Best normalization scheme for portable mp3 player (flac->AAC)

[higgzbossunova]

Hi, I have done searches on this forum but I am still a bit unsure. I thought i'd seek help before converting from my lossless library to AAC for my portable player. Which like most does not support replaygain tag normalization. And while commuting I'm tired of lowering and raising the volume between tracks.
I mostly listen to a mix of files from different albums, and maybe whole albums 25% of the time. Majority of the music is from pre-loudness era, before 2000).

I'm wondering which i should choose:
a) replaygain(apply) - track (flac already have replaygain tags using 'rsgain': LUFS -18dB, true peak, clipping protection for positive gain/-1dB max peak)
b) volume normalize/Replaygain (track gain)
c) volume normalize/EBUR128 (track gain)
d) volume normalize/Adaptive (EBUR128)

I've read the DSP help page but still not sure which is most suited for my purpose, listening with earbuds amidst loud commuting noise on my player.
Grateful for any advice as I have never before normalized my music.

[Spoon]

>d) volume normalize/Adaptive (EBUR128)

This one if likely in a noisy environment

[higgzbossunova]

>d) volume normalize/Adaptive (EBUR128)

This one if likely in a noisy environment

Thank you, that's what I will use. If doing a batch convert, what order should i use to also resample any hires files to 16-44?
I did a forum search and you advised Conditional Encoding. Can i in one go set up a conditional encode for any hires to be downsampled, majority are 16-44. And also normalize using Adaptive (EBUR238)? Or maybe it's not possible.

If it is doable, would the DSP order be something like:
Conditional Encoding ->
Bit Depth -> 16, triangular dither
Resample -> 44.1
Volume Normalize -> Adaptive (EBUR128)

For the Conditional Encoding step, would it be:
Action-> Skip, If Any Condition Matches
Bitdepth -> 16 bit or below
Sample Rate -> 44kHz or below
General -> not checked

I've painstakingly selected many albums and am afraid of losing the selection of i do something wrong. the sage thing may be to just volume normalize and fish out the hires later.

[Spoon]

Yes highres to 44KHz, bit depth does not matter for AAC, resample can be set to 44KHZ if already is it is not done.