x1]+F[x1,x2]+F[x2,x3])+ 1/5*(H[x0,x1]+H[x1,x2]+H[x2,x3]),x0≤x1&&x1≤x2&&x2≤x3},{x1,x2}]

Out[4]={7.388367735069655*10^6, {x1 -> 2.4650125379662637, x2 -> 3.9306108222644514`}}

So very tiny changes in your coefficients give very different results.

I would suggest you spend a few days reading about the details of approximate and exact and machine precision mathematics as provided by Mathematica. Much of that is probably not what you would expect it to be.