The problem with overly general design questions like this is that there's interesting territory to explore in either direction! If you provided more detail, maybe answers to some of these questions would flow naturally from the decisions you've already made.
1. How do you feel about auto-battle computing? (So you don't have to fight the battle yourself)
If your players feel like fighting battles is dull or a burden, you fucked up.
As Veracity says, if the battle's definitely going your way, the enemy can just surrender. You can even make getting them to surrender earlier a bonus condition, either preserving your forces or granting you some of their resources in the broader game.
2. How should units be deployed into the battlefield? Dropships or something else?
In a manner which is appropriate to both your fiction and the scale of your battlefields. Realistically speaking aircraft have very limited carrying capacity, are easy to see coming, and since the 70s have been too vulnerable to missile strikes to use en masse. On the other hand, dropships are cool as hell.
3. Since my game is more combat based, what is your opinion of handling all resources harvesting in a home base? In other words, you can't build every structure in every battlefield, but only in your home base.
You could drop all structure building on the battlefield and only
have a home base, like X-Com. Depends how long, involved, and large in scale you want the battlefields to be.
4. Should the world map be in real time? Your enemy can be fighting another battle in another part of the world map while you can opt to fight another one simultaneously elsewhere.
You could definitely build an interesting game this way. You could allow the player to hop back and forth between ongoing battles, potentially overwhelming them with multitasking and forcing them to choose where to use their time as in an RTS, but you would probably have to simplify the tactical game to compensate. You could lock the player into the tactical game they've started and have other things play out in real time, but then you risk the player losing deep swathes of territory while they're busy in a long battle, making them lose the game without a chance to respond. That'd piss people off.
Atom Zombie Smasher has a turn-based metagame - you pick your battles, and the asymmetric enemy makes fixed advances into neutral territory during each battle. I think this is an interesting mix of forcing players to choose their battles without letting things get out of control while they're busy.
5. What do you think of AI assistants fighting a battle for you or assisting you? They would automate resource gathering, or even combat if you choose.
If a decision is usually interesting, have the player make it.
If a decision is usually obvious, automate it or remove the need for a decision in the design, as it's just busywork.
If a decision is sometimes interesting and sometimes obvious, consider triggering the obvious option by default and letting the player override it. Examples:
- When a weapon runs out of ammo in an FPS, modern games begin a reload automatically. Maybe the best option is to change weapon instead, but the player can override the reload to do that. Simply reloading is always better than standing around with a normal weapon.
(Counterpoint: in Deus Ex: HR, you can't sprint while reloading. Because of this, in certain situations, automatic reload can hurt you.)
- Units that are standing around in an RTS which are fired upon will generally fire back. The best decision may be to run away, but simply firing back is *always* better than just standing there.
There are simple, predictable ways of automating easy decisions, and you should use them. But never, ever
ask the player to rely on a remotely intelligent AI. If it's seen to be making decisions, even if it's smart and comprehends the game perfectly, it will occasionally make decisions the player disagrees with or doesn't understand and the player will get pissed off with that (even if it was a good decision!). The player is forced to constantly keep track of what the AI's doing, which is actually more work than just doing its job, and learns to resent it for this.
See squad AI in any number of games. If squadmates are invincible or expendable, they're allowed AI which advances and takes up positions, because there's no way that can cause serious problems. If they can die easily and that death would seriously hurt the player, squadmates are allowed almost no autonomy and largely move and fire as instructed.
6. Do you prefer squads or individual infantry?
Depends on the scope of the battles, the setting, the importance of individual unit control in the combat mechanics... this links into too many other factors.
7. What do you think of grand strategy in RTS games in general?
I think there's much more room to bring new ideas to such a concept than there is to a traditional RTS. I also think it's the better choice as an smaller developer, since longer-form multiplayer games like Civilisation are best played with friends (whereas shorter-form multiplayer games like Starcraft or Team Fortress die without a vibrant player community; arranging a match with friends is much less worthwhile).