US 8,429,104 B2 · Concurrent two-phase completion genetic algorithm system and methods

Abstract

A genetic algorithm architecture implements a two-stage completion genetic algorithm with respect to an evolving current population data set. The two-stage completion genetic algorithm that includes genotype and phenotype completion loops. The genotype completion loop operates to compete the current population data set based on genotype field fitness scores. The genotype completion loop also implements a phenogenesis operator used to generate a current phenotype set. The phenotype completion loop operates, concurrently with the genotype completion loop, to evaluate the current phenotype set, constrained relative to the current population data set, against a fitness function to produce phenotype fitness scores. The phenotype completion loop implements a genotype reduction operator that then determines corresponding genotype fitness scores for use as the basis for competition in the genotype completion loop.

Claims20

Dependent claims

A method, executable as a data analysis process on a computer system, for determining an optimal solution set for a population data set using a defined fitness model, said method comprising the steps of:

a) receiving a predefined population data set and a predefined fitness model;
b) first competing, using a first genetic algorithm loop, genotype representations of phenotypes within a source population data set initially derived from said predefined population data set;
c) second competing, using a second genetic algorithm loop using said predefined fitness model, phenotypes of said source population data set, wherein a current generation of genotype representations determines a current set of phenotypes that will compete in said current generation of said second competing step, and wherein said current generation of said second competing step determines genotype fitness scores to be used in said first competing step for said current generation of genotype representations;
and
c) providing said current set of phenotypes as an optimal solution set dependent on closure of said second competing step as determined using predetermined closure criteria.

2 more elements

A computer system implemented data analysis process for determining an optimal solution set relative to a given population data set and fitness model, said computer system implemented data analysis process comprising the steps of:

a) receiving, from a control processor, a source population data set;
b) processing said source population data set through a two-stage completion genetic algorithm wherein a current population data set, initially derived from said source population data set, is transformed through
i) a genotype completion loop, including a phenogenesis operator, wherein genotypes of said current population data set are competed based on genotype field fitness scores, said genotype completion loop providing for a cyclic revision of said current population data set;
and
ii) a phenotype completion loop, including a genotype reduction operator, wherein a current phenotype set is generated by said phenogenesis operator relative to said current population data set, wherein a fitness operator evaluates a predetermined fitness function against each phenotype of said current phenotype set to produce phenotype fitness scores, wherein said genotype reduction operator determines genotype fitness scores from phenotype fitness scores, and wherein said current phenotype set is determined an optimal solution set based on predetermined closure parameters;
and
c) providing, to said control processor, said determined optimal solution set.

4 more elements

Dependencies are inferred from the claim text.

Concurrent two-phase completion genetic algorithm system and methods

Life of this patent

Concurrent two-phase completion genetic algorithm system and methods

Life of this patent

Abstract

Claims20