SU1160407A2 - Controlled random number generator - Google Patents

Abstract

CONTROLLED SENSOR RANDOM NUMBERS of aut. St. No. 690469, characterized in that, in order to expand the functionality of the sensor by providing reproduction with changing expectation of a given distribution function, it contains a power raising unit, a multiplier, a HE element, a switch and an adder, the first input of which is connected to the second output a sensor whose timestamp input is connected to the first input of a raising unit, the second input of which is the input of the first sensor constant, the input of the second constant of which is connected to the first The input of the multiplier, the second input of which is connected to the output of the exponentiation unit, and the output of the multiplier is connected to the input of the switch, the first output of which is connected to the second input of the adder, the output of which is the third output of the sensor, the second output of the switch (L body via the element is NOT connected to the second input of the adder.

Description

cr:

R

O The invention relates to computing and can be used in statistical modeling. On the main auth, sv. No. 690469 is known to control a random number sensor that contains an input block for the distribution function, the outputs of which are connected to the first group of inputs of the parallel number comparison block, the second group of inputs of which are connected to the generator output of uniformly distributed random numbers, and the output is connected to the single inputs of the trigger block, zero the inputs of which are connected to another output of the generator evenly distribute random numbers, and the outputs - to the outputs of the block of elements And, respectively, the outputs of which are connected to the inputs of the output unit, whose other inputs are connected to the output of the quantizer, whose output groups are connected to the first group of inputs of the parallel number comparison block, respectively, the first input of the quantizer is combined with the first input of the differentiation unit and connected to the first auxiliary output of the input function of the distribution function, and the second input to output divider, the first input of which through the shift register is connected to the second additional output of the input block of the distribution function, and the second input of the divider is connected to the first output of the bl Single memory, a second output connected to the second input of differentiating unit, the first output of which is connected to the first input of the memory and the second output - to an input of a maximum determination unit, which soedsh-ieH output to the second input of the memory block p. However, the known sensor does not allow modulation with changing characteristics of the original distribution function, like the mathematical expectation. The aim of the invention is to enhance the functionality of the sensor by providing reproduction with changing waiting times for the distribution function. The change in mathematical expectation is expressed in the shift of the abscissa of the distribution function by the magnitude of X, whose change in time j can be reduced in the form of D X uf t, de ct and b are some constants. The power dependence (1) describes a very wide class of functions. To achieve this goal, the controlled random number sensor further comprises a power exponent, a multiplier, a NOT element, a switch and an adder, the first input of which is connected to the second output of the sensor, the timestamp input of which is connected to the first input of the power raising unit, the second input of which is the input of the first sensor constant, the input of the second constant of which is connected to the first input of the multiplier, the second input of which is connected to the output of the raising unit, and the output of the multiplier is connected to the input of the switch, the first output of which is connected to the second input of the adder, the output of which is the third output of the sensor, the second output of the switch through the element NOT connected to the second input of the adder. The drawing shows the block diagram of the proposed sensor. Sensor Contains block 1 for inputting the distribution function, block 2 for parallel comparison of numbers, block 3 for differentiation, quantizer 4, shift register 5, block 6 for memory, block 7 for determining maximum, divider 8, generator 9 for uniformly distributed random numbers, triggers Yu, and elements 11, the output unit 12, the adder 13, the switch 14, the element 15, the multiplier 16, and the exponential unit 17. The sensor works as follows. The continuous values of the given distribution function from the input unit 1 are transferred to the differentiation unit 3 and the quantizer 4. From unit 3, the values of f (y) are fed to memory unit 6, from which they return to the second input of unit 3, the second output of which is removed derivative f () and. sent to block 7 for determining the maximum. This block determines the highest value of the second derivative of the given distribution function and the corresponding abscissa. According to the signal from block 7, memory block 6 outputs to the input of divider 8 the value of the first derivative corresponding to the section of the greatest curvature of the given distribution function. (1a, another divider input is multiplied by two using shift register 5, the value of the permissible reproduction error of a given distribution function. This value is output to shift register 5 from the output of input block 1. The divider calculates the sampling interval and supplies its value to quantizer 4, which performs discretization of the continuous distribution function with the calculated interval and outputs discrete values of the given distribution function to the first inputs of the unit 2 parallel comparison of numbers and sign Arguments of nodal points to the second inputs of the output block 12. V. clock points from the generator of 9 uniformly distributed random numbers receive random numbers to the second inputs of block 2 parallel comparison, numbers. As a result, the random number is compared with the numbers removed from the input block , only the part of the output buses of the block is excited, and due to the hf onotonicity of the increase in the distribution function between the excited and Unexcited parts of the outputs of block 2, there is a single limit that is determined. the triggers 10 and the elements of the 11. Therewith, of all the elements, only the ONE determining „„ „, part of the distribution function is open, within which a random uniformly distributed number appeared at a given time. The outputs of the And 11 elements excite the corresponding elements of the output block 12, and a random number appears at the output of the controlled sensor, subject to the law with unchanged expectation. As elements of the output block, registers can be used in which the values of the arguments of the nodal points of the distribution function are written. Simulate random numbers are also applied to the input of adder 13 to take into account the change in the expectation of a given distribution law. The heating unit 17 and the multiplier 16 estimate the degree of change in mathematical expectation in accordance with formula (1) (the values of b and t come from the sensor input). The shift x is sent to the input of the transcuser 14. If the expectation increases, then the signal corresponding to the shift from the switch output is fed to the second input of the adder 13, where some shift is added to the modeled random numbers. When the mathematical expectation decreases, the shift value should be negative, so in this situation the signal corresponding to the shift is sent from the second output of the switch 14 through the NOT 15 element (to change its sign) to the input of the adder t3. Thus, random numbers from the output of the adder to the output of the sensor correspond to a given distribution function with a mathematical expectation varying over some dependence. In the case of a discrete representation of the initial distribution function, the operation of the controlled sensor occurs in a similar way. The difference lies in the fact that in this situation, blocks 3–8, which perform the specification of the continuous distribution function, do not work. To increase the digit size of the output numbers, a part of the lower order of the generator of uniformly distributed random numbers can be used. The indicated sensor can also be used as a controllable probable AND pole. The invention allows to model. random numbers when changing the mat, "tic"; Wait for a given distribution function over a certain dependence. Such a problem arises, for example, when simulating random values of the resistance i (properties) of the object’s technical resistance to external effects, so the jcafc of the object’s resistance during operation decreases (aging occurs. Thus, the error of the statistical The economic effect of the use of H3o6jpeTeHHH in the national economy can be estimated by the amount of preventable damage that may occur due to incorrect conclusions based on statistical results. th .modelirovani because of its significant errors due to neglect of variations expectation predetermined distribution function,

8bo8

dpnnyzh

R

--1I - III

7G | I 1g

ivar

fit N

Well ::

15

 f7

H-jl

± 217

Claims (1)

CONTROLLED SENSOR OF RANDOM NUMBERS by ed. St. No. 690469, characterized in that, in order to expand the sensor’s functionality by providing reproduction with a variable expectation of a given distribution function, it contains an exponentiation block, a multiplier, a NOT element, a switch and an adder, the first input of which is connected to the second the output of the sensor, the time stamp input of which is connected to the first input of the exponentiation block, the second input of which is the input of the first sensor constant, the input of the second constant of which is connected to the first input a multiplier, the second input of which is connected to the output of the exponentiation unit, and the output of the multiplier is connected to the input of the switch, the first output of which is connected to the second input of the adder, the output of which is the third output of the sensor, the second output of the switch through the element is NOT connected to the second -L eye the adder input. w II Π 60407 2