Look into the Ogre Frameworks that are availble. GOOF, Yake and a few others are available. They all extend Ogre to be more of a game engine. Obviously this adds more complexity.
Never Tried TrueVision3D first hand but I have heard good things about it. Other options would be C4, Beyond Virtual, Torque (I feel dirty just mentioning it…. but seeing as how it has an RTS component, if you have good programmers it should work well)
Source on the other hand is more than limited. Level sizes are limited, too many lights and it begins to slow down drastically. Too much visibility and it begins to crawl. That said the empires mod has done fairly well with it, but they have suffered from performance issues.
Personally if it was me, and I knew I had a good team of programmers, I would go with Ogre, grab a good framework, and utilize the community and the huge number of available plugins.
If I didn’t have a strong programming team, but had a relatively savy team I would go with Torque with the RTS component. and try to work within the bounds of what Torque offers out of the box. Perhaps adding some community code.
Do what you know will be easiest in the long run. Your course has set you the task of doing a game, there is no point going into the demostrations with a few spinning boxes, and some nice graphical features. They want a game, so figure out the strengths of your team and work on that.
Some words of caution, evaluate the engines workflow before you commit, and ensure you know what tools you are going to use. In the long run it will get you there faster. Pre-production is key here. If you take on Source or Torque, and realise that 1 month from your deadline, X tool does not support Y and this means your game will not come together. You will be kicking yourself.
[Edit] Damn you gizmo :P