I was actually thinking about blogging on this kind of topic, but the OP's problem's as good as any to start with.
As I see it, GMs who try to run simulationist or sandbox games aren't going to be able to do it to the fullest extent of the definition, nor to a "realistic" level. To truly run a simulation, you need a process/methodology for determining everything. And I mean everything, from the exact x,y,z position of every entity in the world, to what they're carrying, to their current mood and goals. At best, a GM can fake it, and that faking it, is actually the same thing GMs in other styles are doing. The sandbox methodology falls into the same category, it's trying to create an open world, which creates a greater burden on the DM to document everything, to create the possibility of what players might do.
A key problem in the sandbox, is like Elder Scrolls: Oblivion does it. A PC can pick up ALL the quests in the game, but never get around to them. And nothing bad happens. And realistically, the GM has to skip applying those consequences, or the campaign world falls into mega-darkness.
What it gets down to is this: Simulationist or sandbox play is trying to achieve the holy grail of RPGs, which is an environment that is completely impartial, yet "realistic". Where it fails, is that it isn't practical for humans to run, and it doesn't always make for a good story. It's too complex to implement human emotion and drive into NPCs, without a GM to fake it.
But let's stop knocking the play-style's presumptions (or my observations about the play-style). Let's offer some solutions.
Problem: Players have analysis paralysis. They're taking too long to plan anything.
Solution: Apply a "realisim" house rule that says if the party takes too long to initiate action (aka arguing), the timeline advances and consequences of inaction will occur. Meaning, it's fine if the party discusses for 5 minutes which mission they want to do, or how to take out the guards at the gate. It's not fine to spend hours on it.
The idea Ydars #2 suggestion is good. Real time discussion planning costs game-time (since there's no rules resolution going on). You don't have to go so far as the "if you say it, your PC says it", but the general gist is translated to the game.
Problem: the players have too many threads/quests/missions going on
Solution: only give PCs missions/quests/plot hooks related to the current game situation and their PCs goals. On the first game with no real PC goals in a new world, that means giving them a "save the princess" quest that their PCs would undertake. Subsequent games should have missions/side quests related to consequences, new developments, and PC drive goals that develop.
Consider it like real life. If you work for somebody (say as a cop), you get sent on a call, or you see a crime in progress. You don't see 10 crimes in progress, and if you did, you'd still only handle one, and call the others in. Your boss is only going to give you a certain number of projects. If you have side goals (like moving to a new department), you work those angles on the side, while doing the main project.
If you're a freelancer, you don't often get multiple requests at the same time.
Consider it this way, in your sandbox/simulation, the NPCs aren't likely to bring their problems to somebody who's overburdened with their own (like the PCs).
To wrap it further, consider that ultimately, you ARE telling a story. Every PC session is a story, some are just more boring than others (just like every day in your life is a story). Don't make it too convoluted with a million threads. Therefore, talk to the players before you WRITE the next session's material. Ask them, based on the current game situation, what they want to do/pursue/attempt in the next session. Then write to that.
That's actually another part of the solution. Get the party to agree to which "quest(s)" they're actively working on for the next session. Are they going to the Dungeon of Disasterous Doom, or the Cave of Collosal Catastrophe. Commit them to one, and write that dungeon in the next session.
Problem: Too many threads currently in campaign
Solution: The previous solutions are how I avoid the problem of too many threads. Once you got too many, you need to trim them down, to really get the benefit of the advice. Step 1, tell the players you are going to do a new format, as above, to narrow dowb the prep work to just material the players are going to use for the session. Step 2, have NPCs complete some of the other threads, or have them resolve themselves. Basically pull a Mark Twain and drop the magnificent twins down a well, because they weren't as important. Optional Step 3 You can even roll the hero NPCs in later, as the party has to deal with them (perhaps saving the NPCs on a future mission).
Step 4, explain to the PCs that you've done all this to clean up the game world to things the party is actively working on, and they don't need to worry about it. If the old quests come back later, its a so that you can bring new challenges to the party.
Problem: Players are too paranoid
Solution: Firstly, players need to understand that anything bad that happens to the PCs, is either they're fault, or the GM is doing it to create a new problem and story. The game doesn't happen in a vacuum. It doesn't wait for the PCs to go left or right. And ultimately, every interesting thing that happens, is because the GM gives it life. This is true, because the GM has the power to make everything the PCs do be uneventful and boring, as well. Therefore, establish a contract with the players that the GM's job is to create challenges, drama, setbacks, and problems that are entertaining, and within the ability of the players to deal with (though it may require some time and effort on the PCs part). If the GM didn't do this, every dungeon would be empty, all NPCs would be content and need no help, and nobody would resist the PCs actions.
This first step establishes that it is the DM's job to screw with the party, and the game would be boring if he didn't. Next, you have to establish the right time to screw with the party. Like Red Herrings, and Bluffing in poker, you need to do it seldomly. Make more missions straight-forward, than crooked. Plot-twists should be done at the right time, not everytime. Consider it like fighting. If you do the same attack 3 times in a row, you set a pattern. Then you do the 4th so it looks like the last 3, but is actually something different.
Problem: Players are argumentative
Solution: Prior suggestions should reduce this, but you may need to directly address it. The players need to understand that discussing things should involve everyone, and it needs to be decisive to be valuable. Otherwise, it's arguing for arguing's sake. Consider the following format: every player should get to make a statement. Every player should try to state their concern about the problem, as well as what problem they percieve (many arguments are because 2 people aren't solving the same problem, and don't know it). If a player repeats himself, he's done, he's not adding anything new. After that, the party needs to vote, and the players need to agree to stick to it.
Problem: players stick to "doing what my PC would do", rather than good of group
Solution: The real problem with this mentality is that it is used to go against the will of the group. It's often used so the PC can act like a jerk to another PC. The reality is, in almost any situation, the player COULD come up with a rationalization for the PC to foster getting along, and working together. Step 1, make sure your Player knows you have a dim view of this behavior, and why. Step 2, avoid creating situations that create conflict that you don't want the party arguing about. It's one thing for a player of an elf-hating dwarf to act like a jerk to the player of an elf. It's another to expect the paladin to help sacrifice babies in order to progress the adventure. The reality is, you may want the players to feel conflicted about NPCs doing something. You should avoid having the players be conflicted with each other. This requires help by the players, to not make mountains of mole-hills, and the DM to not agitate things.
Problem: player is contrary, just to be contrary
Solution: Story: I sat in a communication class for work, where 6 teams had to come to a consensus. It was done by each team send in a representative to a 10 minute "discussion" round, with the end of the 3rd round culminating in a vote. In the 1st round, we learned that nobody was listening, they were just repeating themselves. In the 2nd round, we got more orderly debate. We also discovered that some people were deliberately choosing the "worst" option, just to be contrary. They had their justification. They wouldn't budge. So between the 2nd and 3rd round, some of us figured it out, and we talked to each team to send in specific people for the last round. We convinced each team to send in the more passive member, not the driven aggressive one. The result was, we actually got a vote done. THe exercise wasn't about making the right choice, it was actually getting a vote done, before the time limit was reached (by not wasting time arguing).
The key lesson was this: Some people are Drivers. They set a goal, and they push very hard to achieve it. They usually suck at listening. Some people are contrarians. They try to argue with anything. Combine the two (and they usually do), and you've got a roadblock.
Here's the thing, roleplaying is a form improvisational theatre. In Improv, a blocking action is anything that prevents another actor from getting involved, or moving the scene forward. It turns out real life conversation, is also a form of this. Thus, your best players, and your best conversationalists, are those that include others, and listen to others.
The point then is that habitual blockers don't get invited back to parties. They're boors, they're rude, they're argumentative. After explaing how being personable works, you're going to have to point out that non-personable people don't get to play. The reason is, that almost all gaming groups except any PC. They get a pass, because the game must go on. In reality, nobody would hang with a jerk, or invite them back. So apply real life common sense here.
