>>8667976
The best explanation I've seen for different Samplers and Schedule Types is https://civitai.com/articles/7484/understanding-stable-diffusion-samplers-beyond-image-comparisons

The long and short of it is they're all different mathematical algorithms that can have different pros and cons, but there's no one "best" one.

Some are faster than others, some are sharper with more detail while others smooth thing out more, etc. Notably is Ancestral Samplers versus non-Ancestral. Ancestral Samplers at noise at each step, so the image continues to change as more steps are added. Non-ancestral samplers converge on a single point, so after enough steps to denoise the image there won't be a significant amount of change. Non-ancestral is better if you want more consistency. Ancestral samplers are the ones that end with a like Euler a.

Euler a became popular because it's fast and it's a smooth sampler, so it tends to smooth out jaggy lines and such. It's actually less detailed than other samplers, some are geared towards photorealism with maximum detail but that makes imperfections more notificable in 2D anime style gens.

Pic related is a test of different samplers and schedule types. Really, the best thing you can do is just try them out and see what you like.

Full res test: https://files.catbox.moe/dtzbfs.png