The alignment here is a short alignment, we a priori don't know how close related are the species.

1. So in this case the defaults parameters should work fine, we do not need to do any modification except for the distance measure that should be change to kmer4_6, as we are working with nucleic acid sequences.
2. If you want to run Muscle in a faster way because you are confident about the similarity of your sequences, you can use a limited number of iterations and put a low number in the field: Maximum number of iterations, 2 or even 1 should be OK. The anchor optimization with low number of iterations (1-2) has no effect, but you should activate it for more iterations, it speeded the computation and have little or no effect on accuracy.
3. For huge alignments the first thing you should do is to have a look to the option of Max. number of hours for align. refinement, than reduce the number of Maximum number of iterations if your sequences are not too divergent, and let the default options for distance measure.