One driver, one monitor and one sequencer in one agent is the most common structure I have ever seen. I am curious if one agent can have many above components.
If an agent can contain multiple drivers, multiple monitors and multiple sequencers, is there any benefit to this structure?
I see no benefit to having multiple drivers, monitors or sequencers in one agent.
The most important reason to restrict them to one each is because that is what the methodology expects. It is important to adhere to common practices of a methodology to allow for standardization and reuse of testbench components.
An agent is associated with an interface (a group of related signals). It is common for designs to have several instances of the same interface, like a router. In that case you would use multiple instances of the same agent. The benefit is that each agent driver can act independently of each other, etc.