MCDM and visualization assumption.

[muazhari]

In my optimization use case, one of the objectives is maximization.
To be calculated by Pymoo, the maximization objectives must be multiplied by -1 (converted to minimization).
Based on this, after the minimize() function executed, the result.F which are maximization need to be multiplied by -1 again to back to the original.
Then, I want to supply the result to MCDM and the visualization features in Pymoo.
Does MCDM and visualization always assume the input of objectives is minimization?

Also, can you provide the normalization/preprocess function to be able easily used for MCDM and minimization in Pymoo?
Unknowingly, I do not have any knowledge to do it right.
Especially when faced with edge cases like what if there is a division of zero?
You can directly correct the formula below in this discussion because I need it as soon as possible.

nF = (res.F - approx_ideal) / (approx_nadir - approx_ideal)

i = decomp.do(nF, 1 / weights).argmin()

Thanks in advance!

[blankjul]

The entire framework assumes minimization. So does MCDM.

I recommend fixing the edge case by using an eps value (this is probably the most practical fix), e.g.

max(weights, 1e-16)