SU999144A1 - Random number generator - Google Patents

Description

(54) GENERATOR OF RANDOM NUMBERS

The invention relates to computing and can be used in stochastic computing and computing devices. A random number generator is known which contains a random number sensor, a reminder device, comparison elements, registers, AND elements, OR elements, counters, a control unit. However, the generator allows to obtain a correlation function different from the delta function, but does not allow get random numbers from the control of the autocorrelation function The closest in technical essence to the invention is a random number generator containing random number sensors, a memory device, elements, compare, registers, cients AND, OR elements, counters, control unit allows kschy receive the random sequence of a certain autocorrelation function unchanged probability distribution law disadvantage generator il 2 t ete that has low precision forming and reduced functionality. . The aim of the invention is to increase the accuracy and enhance the functionality by generating random numbers with a controlled autocorrelation function. The goal is achieved by the fact that a random number generator that contains a random number sensor, two blocks of registers, a comparison unit, a device and a control unit, whose outputs are respectively connected to the control inputs of a random number sensor, two blocks of registers and a comparison element block, the first group of inputs of which is connected to the bit outputs of the first block of registers, an input device, a multi-channel correlator, are introduced ;; the meter and the aforementioned device, whose outputs are connected to the first group of memory inputs, the second group of inputs of which are connected to the outputs of a random sensor, numbers, to the outputs of the Memory device are connected to separate inputs of the second block of registers, the outputs of which are connected directly to the first group of inputs an arithmetic unit and a multi-channel correlometer - from the second

a group of inputs of a unit of comparison, the first group of inputs and outputs of which are connected to the second and third groups, respectively, of the inputs of the arithmetic device, and the additional outputs of the control unit are connected to the control inputs of the storage device, the arithmetic device, the multi-channel correlometer and the input device whose outputs are connected to inputs of the first block of registers.

The drawing shows a functional diagram of a random number generator.

The generator contains a random number sensor 1, a control unit 2, a memory 3, which is built on the principle of associative memory, registers 4 and 5, a multi-channel correlometer b, a comparison elements block 7, an input device 8, an arithmetic unit 9. The drawing also shows tires 10, which receive the values of the autocorrelation function, and output tires 11.

Control unit 2 is designed to control the process of generating random numbers in sensor 1 random numbers:; ate, sampling and writing numbers to memory 3, writing and shifting numbers in block 4 registers, writing specified autocorrelation coefficients to block 5 registers, and to control the processing of signals entering the arithmetic unit 9.

The storage memory & 3 is intended for Recording, storing 1, Pack-digit numbers, as well as K-bit sampling samples of one of them. The search K-bit part of each number stored in memory 3 is also K-most significant number of digit numbers.

The storage device 3-works in the search mode for the nearest number.

Register block 4 is designed to store (.41) 3 W-bit numbers sequentially selected from a memory device and is a MIND shift registers each of .4 length 1) bit. Sloxes (.4 1) bit each shift register are generator output.

The register block 5 is intended to store C numbers, each of which. corresponds to the specified autocorrelation coefficients.

The multi-channel correlometer b is designed to measure the autocorrelation coefficients of a sequence of numbers selected from memory 3.

Block 7 Comparison Elements Comp. The measured autocorrelation coefficients are given with specified values.

The input device 8 is intended for writing registers of specified autocorrelation coefficients into block 5.

The arithmetic unit 9 implements an algorithm for extrapolating stationary random processes according to a given autocorrelation function while maintaining the one-dimensional distribution law of the original random process. The generator works as follows. .

From the sensor 1, the initial random sequence enters the memory 3, fills it and the block 4 of registers. Time-shifted values of the sequence X from block 4 arrive at the inputs of the multichannel correlometer b, which goes to the measurement mode of the current values of k; / KjJC ,, .., Kg. In block 7, the values of K, K,.,., K and K ..., K are received, arriving respectively from blocks 5 and 6, and the formation of difference signals Dd, A, ,, ,, Lei Zp,, 2 g .-. Arithmetic unit 9, using the values of prehistory x, x., ..) xg, (hbH) from block 4, ass: 1nn values of the autocorrelation function from block 5 taking into account the difference signals from block 7 forms a sequence of polling signals X which enter into their The queue of steps in; 1 of the polling signs of the memory device Z. At the output of which the numbers of the output sequence appear, each The one of which is the number of the original sequence from sensor 1 and is minimally different from the corresponding number of the sequence X.

Due to this organization of operation of the device blocks, the values of the autocorrelation function of the output sequence X Kp, K, n,, Kg are almost the same as the values of Kp, K,.

The feedback loop of the memory device 3 and the block 4, formed by the correlometer 6, the block 7 and the arithmetic unit 9, makes it possible to realize the principle of automatic control. For this, it is necessary to use differential signals AO, D, .UL so that feedback is negative.

Claims (2)

Thus, the proposed generator makes it possible to form any preassigned autocorrelation functions (which may exist for a given distribution law) and it is possible to form not only stationary random sequences with a definite autocorrelate cirrhotic function, but also random sequences that are unsteady by autocorrelation. Claims A random number generator comprising a slot number sensor, two register bins, a comparison unit, a memory and a control unit, the inputs of which are respectively connected to the control inputs of the random number sensor, two register units and a comparison unit, the first group of inputs which is connected to the bit outputs of the first block of registers, characterized in that, in order to increase the accuracy and expand the functional capabilities, an input device, a multichannel, is introduced into it relometr first armature and an arithmetic unit which outputs are connected to the first group of memory inputs, a second set of inputs of which is connected to the outputs of the random number, and the outputs of the memory : fi devices: connected to the bit inputs of the second block of registers, output-. The ports of which are connected directly to the first group of inputs of the arithmetic unit, and through a multi-channel correlometer to the second group of inputs of the unit of comparison elements, the first group of inputs and outputs of which are connected respectively to the second and third groups of inputs of the arithmetic unit, and the additional outputs of the control unit are connected to the control inputs an e-memo device, an arithmetic unit, a multi-channel correlometer, and. . input devices whose outputs. The keys are to the inputs of the first register block. Sources of information taken for review during the examination 1. Authorship certificate of the USSR 370601, cl. G 06 F 1/02. 2. Authors certificate of the USSR 516042, cl. G 06 F 15/96.