Genetic Algorithms are part of evolutionary computing that help in solving problems using genetics as a methophor. The GAs have been part of heuristic optimization techniques which includes simulated annealing, tabu search, Ant Colony Optimization etc.
The GA work on simultaneous search for optimal solution in a search space through the metaphor of mutations, reproductions, genetic inversions, cross-over etc. The solution variables are pre-defined and their ranges are also pre-defined which creates the boundaries of the search space. The solution variables are organized as bit-strings that are akin to chromosomes.
At the start of the search - a set of random bit-strings are created (it has to be random) as starting points in a play of evolutionary population simulation. Just like population reproduce, mutate, etc, the solutions population is run through multiple rounds of reproduction etc simulations called generations. Each generation chooses the fittest solutions to create further offsprings. The fitness function acts as the environment where many solutions gets killed in each generation but fittest solutions are propagated to the next geneartion. If one chooses the right fitness function and have proper genetic operators, it is likely that population converges rapidly to the global optimal solution rather than getting stuck in the local optimas.
READER MAY SAY - OK WHATS MY POINT!
Well, Ideation/ problem solving is exactly like this. Can we use the GA metaphor to create multiple starting points (may be random) to converge to solutions through a series of generations - there may be many questions - for example what is the fitness function? How will one represent the idea bit string and how will anyone know the variables upfront?
If we start thinking using TRIZ Inventive Prinicples and TRIZ Trends and combine them with other inventive principles (Vedic Mathematics) and then use the GA engine search for Ideas, I think we have something here!