1) That depends on your degree of certainty regarding the slugs. From an analysis point of view, I'd suggest a time history analysis if you start loading the system with mulitple slugs. This will be more complicated, but I don't think a response spectrum analysis can do this justice if you have more than one slug.
2) The limiting DLF value for a single impact load is 2 (from dynamic theory). Once you have multiple impacts, this limit is no longer applicable.
3) There is a big article in the
November 1994 newsletter on how this is computed.
4) For a single slug I think the spectrum method is a better approach. However, regardless of what you do, your results are only as good as your estimate of the slug.