how-do-i-e​stimate-or​-optimize-​the-parame​ters-of-my​-ode-45 system of equations?

I have three constitutive equations containing variables which need to be optimised..

How to give the equality constraints in optimisation solver ... (LSQNONLIN, FMINCON} for the following system of equations...

since I am having only three equations with initial conditions..

Finally I want to optimise the variables with experimental data with minimum least squares..

I have attached the function file and executable file using ODE45, equations and experimental data...

Please help me to sort out this problem... Help will be appreciated...

[Alan Weiss]

Is this a school assignment?

Have you seen Fit an Ordinary Differential Equation (ODE)?

Alan Weiss

MATLAB mathematical toolbox documentation

[ruban ramalingam]

No sir... This is the inital step of my project work...

Yes.. I solved those ode equations using ODE45 solver.. I want to optimise the variables in such a way that I get the best least square fit with the experimental data...

I don't know how to give constraints in the optimisation toolbox for these equations for optimising the variables using different solver (interior point method, levenberg-Marquidt algorithm) and want to compare the least squares errors of the different method..

Can you tell me how to couple this equations with the optimisation toolbox and by giving the input variables in the optimisation toolbox by showing a pic?

This will be helpful for me..

Answer 1

[Alan Weiss]

Sorry, I don't understand what you said. What do you mean you "don't know how to give constraints in the optimisation toolbox for these equations for optimising the variables using different solver (interior point method, levenberg-Marquidt algorithm) and want to compare the least squares errors of the different method?"

• Do you mean you want to pass options to lsqcurvefit to use different algorithms?
• Do you mean that you have linear or nonlinear constraints that don't show in your problem setup so far?

For the first case, just select the 'Algorithm' option using optimoptions and pass the option to the solver. See the example on the function reference page. For the second case, you will need to use a different solver, fmincon, as explained on the optimization decision table.

Alan Weiss

MATLAB mathematical toolbox documentation

[ruban ramalingam]

Since the equations are of coupled ode equations with no constraints..

It asks for the constraints for getting solved in optimisation toolbox...

Please look at the equations and reply..

[Alan Weiss]

I have no idea what you mean "It asks for the constraints for getting solved in optimisation toolbox..." Sorry. If you are getting an error message, please show what you have input and the complete error message. If you mean something else, feel free to ask again.

Alan Weiss

MATLAB mathematical toolbox documentation