Polyoptimizing genetic algorithm for finding multiple solutions to problems

摘要

A method and system for finding a multiple solutions to a problem is disclosed. The method includes: modeling a problem as an objective function, generating a multitude of candidate solutions for said objective function; and determining a set of acceptable and distinct solutions. The objective function is capable of mapping structures in a search space. The search space includes a multitude of points. For each of said candidate solutions at each of said multitude of points, the set of acceptable and distinct solutions are determined by: calculating a fitness using the objective function; assigning each fitness to its candidate solution; transforming the fitness using a penalty function; determining if the candidate solution is acceptable; determining if the candidate solution is sufficiently distinct; and adding the candidate solution to the set of acceptable and distinct solutions if the candidate solution was determined to be acceptable and distinct.