Object form division method and system

Claim 1

A method, in an object division form system, of generating object division forms into which an inputted form of an object is divided using a genetic algorithm, said method comprising steps, performed by said object division form system, of: providing, as a data structure of a gene, an area form of the inputted form of the object, a division rule number assigned to the gene, and information indicative of at least one other gene generated by further dividing the area form concerning the gene; providing, as a data structure of a chromosome, a tree structure of genes having area forms into which the inputted form of the object is sequentially divided; providing an evaluation rule in which, at least, an evaluation value increases as an internal angle at each node constituting the area form approaches 90°; if the two object division forms selected in accordance with the evaluation values of the object division forms contain area forms that can be exchanged in a state in which the object form can be retained, performing a crossover process of exchanging the area forms so as to generate two new object division forms; performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with any one of prepare division rule for dividing area forms; and performing an end determination process of determining whether the evaluation value for one object division form satisfies a predetermined end determination criterion so as to terminate processing operations, wherein, if the inputted form of the object is a two-dimensional form or a three-dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form is a three-dimensional solid form, a plurality of hexahedron elements are generated based thereon.

Claim 2

A method, in an object division form system, of generating object division forms into which an inputted form of an object is divided using a genetic algorithm, said method comprising steps, performed by said object division form system, of: providing, as a data structure of a gene, an area form of the inputted form of the object, a division rule number assigned to the gene, and information indicative of at least one other gene generated by further dividing the area form concerning the gene; providing, as a data structure of a chromosome, a tree structure of genes having area forms into which the inputted form of the object is sequentially divided; providing an evaluation rule in which, at least, an evaluation value increases as an internal angle at each node constituting the area form approaches 90°; if the two object division forms selected in accordance with the evaluation values of the object division forms contain area forms that can be exchanged in a state in which the object form can be retained, performing a crossover process of exchanging the area forms so as to generate two new object division forms; performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with any one of prepare division rule for dividing area forms; and performing an end determination process of determining whether the evaluation value for one object division form satisfies a predetermined end determination criterion so as to terminate processing operations, wherein, if the inputted form of the object is a two-dimensional form or a three-dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form is a three-dimensional solid form, a plurality of hexahedron elements are generated based thereon.

Claim 4

A system for generating object division forms into which an inputted form of an object is divided using a genetic algorithm, comprising: means for inputting a form of an object; means for displaying at least the object form; storage means for assigning numbers to a plurality of prepared division rules for dividing the object form and storing the division rules with the numbers; area division means for generating N (which is a natural number) object division forms by dividing the object form in accordance with a division rule; evaluation means for finding an evaluation value for each object division form in accordance with a predetermined evaluation rule in which, at least, an evaluation value increases as an internal angle at each node constituting the area form comes close to 90°; end determination means for, if at least one of the found evaluation values satisfies a predetermined division end determination criterion, determining that the object division form corresponding to the evaluation value is an optimum area division form resulting from dividing the object form into areas; selection means for, if none of the evaluation values satisfy the predetermined division end determination criterion, selecting two among the N object division forms in accordance with the evaluation value; crossover means for, if the divided area forms that can be exchanged in a state in which the object form can be retained exist, performing a crossover process of exchanging the area forms to generate new object division forms until N object division forms are generated; mutation means for performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with one division rule arbitrarily selected among the division rules stored in said storage means; and start control means for starting said selection means, said crossover means and said mutation means until the evaluation value for one object division form satisfies the division end determination criterion, wherein, if the inputted form of the object is a two dimensional form or a three dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form thereof is a three-dimensional solid form, a plurality of hexahedron elements are generated based thereon.

Claim 5

A system for generating object division forms into which an inputted form of an object is divided using a genetic algorithm, comprising: means for inputting a form of an object; means for displaying at least the object form; storage means for assigning numbers to a plurality of prepared division rules for dividing the object form and storing the division rules with the numbers; area division means for generating N (which is a natural number) object division forms by dividing the object form in accordance with a division rule; evaluation means for finding an evaluation value for each object division form in accordance with a predetermined evaluation rule in which, at least, an evaluation value increases as an internal angle at each node constituting the area form comes close to 90°; end determination means for, if at least one of the found evaluation values satisfies a predetermined division end determination criterion, determining that the object division form corresponding to the evaluation value is an optimum area division form resulting from dividing the object form into areas; selection means for, if none of the evaluation values satisfy the predetermined division end determination criterion, selecting two among the N object division forms in accordance with the evaluation value; crossover means for, if the divided area forms that can be exchanged in a state in which the object form can be retained exist, performing a crossover process of exchanging the area forms to generate new object division forms until N object division forms are generated; mutation means for performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with one division rule arbitrarily selected among the division rules stored in said storage means; and start control means for starting said selection means, said crossover means and said mutation means until the evaluation value for one object division form satisfies the division end determination criterion, wherein, if the inputted form of the object is a two dimensional form or a three dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form thereof is a three-dimensional solid form, a plurality of hexahedron elements are generated based thereon.

Claim 7

A method, in an object division form system, of generating object division forms into which inputted form of an object is divided using a genetic algorithm, said method comprising steps, performed by said object division form system, of: providing, as a data structure of a gene, an area form of the inputted form of the object, a division rule number assigned to the gene, information indicative of at least one other gene generated by further dividing the area form concerning the gene; providing, as a data structure of chromosome, a tree structure of the genes having area forms into which the inputted form of the object is sequentially divided; if the two object division forms selected in accordance with the evaluation values of the object division forms contain the area forms that can be exchanged in a state in which the object form can be retained, performing a crossover process of exchanging the area forms so as to generate two new object division forms; and performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with any one of prepared division rules for dividing area forms, wherein, if the inputted form of the object is a two dimensional form or a three dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form thereof is a three dimensional solid form, a plurality of hexahedron elements are generated based thereon.

Claim 8

A method, in an object division form system, of generating object division forms into which inputted form of an object is divided using a genetic algorithm, said method comprising steps, performed by said object division form system, of: providing, as a data structure of a gene, an area form of the inputted form of the object, a division rule number assigned to the gene, information indicative of at least one other gene generated by further dividing the area form concerning the gene; providing, as a data structure of chromosome, a tree structure of the genes having area forms into which the inputted form of the object is sequentially divided; if the two object division forms selected in accordance with the evaluation values of the object division forms contain the area forms that can be exchanged in a state in which the object form can be retained, performing a crossover process of exchanging the area forms so as to generate two new object division forms; and performing a mutation process of dividing at least one divisible area form constituting a part of the object division form in accordance with any one of prepared division rules for dividing area forms, wherein, if the inputted form of the object is a two dimensional form or a three dimensional shell form, a plurality of quadrangle elements are generated based thereon, and wherein, if the inputted form thereof is a three dimensional solid form, a plurality of hexahedron elements are generated based thereon.