I am beginner at the area of optimization. Normally, I download the GA, PSO codes from internet and then I modify them according to my problem and run it on Matlab.
In binary GA, most of the codes available in the internet does not mention about the variable limits, or the codes are given for only one or two variables. But, when we want to set multiple variable limits like upper and lower bounds of 10-20 variables. How to do it and how I should modify the basic crossover, mutation function for large number of variables. Any suggestion or help will be appreciated.
Dear Rasel,
Have a look at https://www.mathworks.com/help/gads/genetic-algorithm-options.html
https://www.mathworks.com/matlabcentral/answers/411298-genetic-algorithm-lower-and-upper-bounds
Try making up a fitness function that is a function of the bounds as well as a function of what you want to optimise.
Suppose A is between 1 and 10, B is between 1 and 20 and C is between 0 and 100 and you want to A and B to be big and C to be small.
LT means Less Than, GT means Greater Than
FITNESS FUNCTION(A,B,C)
IF A is LT 1 OR A GT 10 THEN
fitness = 0
GOTO LabelEndofBounds
ENDIF
IF B is LT 1 OR B GT 20 THEN
fitness = 0
GOTO LabelEndofBounds
ENDIF
IF C is LT 0 1 OR C GT 100 THEN
fitness = 0
GOTO LabelEndofBounds
ENDIF
****Write code here for fitness as a function of the variables you want to ******optimise, eg you might want A and B to be big as possible and C to be small ******as possible then
fitness =A*B/C
: LabelEndofBounds
FITNESS FUNCTION=fitness
END FUNCTION
https://www.sheffield.ac.uk/polopoly_fs/1.60188!/file/manual.pdf
- Badges
- Science topic