High redo generation - suggestions required

Hi,

My database is generating high number of redo logs ( average : 30 per hour with each redo log 100 MB size. i.e 3 GB per hour) . I can not avoid this number as we do generate lots of DML.

Please do suggest me the best optimizing techniques:

For example,:

1. increase the redo file size to 200 MB -> generates less number of redo logs. Oracle suggest 5 to 6 per hour only.

2. increase the redo log groups, so that server is not waiting on arch process to complete writing.

3. increase lgwr processes.

4. log_checkpoint_interval / log_checkpoint_timeout values..

Please also correct me the above points if I am wrong. Please also suggest me the best practices you follow in this case.