Selecting the best array of values after running through loops

See more:

Python

def func(number1,number2,number3):
    
    add_div_1 = (number1 + number2 + number3) / 3
    add_div_2 = (number1 - number2 - number3)/ 3
    add_div_3 = (number1 + number2 - number3) / 3
    add_div_4 = (number1 - number2 + number3) / 3
    
    cond1 = abs(add_div_1/add_div_3)
    cond2 = abs(add_div_2/add_div_4)
    
    opt = cond1*cond2
    return opt

one_num = func(1.5,2.5,3.5)
print one_num

array_n = [10,20,30]

b = 0
bt = 1   # step

while b < 10:

    b = b + bt

    for i in array_n:
        MAX = 31
        r = rand(3)
        r = abs(2*MAX*(r-0.5))

        num1 = array_n[0]
        num2 = array_n[1]
        num3 = array_n[2]

        n1,n2,n3 = r

        new_num1 = num1 + n1
        new_num2 = num2 + n2
        new_num3 = num3 + n3

        newnums = [new_num1,new_num2,new_num3]
        print newnums

one_num = func(new_num1,new_num2,new_num3)

if one_num < 27.0:

    print "The number is lower than 27.0 .","The number is:", one_num
    print "The number values are:", newnums

else:

    print "The number is larger than 27.0 .","The number is:", one_num
    print "The number values are:", array_n

The code above works with no problems, but does not do the job I want it to do. And, I am not sure how to do that. You will need to run the code in order to understand what I am about to explain.

At the end, after you see the thirty arrays of values, you will see either the two statements with "The number is lower..." or "The number is larger...". For either pair of statements, the values are selected from the last array of those thirty arrays. Instead, I want the program to choose the best array that gives me the "one_num" as low as possible from 27. Then, when I run the program again, the program will create all new arrays, and from that it should choose the best array that gives me the "one_num" as low as possible from 27. But, if the "one_num" in the 2nd second run or any in other run after its previous is higher than its previous "one_num", then keep the good value (lower one_num).

I am sorry, if that is confusing, but it is hard to explain programming.