There is nothing like pumping some “cranial iron” to get revved up for the weekend (That does make me sound like a boring person, doesn’t it? But it was fun!)

Have a good one.

You are correct with your fitness calculations.
The crossover pairs can be chosen in a variety of ways. Here I went for a biased random (I picked at random and, if the two didn’t produce a useful new individual went back and picked two others)

Even though I don’t fully understand what’s going on, I tried doing the reader’s exercise and was stumped for quite a while trying to come up with your results for the 2nd generation. This carried on for a while until I realised that the fitness for 1010 0101 0010 0101 isn’t (10-5-4+5) = 6, but rather (10-5-2+5) = 8.
Lacking natural intelligence, not to mention the inability to comprehend the artificial sort , I was concentrating on the numeric values and ignoring the binary ones.

Two questions:
1. Am I on the money with the above?
2. How does the algorithm decide which two individuals are to be crossover partners?

Otherwise, thanks for your blog which is always an interesting read.