The ‘dydt’ definition should be:
dydt = [0; 0];
that would define a column vector. The differential equation runs correctrly with that change.
The problem is that you are fitting one column off data (‘f’), however your differential equation returns two columns. This throws:
Error using lsqcurvefit (line 269)
Function value and YDATA sizes are not equal.
That can be solved by using ‘t’ instead of ‘tspan’:
[T,y] = ode45(@MSolveFit,t,conds);
This works:
D = load('fittingdata_t_f.txt');
t = D(:,1);
f = D(:,2);
pguess = rand(4,1)*1E+6;
p = lsqcurvefit(@MSolve,pguess,t,f);
function y = MSolve(p,t)
cond1 = 0;
cond2 = 0;
conds = [cond1, cond2];
tspan = [t(1) t(end)];
% y(1) = P(t), y(2) = M(t)
[T,y] = ode45(@MSolveFit,t,conds);
y = y(:,2);
function dydt = MSolveFit(t,y)
nc = 2; % # monomers for primary nucleation
m0 = 10; % initial concentration of tau
mt = m0 - y(2);
% parameters :
% p(1) = kn
% p(2) = km
% p(3) = ka
% p(4) = kp
% dydt(1) = dPdt(t)
% dydt(2) = dMdt(T)
dydt = [0; 0];
dydt(1) = p(1).*mt.^nc + p(2).*y(2) - p(3).*y(1).^2;
dydt(2) = p(4).*mt.*y(1);
end
end
figure
plot(t, f, 'p')
hold on
plot(t, MSolve(p,t), '-r')
hold off
grid
However I cannot get a good fit to your data, probably because I have no idea what you are doing or what the initial parammeter estimates should be.
