Question about the best fitness function value in genetic algorithm

I am using the genetic algorithm from Global Optimization Toolbox to sovle a non-smooth optimization problem. The level of display is set to be 'iter', and I also add "@gaplotbestf" in the plot options.

However, the best fitness function value shown in the display is different from the one shown in the @gaplotbestf plot.

For example, the best fitness function value shown in the display is -0.00739942, and the best fitness function value shown in @gaplotbestf plot is -20.3323.

I have no clue why this happens and which number I should trust. Thanks!

[Walter Roberson]

I wonder if gaplotbestf takes into account locations that turn out to be rejected due to constraints?

[Alan Weiss]

Do you have nonlinear constraints? Please show us your call to ga, including all options.

Alan Weiss

MATLAB mathematical toolbox documentation