Whilst re-implementing the paper “A Cooperative Coevolutionary Approach to Function Optimization” by Mitchell Potter and Kenneth De Jong for a Computer Science assignment, I noticed a typo in one of the benchmark functions.
In the paper, the Schwefel function is listed on page 5 as
f(x) = 418.9829n + sum(x_i * sin(sqrt(abs(x_i))))
but the actual function is
f(x) = 418.9829n - sum(x_i * sin(sqrt(abs(x_i))))
Notice the sign change in the middle.
I have found 3 references that back this up:
It’s probably also worth noting that one of these sources give the constant to a greater accuracy; 418.982887.