best penalty value and mean penalty value

Hello everyone,

I need to interpret the solutions of a problem i have been working on. My problem is a maximitazion problem and i am using genetic algorithm. There is best penalty value and mean penalty value. I will share a screenshot of the result.

The best value is acceptable by me but I do not understand the difference between best and mean. I have read that it is about feasibility of the population. But i do not understand if the plot is the way that it should be or the infeasibility of the population members is very high?

If you can help me to interpret this plot i would be very appreciated.

Thank you in advance.

Answer 1

[Star Strider]

From the documentation:

'gaplotbestf' plots the best score value and mean score versus generation.

So in every generation (or interval defined by the 'PlotInterval' name-value pair), the 'gaplotbestf' option reports the best fitness value in that generation (or 'PlotInterval' value) and the mean of all the fitness values of that generation (or 'PlotInterval' value).

I believe that is what you have chosen to plot (and the plot I use to report the progress of ga during the optimisation). Since you are maximising the objective fuunction, the fitness values will be negative, as they should be in that situation.

.