problem in finding inverse

%function[el5,f0]=data_cholesky(sigma,nel,l)
clc
clear all
sigma=0.15;
  
    l=300;

 nel=200;	% number of elements 
nnel=2;	% number of nodes per element 
ndof=1;	% number of dofs per node 
nnode=(nnel-1)*nel+1;	% total number of nodes in system 
sdof=nnode*ndof;	% total system dofs 
index=zeros(nnel*ndof,1); 
mm=zeros(sdof, sdof);
bcdof(1)=1;	% first dof (deflection at left end) is constrained 
bcval(1)=0;	% whose described value is 0 
bcdof(2)=2; % 12th dof (slope at the symmetric end) is constrained 
bcval(2)=0;	% whose described value is 0
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
T = zeros(nel+1, 1); % Replace with your initial guess for x
i=1; %boundary condition
T(i) = 300;    
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
keq=zeros(nel+1,nel+1,l);
keq1=zeros(nel+1,1,l);
keq2=zeros(nel+1,nel,l);
keq3=zeros(nel,nel,l);
A1 =zeros(nel+1,1,l);
A2 =zeros(nel,1,l);
el40 = zeros(nel, l);  % Initialize el4 as a 2D matrix
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
a=nel;
i=1:a;
b0=11;
mean0=zeros(3,b0);
vari0=zeros(3,b0);
kk(:,:,:)=zeros(sdof, sdof,l);

ss(:,:)=zeros(sdof, sdof);
m=zeros(2,2,l);
s=zeros(2,1,l);
k0 = zeros(2,1,l);
el0=zeros(nel,b0,l);
%for b0=1:b0
b=b0-1;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ka =  200 ; % conductivity
tleng=50*10^-3; % total length
d =1*10^-3;
le=tleng/(nel);

k0 = (ka/le)*[1 -1;-1 1];%
P =3.14*d ; % perimeter;
h = 20; % convective heat transfer
Ac = 3.14*d^2/4; % cross area crosssection
q0 = 0;
m = (P * h * le) / (6 * Ac) * [2 1; 1 2];%
Tinf=30;
s = (q0 + (P*h/Ac)*Tinf)*[le*0.5;le*0.5]; %
s=diag(s);
s = repmat(s, [1, 1, l]);
m = repmat(m, [1, 1, l]);
%k0=repmat(k0, [1, 1, l]);

Q=zeros(nel+1,1);
Q(1)=1;
Q(nel+1)=0;








for iel=1:nel	% loop for the total number of elements 
    edof = nnel*ndof; 
start = (iel-1)*(nnel-1)*ndof; 
% 
for i=1:edof 
index(i)=start+i; 
end 







edof = length (index);
for i=1:edof 
ii=index(i) ;
for j=1:edof 
jj=index(j) ;


mm(ii,jj)=mm(ii,jj)+m(i,j,l);
ss(ii,jj)=ss(ii,jj)+s(i,j,l);
end 

end

end





for l=1:l
   

d1=tleng/nel;
d=1;

i=1:nel;
j=1:nel;
ee=zeros(nel,nel);
R=zeros(nel,nel);
e(i)=tleng/2*nel+(tleng/nel)*(i-1);
for i=1:nel
for j=1:nel
ee(i,j)=e(i)-e(j);
R(i,j)=sigma^2*exp(-abs(ee(i,j)/d));
end
end
C1=chol(R,'lower');
Z1=randn(nel,1);
alpha=C1*(Z1);
E0=1;
el4= ((1+alpha));
 el40(:, l)= el4;
el=el4(1:nel);

el2= 1 + zeros( 1, 100 );
el3=[0.90991  0.98608 0.90314 1.1458 1.0554 0.91364 1.3918 0.61164 0.74829 1.369];
el5=E0*ones(1,20);
for iel=1:nel	% loop for the total number of elements 
    edof = nnel*ndof; 
start = (iel-1)*(nnel-1)*ndof; 
% 
for i=1:edof 
index(i)=start+i; 
end 



c0(iel)=el4(iel);
k(:,:,iel)=c0(iel)*k0;

edof = length (index);
for i=1:edof 
ii=index(i) ;
for j=1:edof 
jj=index(j) ;

kk(ii,jj,l)=kk(ii,jj,l)+k(i,j,iel);
end 

end



end
sss=diag(ss);
keq=kk+mm;
keq1=T(1)*keq(:,1);

keq2 =keq(:,2:end,:);
keq3=keq2(2:end,:,:);
A1(:,:,l)=sss+Q-keq1;
A2(:,:,l)=A1(2:end,:,l);
%A3(:,:,l)=inv(keq3(:,:,l));
%temp(:,:,l)=A3(:,:,l)*A2(:,:,l);

temp(:,:,l)=keq3(:,:,l)\A2(:,:,l);




 


 


end
% Define the length and number of elements

temp1=squeeze(temp);
T0 =300 * ones(1,l);
temp1=[T0;temp1];
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%temp1([2:4],:)=[]

x= linspace(0,tleng,nel+1);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%55
x=x';
for i=1:l
    plot(x,temp1(:,i))
    hold on
end

 mean_temp =mean(temp1,2)   ;
 mean_temp(1:3)

ans = 3×1

300.0000 341.6619 340.4453

the variation of temp vs x should be continously dectreasing. but when i run it several times .sometimes temp increses with x please help

[Image Analyst]

I have no idea what this weakly commented code does. Try adding more comments to explain the various things. Your subject says "problem in finding inverse" so I guess you want to find the inverse of a matrix. About all I can say is that if you want the inverse of a matrix, try inv

help inv

 INV    Matrix inverse.
    INV(X) is the inverse of the square matrix X.
    A warning message is printed if X is badly scaled or
    nearly singular.
 
    See also PINV, COND, CONDEST, LSQMINNORM, LSQNONNEG, LSCOV.

    Documentation for inv
       doc inv

    Other uses of inv

       codistributed/inv    InputOutputModel/inv    sym/inv
       gpuArray/inv         quantumCircuit/inv      symbolic/inv