there is no easy formula for this. from my 2 production databases:
Database 1: 14G used, 10-15 concurrent users.
3 redo logs of 250M each

Database 2: 5G used, 100 concurrent users.
3 redo logs of 15M each

Both database turn over a redo log about once an hour. As you can see, there is little correlation.

fortunately this is easy to remedy if your redolog turn over too quickly (i like to see logs turn over every 60-90 minutes during normal activity). You can do this online with an ALTER DATABASE command. Just be *very* careful about deleting the old under/over-sized files and make a backup of your controlfile if you're not doing a database backup soon.

D.