why can I have a valid output when the input is less than 1000-dimensional

[HelloDuoLA]

F15 is a 1000-dimensional function, why can I have a valid output when the input is less than 1000-dimensional

[dmolina]

I am sorry, this code is a wrapper of the original C++ core, actually it is evaluating a 1000-dimensional vector, in which the first positions of the vector, and the rest of values are the consecutive positions in the memory to achieve 1000 values. The problem is that using the simple API allowing both numpy and list, it is not easy to check the dimensionality. Do you consider it important?

[HelloDuoLA]

OK, thank you for your answer.

[Wukong-SCUT]

This is a point worth emphasizing. Because for F13 and F14, they have overlap=5, so only 905-dimensional vectors are needed. But info tells us that we need a 1000-dimensional vector, but in fact the following 95-dimensional vectors are useless. This can be misleading because when checking the VI, the algorithm may change the values ​​of the next 95 dimensions but get the same fitness, which can be a disaster for the algorithm results.