The OP may also be interested in this:

Case #1
select *
from T
where a = 1
and b = 1
and c = 1
/

vs.

Case #2a
select *
from T
where a = 1
or b = 1
or c = 1
/

vs.

Case #2b
select * from T where a = 1 UNION
select * from T where b = 1 UNION
select * from T where c = 1
/

What is best index strategy for each query?

Well, this is a bit simplified but best strategy might be the following:

Case#1 -- T(a,b,c)
(one composite btree index)

Case#2a -- Bitmap=>T(a), Bitmap=>T(b), Bitmap=>T(c)
(three bitmap indexes one on each column)

Case#2b -- T(a), T(b), T(c)
(three btree indexes one on each column)

I said it was a bit simplified, but it does get to the basic point at hand. Different variations of the query (Cases 2a and 2b being symatically equal (they return the same rows)), can be optimized differently using different index sets. How well each index alternative works depends of course on your code, and how you use your data. This is particularly true when considering bitmaps indexes.

I am not suggesting any specific index set for the OP's situation, nor am I promoting one coding sytle over another, only trying to give some ideas from which you can do your own research. I particularly like CASE#2b because it gets poeople talking.

Good luck, Kevin