Genetic Options
Previous  Top  Next

Max Generations
This controls the maximum number of generations that Optimax will record. Once this number is reached, if the number of permutations specified in TradeStation has not been reached, the strategy will "spin," meaning that the remaining permutations in TradeStation will execute over and over using the same parameters as the final individual in the final generation. During this phase the OMX_InitPass call will return 0 and so no further results will be send to Optimax.

Tournament Group Size
Tournament Group Size is described here.

Roulette Scaling Factor
Roulette Scaling Factor is described here.

Initial Population Size
The number of individuals in the first generation, and in all subsequent generations. The number of individuals in a generation determines the maximum possible population diversity, and thus the probability of searching all features within the entire fitness landscape.

The greater number of permutations and inputs, the larger your initial population needs to be to cover the entire fitness landscape successfully without getting stuck on local maxima.

Maximum Life Span
Maximum Life Span is described here.

Elite Percent Taper
Elite Percent Taper is described here.

Min nHamming Distance
Min nHamming Distance is described here.

Top Percent
Top Percent is described here.

Crossover Rate
A fractional value between 0 and 1 controlling the frequency with which chromosomal crossover occurs during mating. When set to 1, crossover occurs for every chromosome in the mating operation. When set to 0, crossover never occurs. At 0.5, crossover occurs half the time. In nature, crossover almost always occurs, so a value very close to 1 is normal, for example, between .90 and .99. For those chromosome operations where crossover doesn't occur, Optimax simply copies the chromosome from either the mother or father; the donating parent is chosen at random.

Mutation Rate
A fractional value between 0 and 1 controlling the rate at which mutation occurs during a chromosomal crossover operation. When set to 0, mutation never occurs. When set to 1, every gene in a chromosome will mutate. In nature, mutation occurs at a very low frequency. A small amount of mutation is beneficial to the evolutionary process, causing Optimax to discover fitness terrain features it might not otherwise find. A high probability of mutation is destructive to the mating process and will prevent convergence from ever occurring.

Genetic Complement Number
The number of genetic complements that will be generated in each generation. A very small number of complements benefit the evolutionary process by increasing the genetic diversity of the mating pool in quantum increments. A large proportion of genetic complements in a generation cause chaos during mating.

The genetically engineered complement's fitness is unknown, so Optimax gives it a fighting chance by always assigning to it the median fitness of the current population.

A value of 1 is common.

Convergence Length
The number of generations over which convergence is measured.

Convergence Std Dev
The standard deviation value used when calculating convergence.