Prerequisites:

Description:

    This is the science of manufacturing artificial ecosystems, generally for the purpose of supporting humans (or one's own species, if other) at the top of the food chain.  Specifically, it is the concept of creating a completely artificial combination of life forms in an artificial habitat, and then of optimizing it.
    There are many things to be optimized for, and in general there will be a tradeoff between the two.  The first optimal condition is an ecosystem that will support a maximum number of humans in a minimum volume(energy consumption).  Spacecraft intended for deep-space work (as opposed to those which must sometimes enter environments hostile to large mobile structures, like planetary atmospheres) can quite often be designed to allow for truly enormous volumes within.
    Another ideal is to minimize the mass of the ecosystem in question.  Since mass is always at a premium for space travel, reducing the mass associated with keeping one crew member alive and active by even a few percent could mean far more available crew on a large multi-thousand-crew ship.  Substantial improvements in this technology would require genetic engineering to achieve, thus the Genetic Engineering (dev) contribution in this research topic.
    A third thing to minimize is the energy output required to maintain a single person.  Near a suitable star, one might be able to use sunlight directly (such as for a stationary space station-style habitat), but in other locations, lighting and heating will be necessary to support the plantlife (or alien equivalent).  The ability to reduce this requirement without reducing the crew would be of substantial value if the technology set in question does not include an energy supply so plentiful as to make the requirements of life support insubstantial.
    A fourth optimization is the halflife of the survival of the ecosystem.  Real ecosystems are reasonably resilient, but can break down under large stresses.  The smaller an artificial ecosystem is made, the less resilient to the stresses imposed on it it can be expected to be.  At each Engineered Ecosystems (dev) level there should, therefore, be a tradeoff between the size of an ecosystem and the amount of time it is able to exist alone before it will break down and become unable to support higher life forms.  This could lead, for example, to huge city-ships being required for early interstellar travel even when other technologies allow for smaller ships to do the job.  Also, a ship that is undermanned (ie if most of the crew are in some form of stasis, or the ship is fitted with a larger-than-necessary ecosystem)  should be able to survive for a longer period of time between refits, as the ecosystem in question will be under less stress. Honestly, the stress should be represented as a half-life of the ecosystem, but that would lead to quite a bit of unpredictability in the game, and a simple lifetime might be more suitable for a video game.
    Fifth is the problem of who maintains the ecosystem and actually grows the food.  If people are to do it by hand, then that adds to the total mass of ecosystem needed (to support the farmers).  Robotics is an answer to this problem: robotic farmhands require only electricity and maintenance, not food, and thus would reduce the number of farmers dramatically.  The other alternative, as I see it, is that if genetic engineering development is sufficiently advanced, it may be possible to put together ecosystems that do more and more of the job of taking care of themselves by themselves.  For example, it might be possible to engineer an insect colony to harvest grain (part for themselves, part to be eaten by the crew), or to arrange for a certain number of animals to find their way to the slaughteryard section of the food preparation area each day.  The difficulty of automation is partly a problem of the development level of genegic engineering/robotics (whichever is most highly developed) and partly a problem of a tradeoff with the time of survival of the ecosystem (Fewer people watching the ecosystem means greater probbalility of it breaking down.).
    Probably the best detailed way to deal with all of these tradeoffs is to give an ecosystem design a certain number of points to work with (based on Engineered Ecosystems (dev) level) to divide among the competing ideals.  A simpler way for low-complexity games would simply be to adjust the halflife of the ecosystem based on development level, and ignore the rest.