parallelismConfig

Creates ParallelismConfig with ParallelismConfigScope and apply it to the current PanmicticConfig.

Panmictic Parallelism is based on the idea of dividing PanmicticGA stages (selection, crossover, mutation, evaluation etc.) into independent processes that can be executed in parallel mode. That's why it is called also Genetic Operator Parallelism - parallelism is used at the level of the genetic operator, transitions between stages occur sequentially.

There cannot be such a situation that some chromosomes go through the crossover stage, and other mutation stage - the entire population can be under the influence of only one operator at one time.

Base selection, crossover, mutation, evaluation stage functions support Panmictic Parallelism by default using process functions - it's safe to use!

NOTE In most cases, the parallelism overhead for the selection stage exceeds its efficiency, so select operators have parallel mode disabled by default. Set parallelismLimit to non-zero value to make parallelism available.

Example:

pGA {
    // configure specific params of PanmicticGA

    // configure Panmictic parallelism
    parallelismConfig {
        workersCount = 5 // max count of parallel coroutines
        dispatcher = Dispatchers.Default
    }

    evolve {
        // selection stage NO_PARALLELISM by default
        selTournament(size = 3)
        // crossover stage turn off parallelism by 0 NO_PARALLELISM
        cxOnePoint(chance = 0.8, parallelismLimit = 0)
        // mutation stage parallelism available (workersCount limit)
        mutFlipBit(chance = 0.1, flipBitChance = 0.01)
        // evaluation stage parallelism also available
        evaluation()
        stopBy(maxIteration = 50) { bestFitness == 100 }
    }
}

See also