3. random mutation”nucleotide rearrangement, sufficiently to cause a gene to express a different amino acid in a protein sequence (NON-SELECTIVE).
I agree with you that mutation is non-selective, but this is a rubbish description of mutation. For a start there is not need for a change in amino acid sequence to produce a potentially adaptive mutation and for another 'rearrangement' simply suggests the sort of recombinatorial shuffling you get with sexual reproduciton and totally fails to encompass the wide spectrum of potential genetic mutations.
Your cases 2 and 4 I would argue are, or at least can be, selective. Assortative mating and other forms of sexual selection certainly are. It should be obvious from your own definition of Natural selection that this is the case since sexual selection will clearly lead to differential reproductive success of the sexually favoured traits. The gene flow one is more arguable since there is a clear overlap with the issue of mate selection as to the degree of gene flow between populations.
So really drift and mutation are the only solid candidates I see here for clearly non-selective mechanisms of allele frequency change.
With only the non-selective factors you may get evolution in the trivial sense of changes in allele frequency but never the sort of adaptive evolution that is what makes evolutionary biology so interesting.
TTFN,
WK