Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  

Author Topic: Living systems theory and organizations in games  (Read 3508 times)


  • Bay Watcher
  • He has a fertile imagination.
    • View Profile
Living systems theory and organizations in games
« on: September 07, 2012, 02:46:34 pm »


I have always liked the games which let the player manage people. Most just simplify the process, leaving him with mindless automatons who just follow orders and, sometimes, maybe leave or attack when conditions become unbearable. Very rarely your underlings will have any ambitions beyond serving you and getting paid. Rome and Crusader Kings 2 are precious examples of the game where your vassals may plot against you.

Still, I sometimes dream about a game where the main difficulty would be managing an organization. Dealing with internal backstabbing, competition, disloyalty, corruption and inefficiency. Orders could be subverted, reports falsified, work deliberately sabotaged. I would also like it had some more forms of intra-organizational hostility than outright violence and hard-coded espionage options: AI should be able to see weaknesses in the enemy's organizational structure and act accordingly. Which manufacturing process to sabotage, which underling to bribe to turn a blind eye, which manager to best approach to strike a lucrative deal. Of course, this won't happen nowhere in near future.

One of the things we lack is a nice, simple model to describe an organization. Something standard enough that could incorporate anything from primitive tribal villages to corporations, governments and espionage organizations. Sounds hard, but I think I have recently found something.

Living systems theory is quite old (1978). Its main principle is that all living things operate at least a bit similarly. They need to take matter and energy from the outside any organize it using information to resist entropy, maintaining a stable state. Moreover, they tend to organize in levels: cell, organ, organism, group, organization, society, supra-national system, each level behaving somewhat similarly.

Have you ever criticized corporate personhood? Well, according to this theory, they are at least living beings. The same with nations or churches.

The book has over 1000 pages, so I'm just going to sum up the things that may be useful.

1. Variables and stresses

You can point out some numeric that a living system just needs to keep stable. A cell needs to keep its cell barrier intact to prevent the contents to spill out, enough food and oxygen to keep its internal processes running, proper temperature, etc. A corporation needs enough workers on proper positions, liquid assets to pay them and buy materials, enough materials to produce and products to be sold. If one of them goes beyond their normal boundaries it causes a stress. Most, if not all, activity of a living systems is reacting to existing and anticipated stresses.

This is important, because it tells a lot how an organization works. They are pretty inactive unless one of their important variables is, or will be, threatened. Even the most benevolent organizations will eventually evolve to care mostly for their own survival mostly, or dissolve. Moreover, they usually tailor their response to the threat magnitude. Until it happens, they are pretty content to just sit there.

Why don't all organizations just focus on gathering resources? Why are there churches, fund-raising organizations, hobbyist groups? According to the theory it is an example of specialization. An organization is composed of groups of people, so it needs to bring them in somehow. Some of them are created by the society (their suprasystem) to perform some necessary tasks and processes. If they cease to be needed, people will leave them and the society will eventually stop to support them. So churches need to satisfy their members' religious needs, soup kitchens need to feed the poor. Otherwise they are maladapted and won't survive.

In short: just stick to what makes an organization thrive. Ignore the rest.

2. Subsystems

Every living system needs to perform some particular functions to survive, so we can expect them to have some particular subsystems.

Obviously they need to react. Otherwise we have no living system, just a lump of concentrated matter. This is where the decider comes handy.

All living systems need to separate themselves from the outside world a bit, keeping harmful matter-energy and information outside. The boundary does that. on higher levels, from groups upward, it is much more fluid, but it is there. A family won't just let anyone into their house or let their valuables to be taken away. Matter-energy flows in through ingestor, is distributed by distributor, processed by converter and, eventually, producer or stored in matter-energy storage. Wastes and completed products go through the extruder outside.

Information flows in through input transducer or is generated inside with internal transducer, then is translated by decoder into a form comprehensible to the decider.  Channel and net transmit the information, associator and memory make learning possible. Encoder re-translates the information that needs to go out into something else (like speech) and output transducer transmits it outside.

Last, but not least, the reproducer lets the system, well, reproduce.

What's important with these subsystems that they are abstract. The heart, for example, is only a part of the distributor. Lungs act as parts of ingestor (inhaling oxygen), extruder (exhaling carbon dioxide) and converter (binding oxygen with blood cells). Same goes for groups and upward.

Components are probably the most valuable part of this theory. Game AI could be programmed to recognize crucial parts of organizations and act accordingly without hard-coding everything. If a human village needs water to survive, then - in a siege - an obvious solution would be to cut off the stream, or send someone to poison the well. If breaking into the bank vault is too difficult because of the walls, then perhaps entering through the ingestor disguised will be more fruitful. If a nation is bothered by terrorist cells popping out, it needs to learn how they reproduce and prevent that - by arresting potential members or spreading its own propaganda to counter their influence.

There are also various hypotheses about how living systems govern themselves - how power and authority work, what are the most common stresses, how information-processing components distort the information (almost always to maximize their rewards and minimize punishments from the suprasystem), how are the signals encoded, etc. I am currently just walking around the book and thinking, how to best tackle the problem. Matter-energy processes seem pretty simple, but information will need a lot of work.

As all living systems have a genetic code, it seems almost natural to borrow the concept of memes from Dawkins. Decoder would translate gathered data to some simple symbols (hunger, danger, food, peace, etc.), which could be connected by the associator and stored. The decider could make the decision based on symbol being given - an individual who connects "corporation" with "danger", approached by its executive will do something else, than the one who connects it with "money". The main problem is that I'm not really sure how the symbols should interact with each other to even have some semblance of a coherent thought process. I will need to research more.

Still, before I'll get into this, it would be best to - at very least - assess the territory? Anyone here knows this particular theory or can suggest something else where could I look? Or maybe this concept has so many holes that it's not worth to look into?