Difference between revisions of "Group 2"
From Self-Organization Wiki
| Line 4: | Line 4: | ||
** Non-trivial but approaches exist. | ** Non-trivial but approaches exist. | ||
* '''Chaos nature''': Stochastic input. Description of complex current state is impossible, which makes it impossible to predict future states. | * '''Chaos nature''': Stochastic input. Description of complex current state is impossible, which makes it impossible to predict future states. | ||
| − | |||
* Malicious nodes, faults, defects | * Malicious nodes, faults, defects | ||
* Rare events may lead to major global effects. Repeatability of results | * Rare events may lead to major global effects. Repeatability of results | ||
| Line 13: | Line 12: | ||
* "Immune system" as an umbrella around self-organizing system | * "Immune system" as an umbrella around self-organizing system | ||
* Optimize probability to end up in desired state | * Optimize probability to end up in desired state | ||
| + | |||
| + | ===Side note=== | ||
| + | * Heylighen: "interaction pattern are not specified" | ||
Revision as of 12:01, 12 July 2010
Challenges
- General:
- How to design local rules achieving the desired global properties?
- Non-trivial but approaches exist.
- Chaos nature: Stochastic input. Description of complex current state is impossible, which makes it impossible to predict future states.
- Malicious nodes, faults, defects
- Rare events may lead to major global effects. Repeatability of results
- Modeling and design
- Testing a proposed self-organizing system with respect to a given goal can be difficult (many entities, large operational range)
Ideas
- "Immune system" as an umbrella around self-organizing system
- Optimize probability to end up in desired state
Side note
- Heylighen: "interaction pattern are not specified"
