Is This Optimization Objective Function Feasible?

Hi I have been using fmincon for constrained optimization with no
problems until I increased the complexity of the objective function.
At this moment the objective function's algorithm is as follows with
the only 2 variables to optimized being k1 and k2. No nonlinear
constraints and only some reasonable bounds:

t=1:100;
a = k1*exp(-k2*t);
b = some curve
c = conv(a,b);

for i=1:180 (representing degrees/time)
add c(i) value to a region on images{i} % images predefined
radon transform on images{i} at angle i
end

reconstruct from 180 columns of radon data

% cost function
f = sum(sum((wanted image - reconstructed image).^2));

the result is that the directional derivative starts off at 0 and the
optimization stops without adjusting any parameters. I have tried
with similar algorithms in which many calculations are done one after
the other, all depending on 2 optimization parameters, but with no
reconstruction. They were fine and did not begin with direction
derivative being 0. What might the problem be? Thanks in advance.
.

Relevant Pages

  • Re: Optimal Power Flow with fmincon
    ... I don't know anything about OPF, but I have used fmincon a fair amount. ... Depending on what my objective function looked like, I sometimes had to change some of the default optimization parameters. ... The other thought that I have is that if the different variables of the objective function are scaled very differently you might have problems. ...
    (comp.soft-sys.matlab)
  • Re: Optimization using a *Bounded* Levenberg-Marquardt method
    ... The problem is that I don't understand how I must modify my algorithm to add ... penalty to the value you compute for the objective function. ... This should make the optimization algorithm avoid the ... triangle into the upper half plane, ...
    (sci.math.num-analysis)
  • Re: problem with fminunc: roots containing NaN or Inf
    ... Input to ROOTS must not contain NaN or Inf. ... Sometimes the optimization procedure will last only 2 iterations, other times it can last for>100 iterations. ... the problem is your objective function is returning a illegal values at some points. ... For more information on which solver to use, and on writing constraints, see ...
    (comp.soft-sys.matlab)
  • Re: optimization problem
    ... that fmincon as well as ga cannot solve an optimization problem for a given proprietary function. ... I checked, whether the objective function really works, meaning, whether it really produces different output when changing input parameters: ... This is one of the best ways to debug an optimization, ... that I have to include the constraint sum w_i =1 and other constraints properly. ...
    (comp.soft-sys.matlab)
  • Re: optimization problem
    ... that fmincon as well as ga cannot solve an optimization problem for a given proprietary function. ... I checked, whether the objective function really works, meaning, whether it really produces different output when changing input parameters: ... This is one of the best ways to debug an optimization, ... Using ga, I further have not the problem of differentiability or chosing the right starting value, since it does the latter automatically. ...
    (comp.soft-sys.matlab)