1 view (last 30 days)
Tsogerdene on 3 Aug 2011
I just wondering that why this code working wrong. Please help me! I can't manipulate the ODE function.
options=odeset('RelTol',10^-4);
[T,Y]=ode23('turshilt23',[0 300],[0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5],options);
plot(T,Y);
function dy=turshilt23(t,y,flag)
%here is some code
if(k==0)
y(1)=y(1)/2;
end
dy(1)=mu*y(1)*(1-(y(1)/m_max));
%here is some code
end

Arnaud Miege on 3 Aug 2011
Also, it seems that in your function turshilt23, you haven't defined what k, mu or m_max are. Finally you compute dy(1), but you call the ode solver with an initial condition vector of length 9. You need to compute dy(2), dy(3), ... dy(9) in your function as well.
Have a look at the documentation for the ode solvers, there are various examples you can inspire yourself from.
HTH,
Arnaud

Tsogerdene on 3 Aug 2011
function dy=turshilt23(t,y,flag)
global Global_CycB1;
global Global_CycB2;
global k;
global y_1;
dy=zeros(9,1);
% size of protein
mu=0.005;
m_max=10;
dy(1)=mu*y(1)*(1-(y(1)/m_max));
if Global_CycB2 >=0.099 & Global_CycB2<0.1
if Global_CycB2>Global_CycB1
y(1)=y(1)*(1/2);
end
end
%CycBT
k1=0.04;
k2_1=0.04;
k2_2=1;
k2_3=1;
dy(2)=k1-(k2_1+k2_2*y(3)+k2_3*y(5))*y(2);
Keq=1000;
beta=y(2)+y(7)+Keq^(-1);
Trimer=(2*y(2)*y(7))/(beta+(beta^2-4*y(2)*y(7))^(1/2));
CycB=y(2)-Trimer;
Global_CycB1=Global_CycB2;
Global_CycB2=CycB;
%Cdh1
k3_1=1;
k3_2=10;
k4_1=2;
k4=35;
J3=0.04;
J4=0.04;
dy(3)= (((k3_1+k3_2*y(5))*(1-y(3)))/(J3+1-y(3)))-(((k4*y(1)*CycB+k4_1*y(8))*y(3))/(J4+y(3)));
%Cdc20T
k5_1=0.005;
k5_2=0.2;
k6=0.1;
J5=0.3;
n=4;
dy(4)= k5_1+k5_2*((y(1)*CycB)^n/(J5^n+(y(1)*CycB)^n))-k6*y(4);
%Cdc20A
k6=0.1;
k7=1;
k8=0.5;
J7=1*(10^(-3));
J8=1*(10^(-3));
%IEP
k9=0.1;
k10=0.02;
dy(6)= k9*y(1)*CycB*(1-y(6))-k10*y(6);
%CKI
k11=1;
k12_1=0.2;
k12_2=50;
k12_3=100;
dy(7) = k11-(k12_1+k12_2*y(8) +k12_3*y(1)*CycB)*y(7);
%SK
k13_1=0;
k13_2=1;
k14=1;
dy(8)= k13_1+k13_2*y(9)-k14*y(8);
%TF
k15_1=1.5;
k15_2=0.05;
k16_1=1;
k16_2=3;
J15=0.01;
J16=0.01;
dy(9)= (((k15_1*y(1)+k15_2*y(8))*(1-y(9)))/(J15+1-y(9)))-(((k16_1+k16_2*y(1)*CycB)*y(9))/(J16+y(9)));

Arnaud Miege on 3 Aug 2011
It works for me. What error message do you get?
* Your function is a function of t, y and flag but t and flag are not used. Consider replacing those with ~
* Your function defines k and y_1 as global variables, but those aren't used in the function. You can remove those two lines.
The only changes I made were therefore:
function dy=turshilt23(~,y,~)
global Global_CycB1;
global Global_CycB2;
% global k;
% global y_1;
I used 0.1 for the other two global variables Global_CycB1 and Global_CycB2. The solver solved OK.
Walter Roberson on 3 Aug 2011
Note: Using ~ in that way is not available in 2008 and earlier.

Tsogerdene on 8 Aug 2011
Thank you for helping. I just asking that y(1)=y(1)*(1/2)? I would like to divide actual value of y(1) when some conditions are satisfied. I used y(1)=y(1)*(1/2), but nothing is changed.

Walter Roberson on 8 Aug 2011
When you modify y(1) in your code, you are only modifying the _local_ y(1) . It is not possible to permanently change the y for figure iterations.
Tsogerdene on 8 Aug 2011
Oh I see. Thank you very much!. I understood that.
Walter Roberson on 8 Aug 2011
Typo correction: I meant for "future iterations"