Prev: imfill fuction source code in java
Next: lsqnonlin for 2D problems
From: kp pk on 5 Jul 2010 17:13
n=20;
t=0.1;
G=5;
syms x;
syms y;
for i=1:n
pl=(x^2-1)^i;
temp=diff(pl);
for m=1:i-1
temp=diff(temp);
end
f=factorial(i);
P(i)=temp/((2^i)*f);
end
for i=1:n
P(i)=subs(P(i),x,2*y-1);
P(i)=P(i)/sqrt(int(P(i)*P(i),'y',0,1));
end

syms c;
S=0;
for i=1:n
c(i)=sym(sprintf('c%d',i));
end
for i=1:n
S=S+P(i)*(c(i+1)-c(i))/t+G*c(i)*diff(P(i));
end
M=coeffs(expand(S),y);

This is the program im running.
M is now a array with 20 elements,each of which are equations in c1,c2,c3,..,c20
Now i want to equate each of these equations to 0 and solve for c1,c2,...c20.
can anyone suggest a way to do this?
thanks.
From: Walter Roberson on 5 Jul 2010 17:32
kp pk wrote:

> M is now a array with 20 elements,each of which are equations in
> c1,c2,c3,..,c20
> Now i want to equate each of these equations to 0 and solve for
> c1,c2,...c20.
> can anyone suggest a way to do this?

Possibly

solve(M,c)

I do not have the symbolic toolbox to cross-check this with.
From: kp pk on 5 Jul 2010 17:50
no walter,i already tried that.that doesnt work.
From: Walter Roberson on 5 Jul 2010 18:23
kp pk wrote:
> no walter,i already tried that.that doesnt work.

Hmmm... then the following might work:

Mc = num2cell(M);
cc = num2cell(c);
solve(Mc{:}, cc{:})

The documentation for num2cell is contradictory: it says in some places that
it works for numeric arrays but there is a specific line in the documentation
saying that it works for all array types.


If num2cell does not work then,

Mc = cell(numel(M),1);
cc = cell(numel(c),1);
for K = 1 : numel(M)
Mc{K} = M(K);
end
for K = 1 : numel(c)
cc{K} = c(K);
end
solve(Mc{:}, cc{:})
From: kp pk on 6 Jul 2010 12:55
Yeah,that works.
Thanks a lot ,Robert.
 |  Next  |  Last
Pages: 1 2 3
Prev: imfill fuction source code in java
Next: lsqnonlin for 2D problems