SEARCH (Random)
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(2) |
| ID | ◀ 2946 ▶ |
| Object type | Methodology or model |
A search without pre-instated rules.
This is supposed to be the way mutations and natural selection do operate.
H. PATTEE observes however: “The first criticism is that random search cannot be expected to 'find' successful organization because the search space is so immense. Thus a search for just one sequence of 100 amino acids representing one functional enzyme would require of the order of 20100 trials which is not within any reasonable probability within the age or size of the universe” (1972, p.38)
H. SIMON proposed his Hora and Tempus metaphor about progress towards evolutive complexity by hierarchization as a possible reply to this riddle (1965).
H. PATTEE pursues: “The second criticism is that functional optima or fitness peaks in the adaptive landscape appear to be local, separate optima, so that no evolutionary pathway can be imagined that does not pass through nonfunctional or lethal valleys. Random search and selection can provide a mechanism for adaptation with respect to one of these optima, but for any fitness landscape with a realistic number of dimensions, the trapping problem becomes unsolvable” (Ibid).
His conclusion is that we need a mechanism of self-simplification to escape from this impossibility. Notably, he considers, like SIMON, that self-simplification must be based on hierarchy.