3 ms·
While I like the connection you made between your self-adaptive system and self-adaptive systems by Mother nature, I couldn't see the analogy beyond the fact th
by mck- 14y ago
While I like the connection you made between your self-adaptive system and self-adaptive systems by Mother nature, I couldn't see the analogy beyond the fact that they both use the word 'self-adaptive'.
>> These systems are also decentralized and consist of a number of lower level primitive components, collaboration of which on a higher level produces very complex and intelligent behaviors.
After reading the first paragraph, I was getting excited and put my Emerging Behavior thinking hat on -- Particle Swarms, Ant Colony Optimization algorithms, or even Wolfram's New Kind Of Science and The Game of Life -- only to find out that the system was a if-then rule-based program?
>> While in most biological systems all of those rules are written in cell’s DNA, for a software system I needed to find a good framework for defining those.
Fancy stuff, the likes of Genetic Algorithms, Evolutionary Systems, Memetics? But no... Not those kind of self-adapting systems..
Am I missing something?
PS: Fuzzy control and Neural Nets sound like a whole different ball-game to me ;)
But hey, good stuff nonetheless -- thanks for sharing and glad to hear it was well-received
- avalanche123 14y agoThanks for feedback. Indeed, complex emergent behaviors are not demonstrated in the provided example. But they are possible when changes in environment affect other control loops. Which might result in conflicts, at which point control loop for conflict resolution can be added, this is why MAPE-K tower is a useful abstraction.