Distribution generation system, and optimization system that adopts distribution generation system

Claim 2

A distribution generation system comprising: a plurality of units each of which includes condition storage means for storing a corollary condition, and condition transition means for updating the condition of said condition storage means, wherein in each of said plurality of units, said condition transition means causes a plurality of stochastic processes per unit time for causing the condition of said condition storage means to transit to a predetermined condition to approximately continuous, so that the corollary condition is converged to random variables having a predetermined distribution.

Claim 3

A system according to claim 1, wherein said condition transition means includes means for forming a distribution which is dense near a point indicating the condition stored in said condition storage means, and is sparse at positions far from the point.

Claim 4

A system according to claim 2, wherein said condition transition means includes means for converging the distribution of the random variables to a desired distribution by repeating probabilistic transitions according to the corollary condition.

Claim 5

A system according to claim 2, wherein said condition transition means includes means for generating γ -n -sampling.

Claim 6

A system according to claim 1, wherein each of said plurality of units includes memory means for holding the corollary condition from other units.

Claim 7

A system according to claim 5, further comprising: a transfer device for connecting said condition storage means and said memory means of said plurality of units.

Claim 8

A system according to claim 1, wherein said plurality of units constitute a P-adic tree (where P is an integer not less than 2).

Claim 9

A system according to claim 1, wherein said plurality of units are coupled in a ring shape.

Claim 10

An optimization system comprising: random variable output means for time-serially outputting random variable values having a predetermined distribution; and candidate of optimized solution update means for holding a candidate of an optimized solution, said candidate of optimized solution update means including means for performing optimized solution searching while updating the candidate of the optimized solution on the basis of the random variable values output from said random variable output means.

Claim 11

An optimization system comprising: random variable output means for time-serially outputting random variable values having a predetermined distribution; and candidate of optimized solution update means for holding a candidate of an optimized solution, said candidate of optimized solution update means including means for performing optimized solution searching while updating the candidate of the optimized solution on the basis of the random variable values output from said random variable output means.

Claim 12

A system according to claim 9, wherein said random variable output means includes means for outputting a distribution in which the random variables are distributed densely near a center of the distribution, and are distributed sparsely in a direction to be separated from the center.

Claim 13

An optimization system comprising: distribution generation means for time-serially outputting random variable values having a predetermined distribution; parameter determination means for properly determining various parameters to be supplied to said distribution generation means; and searching means for holding a candidate of an optimized solution, said searching means including means for performing optimized solution searching while updating the candidate of the optimized solution on the basis of the random variable values output from said distribution generation means.

Claim 14

An optimization system comprising: distribution generation means for time-serially outputting random variable values having a predetermined distribution; parameter determination means for properly determining various parameters to be supplied to said distribution generation means; and searching means for holding a candidate of an optimized solution, said searching means including means for performing optimized solution searching while updating the candidate of the optimized solution on the basis of the random variable values output from said distribution generation means.

Claim 15

A system according to claim 11, wherein said distribution generation means includes means for outputting a concentrated sampling distribution in which the random variables are distributed densely near a center of the distribution, and are distributed sparsely in a direction to be separated from the center.

Claim 16

A system according to claim 12, wherein said distribution generation means includes means for repeating probabilistic transitions using the concentrated sampling distribution.

Claim 17

An optimization system comprising: parameter value storage means for storing a parameter value; physical quantity storage means for storing a physical quantity corresponding to the parameter value stored in said parameter value storage means; parameter value sampling means for executing concentrated sampling for densely sampling values near the parameter value stored in said parameter value storage means, and sparsely sampling values far from the parameter value, and for outputting sample parameter values; physical quantity fetching means for outputting the sample parameter values generated by said parameter value sampling means to an external device, and fetching physical quantities corresponding to the sample parameter values from said external device; and parameter value update means for updating the parameter value stored in said parameter value storage means according to a comparison result between the physical quantities corresponding to the sample parameter values fetched by said physical quantity fetching means, and the parameter value stored in said parameter value storage means.

Claim 18

An optimization system comprising: parameter value storage means for storing a parameter value; physical quantity storage means for storing a physical quantity corresponding to the parameter value stored in said parameter value storage means; parameter value sampling means for executing concentrated sampling for densely sampling values near the parameter value stored in said parameter value storage means, and sparsely sampling values far from the parameter value, and for outputting sample parameter values; physical quantity fetching means for outputting the sample parameter values generated by said parameter value sampling means to an external device, and fetching physical quantities corresponding to the sample parameter values from said external device; and parameter value update means for updating the parameter value stored in said parameter value storage means according to a comparison result between the physical quantities corresponding to the sample parameter values fetched by said physical quantity fetching means, and the parameter value stored in said parameter value storage means.

Claim 19

A system according to claim 14, wherein said parameter value sampling means includes means for repeating probabilistic transitions using the concentrated sampling.