| CPC G06Q 40/06 (2013.01) [G06N 3/126 (2013.01)] | 17 Claims |
|
1. A computer implemented method for automatically selecting a portfolio where said portfolio comprises a population of solutions, the method comprising the steps of:
generating, by a processor, a plurality of first generation portfolios from a plurality of available solutions, wherein each of said plurality of first generation portfolios is generated by random selection of one or more of said plurality of available solutions;
applying, by the processor, a random percentage weighting to each of said randomly selected solutions, but where the sum of the random percentage weightings is arranged to equal 100%;
calculating, by the processor, a score for each of the plurality of first generation portfolios, the score for each of said plurality of first generation portfolios of assets comprising an aggregate of a plurality of different evaluation factors;
ranking, by the processor, the plurality of first generation portfolios by their respective scores; and
selecting, by the processor, a plurality of said first generation portfolios as parent portfolios based on their respective scores and using said parent portfolios to generate a plurality of child portfolios,
wherein some of the plurality of parent and/or child portfolios are selected to undergo a mutation process whereby, for a selected portfolio, at least one asset of said portfolio is randomly selected and all or a random percentage weighted portion of said randomly selected asset is replaced by another asset randomly selected from a union of assets comprising a group of all of assets of existing portfolios and a group of other eligible assets and
wherein the parent and/or child portfolios selected to undergo the mutation process are selected by assigning each of said parent and/or child portfolios with a binary value according to a pre-set mutation rate, where the assigned binary value determines whether a portfolio is selected for the mutation process or not.
|