Value Function Iteration Interpolation Error: Input Coordinates must be real

Hello, I've attached the most pertinent aspects of my code. The main issue I am having is lines 70-84 during the process of implementing a value function iteration for a specified objective function. I seem to run into a problem of when I specify my parameter psi=0.357 the interpolation procedure found in valfun_v2.m gives an error that the coordinates must be real. Any suggestions or reasons why would be greatly appreciated.

Thanks.

Answer 1

[Binaya]

Hi David,

I understand that the “interp2” function is receiving complex arguments during execution. This error can occur when you attempt to find the minimum point of 'valfun_v2' between kmin and kmax. As “fminbnd” sweeps through the values between kmin and kmax, it generates a complex value of k to test the function value.

According to the documentation of the “fminbnd” function, the function handle provided as an input argument should accept a real scalar and return a real scalar. However, in your case, the function handle of 'valfun_v2' is returning complex values as output. Since the “fminbnd” function is not designed to handle complex value arguments, your code encounters errors during execution.

You can resolve this issue by following the below troubleshooting steps:

• You can modify the “valfun_v2” function to return the absolute value of the complex number to convert the complex output to a real value suitable for “fminbnd”.
• You create a wrapper function that encapsulates “valfun_v2” and converts the complex output to a real value. This way, you can use “fminbnd” without modifying “valfun_v2” itself.

For more information on “fminbnd” function and its expected input format, please refer to the documentation provided here: Find minimum of single-variable function on fixed interval - MATLAB fminbnd - MathWorks India

I hope this helps.

Regards,

Binaya