I don't know what articles could help us, so let me share my humble experience.
The Active/Active traffic on BIG-IP means that you have one float IP for each device/traffic group.
Therefore, the devices do not share the processing with each other. Is in your hand to decide how to load balance the traffic (E.g. on DNS).
I think the major advantage is to have another BIG-IP to distribute traffic (E.g. separate in external/Internal or incoming/outgoing traffic, business segment, etc).
The major disadvantage is you must constantly monitor and care about the CPU usage on all units.
For example, in a failover group having two devices, each unit should have no more than 40% CPU usage to guarantee the online unit receives traffic when a peer fails and it don't break in collapse.
In another words, the sum from traffic groups could not overtake 80% on a single unit.
Agreed with @Claudio Rocha .
Normally Active-Standby setup is recommended one. If you are configuring Active-Active setup, you need to configure it in such a way that in case one of the node goes offline, other node should take all traffic. Also if you have only LTM then it should be fine. But if you have other modules as well like GTM etc, you need to verify in this regard also.
Also you need think on how you are going to manage nodes. As both are active i.e. Node A is handling A traffic and Node B is handling B traffic. So you need to carefully make daily configuration so it will work as per your expectations.