1. I am not using the tag for now because I do not want to restrict your imagination. However, I will add it if I get mostly speculative answers with no support in actual science. Please provide links supporting your ideas. Please, don't be shy, I love reading.

  2. Please ask for clarifications before you write your answer if you need more details. My worldbuilding project is rather detailed already; if your answer is based on wrong assumptions it would not be as helpful.

The space station is a permanent habitat for the last human survivors in the known universe. If this station fails the humanity will cease to exist forever. The survivors are working on a terraforming project to build a more comfortable and lower maintenance habitat, but it will take centuries if not millennia before they can establish bases on the ground. Please, do not challenge this premise.

My initial design includes these ideas (this can be challenged and improved):

  • the basic structure is a wheel about 10km in diameter;
  • artificial gravity is created by centrifugal force, the maximum g (at the rim) is 1.2;
  • the wheel is separated into self-sufficient compartments that can be fully sealed in case of a hull breach or any other malfunction threatening life-support systems;
  • the station is run by the AI, which is possibly sentient and sapient (this is a topic for a heated debate in my world) and prioritises humanity's survival to the point of obsession; the AI is still young, however, and does not have necessary expertise and experience to handle all tasks related to the station, but they are learning fast;
  • humans can override the AI's decisions on everything that does not relate to the immediate safety of the station and a human;
  • life-support systems are partially organic: water filtration and air oxygenation, for example, are done using algae with the help of vertical farms and small gardens placed strategically around the station;
  • the station has thrusters for orbital corrections, but no big fancy engines to travel around the star system;
  • the station is also equipped with solar panels and sails.

Additional assumptions:

  1. this is a virtually post-scarcity society, there are plenty of resources around (asteroid belt, nearby moons, etc.) that can be mined and delivered to the station's building site; time is the only restriction, resource-wise;
  2. post-scarcity does not apply to organic materials — only those that can survive a 2-century-long journey are present;
  3. the technological level is slightly higher than ours:
    • please assume that any technology showing promise today is available;
    • advanced robotics and almost full automatisation;
    • no FTL or magic-like matter manipulation;
    • cloning technologies exist in theory, but without artificial wombs or some similar technology cloning humans (or animals) is not possible;
  4. the survivors have access to all knowledge and information gathered by humanity by the date of departure from Earth.
  5. people are the most valuable resource, there are very few of them and, ideally, they all need to survive and procreate successfully.

I am mostly interested in these aspects of the space station design (these are not separate questions, they are parts of the overall design):

  • radiation shielding;
  • structural integrity;
  • feasibility of several small eco-systems vs one large eco-system (I am not creating complete biomes, my eco-systems are closer to aquaria: a limited number of species with constant maintenance and fertilisers, but the bigger the aquarium the more stable it is);
  • safety and redundancy features.


What would be the best or optimal design for a space station given all the constraints listed above?

Answers to the clarification requests

  1. Why can't they find a planet or a moon or something bit o plant roots on?

They want to be close to their terraforming project. The station is envisioned as a permanent satellite for their future home. It is not possible to land on their chosen moon since it will undergo heavy bombardment as part of the terraforming project.

The moon orbits a gas giant just slightly beyond the Goldilock zone of an M-class star. The moon itself is bare and has no atmosphere. The gas giant's composition is to be determined in greater detail, but it does have some water. Please do not challenge this. It works.

  1. Young AI? How does that work? It's like copying a book, nothing is messing unless you decide not to copy that page.

AI is not a book. It is a set of self-learning algorithms, various programming libraries, and databanks. The AI needs to learn things to gain expertise and experience. The space station is new, many things are different from what is observed on Earth, so the AI needs time to collect data about the new environment to optimise the performance. The AI is smart, but not a god.

  1. How many people are on that station?

About two hundred on the arrival day. The station must support up to 10000 people in the future.

  1. How do they handle food or water? Without a steady supply of them their farming and water will ground to a halt. Are they "farming" an asteroid?

Yes, they are mining asteroids, nearby moons, and their friendly gas giant. Inorganic materials are not of any concern. They just need to be delivered to the construction site.

  1. Do you want the rotation wheel-like space-station to precess? Or not?

Whatever works best for a station orbiting moon orbiting a gas giant. There are other moons as well. The gas giant does not have rings. The colonists are planning on building a giant magnetic shield to create an artificial magnetosphere for their future home. But the station is their #1 priority. The ship feels way too overcrowded :)

  1. One point of clarification though - is there a question whether the AI is sentient or whether it is sapient?

It is debatable. The AI passes Turing and similar tests. However, the status of the AI as sentient or sapient being is irrelevant. This is only a subject of philosophical debates among humans. If you are interested, the AI does not have anything resembling Asimov's 3 laws.

  1. What's the AI's threat profile? Is it mostly worried about astronomical phenomena that could be dangerous, or about the dumb humans breaking things inside it, or about an external threat (one presumes there's a reason this is the last enclave of humans...) coming after them?

The main dangers are natural astronomical phenomena and human errors. There are no foreseeable external threats. Big scary aliens with even bigger and scarier guns/lasers/biological weapons do not exist in my universe.

Please assume that humanity's self-destruction crusade has finally ended in total and indisputable victory. My people are the last remnants who want to avoid past mistakes.

  • $\begingroup$ Sounds like a fun idea. Why can't they find a planet or a moon or something bit o plant roots on? Water, atmosphere, food...etc would insanely easier there. Is it a part of the story that they reach a point of such tech but can't get to any planet? I think this is the biggest issue. Young AI? How does that work? It's like copying a book, nothing is messing unless you decide not to copy that page. How many people are on that station? How do they handle food or water? Without a steady supply of them their farming and water will ground to a halt. Are they "farming" an asteroid? $\endgroup$
    – Seallussus
    Sep 13, 2020 at 20:56
  • $\begingroup$ I'm sorry but what exactly is your question? $\endgroup$
    – N. Bleac
    Sep 13, 2020 at 22:42
  • $\begingroup$ Good question, (@N.Bleac it reads as "what is the best design within the constraints that meets the 4 dot point considerations at the end?) One point of clarification though - is there a question whether the AI is sentient or whether it is sapient? (I was unaware of the difference until a previous answer of mine was questioned recently.) $\endgroup$ Sep 13, 2020 at 22:45
  • $\begingroup$ What's the AI's threat profile? Is it mostly worried about astronomical phenomena that could be dangerous, or about the dumb humans breaking things inside it, or about an external threat (one presumes there's a reason this is the last enclave of humans...) coming after them? $\endgroup$
    – Cadence
    Sep 14, 2020 at 0:05
  • $\begingroup$ @Seallussus I updated the question to include answers to your questions. Please let me know if you need more details. $\endgroup$
    – Otkin
    Sep 14, 2020 at 0:24

1 Answer 1


I'm no engineer.

Here is a couple of major concepts. That I think influence, and perhaps challenge, the question. They are in no order.

  • The AI, BB short for Big Brother, is a dictator with absolute rule. Why? Because at such depletion of humanity it needs to be. It needs to be above anything including morality.

  • The most logical things is cryogenic sleep with BB handling the multiplications of humans in the most efficient way possible. Depending on the tech it can grow people in pods or bring up a father and mother, based on the best genetic matching, then again put the father to sleep while keeping the mother until the child is being born. This means that your people are asleep until the actual future happens.

  • If you played Fallout or Bioshock then the concept of the human element leading to absolute disaster is as plain as day. Or rather that once you put humans in the mix then nothing, and I mean nothing, is out of the question and the variables and parameters are always changing. You can't protect humanity from itself. Case and point the fact that only 200 people survive. BB is too smart to actually let humans run around and have control over toilet paper let alone anything. I'm no AI but that what I will do. Maybe 2 generations later the old timers will complain about my laws but we will have humanity. This is it's own point as even if you don't put the people to sleep then BB is still aware of that. People will ruin things.

  • Cryogenic sleep also saves resources as humans expend a lot of resource and are absolutely useless at such a setting. A mechanical contraption can extract resources out of an asteroid a 1000% faster than a human and even if it failed it won't endanger the survival of humanity.

  • Commercialization to the maximum. This is the big league here. I'd split the station into equal survivable parts with 0 communication or contact. They are set apart from each other and the protocols of handling even computer data between the parts are like running through a medieval castle made by a doomsday prepper with unlimited resources. So I would split that long wheel into 3 or more, I'm just putting a random number here, and have each capable of survival on their own. So the wheels are their own thing. Not connected.

  • So imagine the Migrant Fleet from Mass Effect. A large flotilla of ships and substation loosely orbiting whatever they orbit or just floating in space. Again this insures that a fire is station A does not spreed to station B and ruin the rest of humanity.

  • However because this is post space travel apocalyptic setting. Would it not make more sense for the station to be made of already existing stations/spaceships? It is far cheaper to incorporate bits and pieces of old tech than make new one, I mean if you already have a factory then problem solved, so won't this station be a mismatched station of parts of all sorts of different ships and stations and older tech? Like Spacehulks from Warhammer 40K. They are basically huge space ships that drifted together to create a large and mismatched ships that drifts in and out of the warp, hyper space, randomly. They include parts of ships 20000 years old as well as the newest stuff. They are so big that they have their own gravity. And they have some valuable stuff as they are so old it is crazy. Though they are usually filled with demons or crazy aliens of a sort or another. Point is why not incorporate the stuff?

  • So while we are at it why bother with gravity or life support? As I said people would be asleep so the station should only be made to simply support people. Also as the resources are limited it would free more resources to terraform the new home of humans.

  • Another problem of letting humans be awake is basically Metro 2033. You know the world is destroyed and the only place to life is the Metro stations, bigger in the world and made to shelter people, underground. only 2 decades later people are split into warring groups, one of which are neo nazis and another one which is communist, and kill each other while crazy level dangers haunt humanity and exist within and without, does that word work here?, the metro. I'm not randomly citing this as again we see this in Rapture, Bioshock games, with the city having a civil war like 2 minutes after it opened. So again the dangers of letting political humans exist and run around doing all sorts of crazy stuff is too much.

  • This is how I think the situation would be. Perhaps a bit of a frame challenge. My point is that slightly more advanced tech and an AI would means that the strength of arms of humanity is useless and even our thinking is not that much better. Even if we are better than BB the potential danger of us ruining things, yet again, is too high. The only thing people will be good at is making babies.

Anyway I don't mean to be against the premise or anything. Just sharing what I think is interesting and pushing things a bit. You can include whatever part you like. Others I'm sure can come up with like actual shapes and equations.

  • $\begingroup$ Cryosleep is not available and growing humans in vats is not possible either. That would be too high tech. A sort of suspended animation is available, but it has a lot of limitations. Old tech is not available, because these people are originally space colonists. They found out that they are the last of humans when they were very close to their destination. I was thinking about building the station around the original colony ship, but since most of the colonists are scientists they want to start from scratch and build a perfect habitat :) $\endgroup$
    – Otkin
    Sep 14, 2020 at 5:24
  • 1
    $\begingroup$ Where is the AI stationed if the station comprises multiple unconnected wheels? Or does it work as a distributed network with multiple datacentres and backups in each wheel? $\endgroup$
    – Otkin
    Sep 14, 2020 at 5:26
  • $\begingroup$ @Otkin, My thoughts exactly. Backups for days. $\endgroup$
    – Seallussus
    Sep 14, 2020 at 15:48
  • $\begingroup$ @Otkin, Does not really address the core problem of letting humans have control. Being scientists dose not exclude them of human politics, self destruction or good old mistakes. "prioritises humanity's survival to the point of obsession" If what I said what I as a human would do, then an AI is more likely to be more extreme. $\endgroup$
    – Seallussus
    Sep 14, 2020 at 15:52
  • $\begingroup$ This is a very good point. But the AI will go about it a slightly different way. Humans are not really good at being commanded, but they are very good at doing what they believe they want to do. So, the AI is going to suggest ideas and make sure that the humans borrow them and push as their own. Instead of BB, the AI is going to become the Master Manipulator :) $\endgroup$
    – Otkin
    Sep 14, 2020 at 20:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .