[Docs] Add source of LogEI formula

[benjamc]

Why: LogEI looks weird, we need an explanation

Assumptions: log-normal observations

Ancient: Efficient Global Optimization of Expensive Black-Box Functions [Jones et al., 1998]: EI formula ("derived by integrations")
Source: An Experimental Investigation of Model-Based Parameter Optimisation: SPO and Beyond [Hutter et al., ???]
Reference: Sequential Model-Based Optimization for General Algorithm Configuration [Hutter et al., 2011]

[benjamc]

@KEggensperger'

Wouldn't EILog be a descriptive name (to emphasize that this is EI of a logged value)? And then we can also have LogEILog.

Derivation from @nabenabe0928 attached
memo.pdf

[nabenabe0928]

memo.pdf

@benjamc @KEggensperger
The derivation above would be more precise and it is with the citation of the original paper as well.

[nabenabe0928]

@benjamc
I published the derivation as an arXiv paper!

https://arxiv.org/abs/2411.18095

[benjamc]

Awesome, thanks @nabenabe0928 !