vectorize the genetic algorithm optimization

Hi,

I am using the genetic algorithm (GA) to fit the kinetic parameters of an ode system. Using the default settings, I could not finish one iteration in 3 days on a MacPro laptop. Since this is the first time that I code a fitting program and use the GA, I am not sure if this is normal. I want to try to vectorize my fitting objective to speed up the process. Does anybody know how to vectorize the ode part of the process? My fitting objective function is attached.

Thank you for your time. Any suggestion is appreciated.

---- fitting objective function starts

function error = IL6_STAT3_fittness(param)

   global lags tspan
   global time_obs pSTAT3_obs yy0

   sol = dde23(@(t,x,lags,param) IL6_STAT3_reduced(t,x,lags,param), lags, yy0, tspan, [], param);
   pSTAT3_sim = sol.y(6,:) + sol.y(7,:);

   % interpolate dde23 output at observed timepoints
   pSTAT3_int = interp1(sol.x, pSTAT3_sim, time_obs); % simulated data

   error = sum((pSTAT3_int - pSTAT3_obs).^2);

---- fitting objective function ends

Thanks, Wendy

Answer 1

[Alan Weiss]

I answered in this thread.

Alan Weiss

MATLAB mathematical toolbox documentation