RU2780800C1 - System for optimising the frequency allocation in a set of compactly arranged diverse radioelectronic tools - Google Patents

Abstract

FIELD: computing equipment.

SUBSTANCE: invention relates to the field of computing equipment and telecommunications and can be used for resource allocation in the event of conflicting resource combinations. System for optimising the frequency allocation implements an adaptive genetic algorithm with rank-proportional selection for assigning frequency sets to radioelectronic tools based on measuring the power of mutual interference between the radioelectronic tools, the radio scanning unit, and the corresponding frequency activation/deactivation control unit.

EFFECT: optimised frequency resource allocation between the radioelectronic tools of a set of compactly arranged diverse radioelectronic tools with minimised mutual interference.

1 cl, 4 dwg

Description

FIELD OF TECHNOLOGY

The invention relates to the field of computer technology and telecommunications and can be used to allocate resources in the presence of conflicting combinations of resources, in particular, to distribute frequencies in a complex of compactly located heterogeneous radio electronic equipment and minimize interference that occurs when using close frequencies of radio signals of compactly located radio electronic equipment.

Consider the resource allocation problem in the following formulation. The complex of electronic means contains several subjects, which, for example, can be used as a radio transmitter or radar. An example of a resource would be a radio signal in a given frequency band, where each entity may use a plurality of valid resources - for example, operate in different frequency bands.

Resources, due to their physical, material nature, conflict with each other and create mutual interference. For example, radio signals in adjacent frequency bands have mutual influence on each other, and in the case of a compact arrangement of radio electronic means, specific quantitative values of mutual influence are determined by measuring the parameters of specific radio signals over a set of subjects and over a set of frequencies, since the degree of mutual influence depends both on the location of the subjects and and on the frequency parameters of the corresponding signals.

In practice, the problem of choosing such an option for distributing resources is relevant, in particular, assigning frequency ranges to radio electronic means, in which each subject receives at least one frequency range from its set of allowable resources and minimizes the total level of mutual interference in a complex of compactly located heterogeneous radio electronic means.

BACKGROUND OF THE INVENTION

From the prior art D1 (William K. Hale. Frequency Assignment: Theory and Applications. - Proceedings of the IEEE, Vol. 68, No. 12, December 1980), a method for solving the problem of frequency resource allocation is known in which the frequency assignment problem is formulated as a search problem mapping a subset of the plane representing the possible locations of radio electronic means into a set of natural numbers representing the assigned frequencies, and it is shown that the indicated problem is NP-complete, which leads to the fact that a strictly optimal frequency assignment is possible only for a small number of radio electronic means, which does not represent practical interest.

In view of the above, in practice, quasi-optimal methods are used to allocate resources. In particular, the source D2 (patent RU 2291459) considers a system for protecting pulsed radar stations from active noise interference, which uses a genetic algorithm to control digital phase shifters. The genetic algorithm used in D2 uses an output power vector that is ranked by increasing interference output power, and a ranked output power matrix is constructed accordingly. The next step is to discard half of those strings that correspond to the maximum output power, and then two strings are randomly selected as "parents" to create two code strings by exchanging a part of the codes, and the amount of the part of the codes is chosen at random.

Thus, the applied genetic algorithm is not adaptive, it uses a mechanical approach (rejection of half of the lines) and random generation of new code lines.

Due to the presence of these shortcomings, D2 does not achieve a stable result of suppressing active noise interference, which is random in nature, especially if there are several interferences in the same interference situation.

The claimed invention is aimed at eliminating the above disadvantages.

DISCLOSURE OF THE INVENTION

The technical result of the claimed invention is the implementation of a method and a system that optimizes the distribution of frequency resources between radio electronic means of a complex of compactly located heterogeneous radio electronic means and ensures the minimization of mutual interference in the complex by using an adaptive genetic algorithm with rank-proportional selection for assigning sets of frequencies.

The claimed technical result is achieved due to the claimed method for optimizing the frequency distribution in a complex of compactly located radio-electronic means of various types, comprising the steps of which:

- carry out scanning of the air and measurement of the power of mutual interference occurring between electronic means, determine the size of the population n;

- measure the frequencies used by radio-electronic means of the specified complex;

- set the coefficient of rank selection α;

- for each individual, a binary random vector of frequency assignment V _i , i=1...n with a binomial distribution Bi(1, P ₁ ), where P ₁ is the frequency resolution probability in the initial population;

- using the measured powers of the mutual interference of the complex, the objective function F _i is calculated for each of the vectors V _i , the frequency assignments of the population of individuals;

- determine the highest value of the objective function for the generated population and the frequency assignment vector corresponding to it, if the objective function reaches the required value or performs a given number of iterations, they proceed to the stage of forming the vector of frequency usage features;

- arrange individuals according to their target function;

- choose al individuals of the highest ranks;

- from the selected al individuals form a new population of n individuals as follows:

for each i-th individual calculate the value

X _i =n*(F _i -min(F _i ))/((sum of all F _i from the set of αn individuals of higher ranks)-n*min(F _i )),

the vector V _i is copied X _i times into the new population,

n/2 pairs of randomly selected individuals from the new population are crossed,

for each individual of the new population invert randomly selected with the probability of mutation P _m binary digits of the vector V _i ;

- go to the stage of calculating the objective function;

- form a vector of signs of the use of frequencies for each radio-electronic means, using a specific frequency assignment vector;

- turn on active frequencies and block passive frequencies on the radio-electronic means of the complex in accordance with the generated vector of signs of frequency use.

The claimed technical result is also achieved by using a system for optimizing the distribution of frequencies in a complex of compactly located heterogeneous radio-electronic means, made with the possibility of implementing the method for optimizing the distribution of frequencies according to the claimed invention and containing a radio air scanning unit, a frequency measurement unit, a frequency assignment vector generation unit, an objective function calculation unit , a block for generating a vector of signs of using frequencies, a control block for turning on / off frequencies, a block for radio electronic means, moreover, the radio air scanning block is connected to the frequency measurement block and the objective function calculation block,

the frequency measurement unit is connected to the frequency assignment vector generation unit, the frequency assignment vector formation unit is connected to the objective function calculation unit,

the objective function calculation unit is connected to the frequency usage feature vector generation unit and the frequency assignment vector generation unit,

the frequency use feature vector generation unit is connected to the frequency on/off control unit, and the frequency on/off control unit is connected to the radio electronic equipment unit and controls the on/off frequencies of the radio electronic equipment of the complex.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig 1. Flowchart of a genetic algorithm with rank-proportional selection for frequency set assignment

Fig. 2. Dependence of the quality indicator on the coefficient of rank selection

Fig. 3. The process of approximation of genetic algorithms to a quasi-optimal solution

Fig. 4 Block diagram of the system for optimizing the distribution of frequencies in a complex of compactly located heterogeneous radio-electronic facilities

IMPLEMENTATION OF THE INVENTION

The solution to the problem of eliminating the distribution of frequencies and reducing mutual interference in a complex of compactly located heterogeneous radio electronic means (RES) has its own specifics, which is due to the fact that in the case when radio electronic means are located compactly and are of different types, there is no way to reduce mutual interference by removing radio electronic means from each other. from friend.

In addition, in practice, the radio electronic means used, for example, shipborne electronic equipment, are of various types, which does not allow the use of existing standard resource separation methods, such as code separation. It follows that in order to eliminate mutual interference, it is necessary to optimize the distribution of available frequency resources between means.

In the claimed invention, the technical result is achieved through the use of a new genetic algorithm that uses the efficient distribution of a plurality of frequencies between compactly located radio electronic means in an iterative process at each stage of which an assessment of the levels of mutual interference of the complex is carried out, on the basis of which the objective function is calculated.

As an objective function that ensures the efficient assignment of frequencies to compactly located radio electronic means, you can use the amount of information contained in the set of signals received by the radio electronic means of the complex. It is known that the amount of information contained in the set of signals processed by the radio electronic means of the complex depends on the overall level of interference affecting the operation of the complex in such a way that the lower the overall level of interference, the higher the amount of information. The optimal distribution of frequencies between the means of the complex leads to the maximization of the objective function and minimizes the level of interference.

Various options for assessing the level of mutual interference and implementing the objective function are considered, for example, in the article “Information criterion for the effectiveness of systems for estimating signal parameters”, Electrosvyaz, 2019 No. 10.

However, in the claimed invention, the essential feature is not the specific form and method of calculating the objective function, but the use of this function to minimize the noise level.

The term “population” used in this document is a set of options for distributing frequencies between the radio electronic means of the complex, an individual is a set of radio electronic means of the complex for which a vector of frequency assignments is established, crossing is the formation of an individual that has new signs of frequency assignments based on the signs of frequency assignments of others individuals, selection - selection by some criterion of individuals to be crossed.

All definitions used herein take precedence over definitions in dictionaries, definitions in documents incorporated by reference, and/or the conventional meanings of certain terms.

There are a number of strategies for selecting (selection) individuals for crossing - proportional selection, the roulette method, the tournament method, as well as the method disclosed in D2. However, the known strategies do not allow to achieve the claimed technical result.

The proposed genetic algorithm uses a new selection operator that provides more efficient operation of the genetic algorithm when assigning frequencies, and also uses objective function estimation to ensure frequency distribution optimization.

Let us represent the vector of frequency assignments of an individual (a set of RES frequencies) as a binary vector V, in which zero means that the frequency is not assigned (prohibited), and one means that the frequency is assigned (allowed):

where ƒ _i,j is a binary indication of assigning the i-th frequency of the j-th RES. The initial population will be formed from random binary vectors whose elements are equal to 1 with probability P ₁ .

As a crossing operator, we will choose the exchange of vectors of individuals V ₁ and V ₂ values in randomly selected positions, as a result of crossing, vectors V ₃ and V ₄ are obtained, for example:

Crossing in the example is carried out in positions (1,2), (1,3), (2,1), (2,3) and (N,2). The vector V ₃ is the vector V ₁ , which has the elements of the vector V ₂ in the given positions. The vector V ₄ is the vector V ₂ , which has the elements of the vector V ₁ in the given positions. Mutation is carried out by inversion of bits in the binary vector of the individual with a probability R _m .

For each variant of frequency distribution over the radio-electronic means of the complex, it is possible to calculate the value of the objective function that characterizes the efficiency achieved when using each specific variant.

The rank of an individual is its position in the set of individuals ordered in ascending order of the objective function. Let's set the number α (1/n ≤ α ≤ 1) - the coefficient of rank selection.

Let us choose from the population αn the best in rank individuals. From them, using one of the selection methods, we will form a population for subsequent crossing. Thus, in contrast to the roulette, tournament and proportional selection methods and the method disclosed in D2, individuals from among the best in rank necessarily participate in crossing, and the formation of a new population is not carried out randomly (as in D2), but using strategies selection, presented below in the description of step 109 of the claimed method, as well as using the target function, ensuring the achievement of the claimed technical result due to a significant improvement in the quality of eliminating mutual interference in a complex of compactly located radio-electronic means of various types.

The proposed genetic algorithm for assigning sets of frequencies is shown in Fig.1.

At step 101, the radio is scanned and the power of mutual interference occurring between radio electronic means is measured and the size of the population n is determined;

At step 102, the frequencies used by the electronic means of the specified complex are measured;

At step 103, the rank selection coefficient α is set;

At step 104, for each individual, a binary random vector V _i , i=1...n with a binomial distribution Bi(1, P ₁ ), where P ₁ is the frequency resolution probability in the initial population, is set;

At step 105, using the measured mutual interference powers of the complex, an objective function F _i is calculated for each of the vectors V _i of frequency assignments of the population of individuals;

At step 106, the largest value of the objective function for the generated population and the corresponding frequency assignment vector are determined,

At step 106, if the target function reaches the required value or performs a predetermined number of iterations, go to step 111 of the formation of the frequency usage feature vector;

At step 107, individuals are ordered according to their objective function;

At step 108 select an individuals of the highest ranks;

At step 109, a new population of n individuals is formed from the selected an individuals as follows:

109-1: for each i-th individual calculate the value

X _i =n*(F _i -min(F _i ))/((sum of all F _i from the set an of individuals of higher ranks)-n*min(F _i )),

109-2: the vector V _i is copied X _i times into the new population,

109-3: n/2 pairs of randomly selected individuals from the new population are crossed,

109-4: for each individual of the new population, bits of the vector V _i randomly selected with a probability of mutation P _m are inverted;

At step 110 go to step 105 calculation of the objective function; At step 111, a frequency usage feature vector is generated for each radio electronic means using a specific frequency assignment vector;

At step 112, active frequencies are turned on and passive frequencies are blocked on the radio-electronic means of the complex in accordance with the generated frequency usage feature vector.

Figure 2 shows the dependence of the quality index (objective function) on the coefficient of rank selection. The proposed new principle of rank selection was used to improve the quality

proportional, tournament and roulette selection principles. It can be seen from the graph that in the absence of rank selection (in the case of α=1), as well as in comparison with the use of random search (in the case of α=1/n), the efficiency of the new principle of rank selection is much higher compared to traditional methods.

The best results are achieved at 0.1 ≤ α ≤ 0.75. For example, with α=0.3, the ranking method is 20% more efficient than using a random search, as well as proportional, tournament and roulette selection principles.

Figure 3 is a graph illustrating the process of approaching genetic algorithms to a quasi-optimal solution. It follows from the graph that when using the proposed genetic algorithm with preliminary rank selection, the objective function has a higher value and in the process of evolution grows much faster compared to the known algorithms.

Thus, the proposed genetic algorithm with rank-proportional selection shows the best results in terms of quality of work compared to genetic algorithms that do not use rank selection and ensures the distribution of frequency resources between the radio-electronic means of the complex, minimizing mutual interference in the complex.

Thus, the set of essential features of the claimed invention ensures the achievement of the claimed technical result, and the specified technical result is ensured through technical operations with material means, including the measurement of physical parameters and the installation of technical modes of operation of radio electronic means.

The new genetic algorithm disclosed in the claimed invention is an integral part of this method and is intended to generate material control signals that activate or block the frequencies of the radio electronic means complex and provide a result (frequency distribution optimization), which is technical.

The claimed method is implemented using the system, the block diagram of which is shown in Fig.4.

The claimed system contains a radio scanning unit 400, a frequency measurement unit 401, a frequency assignment vector generation unit 402, an objective function calculation unit 403, a frequency usage feature vector generation unit 404, a frequency on/off control unit 405, and a radio electronics unit 406.

A system for eliminating mutual interference in the distribution of frequencies in a complex of compactly located heterogeneous radio electronic means implements the main functions according to the method disclosed in paragraph 1 of the claims of the claimed invention.

In particular, the radio scanning unit 400 scans the radio air and measures the power of mutual interference occurring between electronic means, information about the frequencies used is sent to the frequency measurement unit 401, which performs the frequency measurement, the measurement results are sent to the unit 402 and used to generate a frequency assignment vector , which includes performing steps 103, 104, 107-109 of the claimed method, block 403 performs steps 105, 106 and controls block 402 to generate a frequency assignment vector value that satisfies the requirements of the objective function, which optimizes the frequency allocation.

When the required values of the objective function are reached, block 403 issues a control signal to block 404 to start generating the frequency usage feature vector in accordance with step 111. The generated frequency usage feature vector is fed to the frequency on/off control unit 405, which controls the electronic means 400 to implement the step 112, which provides an optimized distribution of frequency resources between the radio-electronic means of the complex.

Claims (13)

A system for optimizing the distribution of frequencies in a complex of compactly located heterogeneous radio electronic facilities, containing a radio air scanning unit, a frequency measurement unit, a frequency assignment vector generation unit, an objective function calculation unit, a frequency usage feature vector formation unit, a frequency on/off control unit, a radio electronic equipment unit , wherein the radio scanning unit, which scans the radio air and measures the power of mutual interference occurring between electronic means, is connected to the frequency measurement unit and the objective function calculation unit, the objective function calculation unit is connected to the frequency assignment vector generation unit; the frequency measurement unit measures the frequencies used by the radio electronic means of the complex, and is connected to the frequency assignment vector formation unit; the frequency assignment vector generation unit is configured to implement the following functions: - setting the coefficient of rank selection α, - settings for each radio-electronic facility of a binary random vector of frequency assignment Vi, i=l...n with binomial distribution Bi(l, P1), where P1 is the frequency resolution probability in the initial population, - ordering the radio-electronic means of the complex, for which the vector of frequency assignments is established, according to their target function, - selection of αn radio-electronic means of the complex, for which the vector of frequency assignments of the highest ranks is established, - formation from the selected αn radio electronic facilities of the complex, for which the vector of frequency assignments is set, a new population of n radio electronic facilities of the complex, for which the vector of frequency assignments is set with the ability to perform operations, including: calculation of the value Xi=n*(Fi-min(Fi) )/((the sum of all Fi from the set αn of radio electronic facilities of the complex, for which the vector of frequency assignments of the highest ranks is set)-n*min(Fi)) for each i-x radio electronic facilities of the complex, for which the vector of frequency assignments is set, copying the vector Vi into a population Xi times, the formation of radio electronic facilities of the complex with new signs of frequency assignments, based on the signs of frequency assignments of other radio electronic facilities of the complex for n/2 pairs of randomly selected radio electronic facilities of the complex, for which a vector of frequency assignments from the new population is established, inversion of randomly selected probability of mutation Pm binary digits of the vector Vi for each radio-electronic means of the complex, for which the vector of frequency assignments of the new population is established; the frequency assignment vector generation unit is connected to the objective function calculation unit; the objective function calculation unit is configured to implement operations: calculating the objective function Fi using the measured mutual interference powers of the complex for each of the vectors Vi of frequency assignments of the population of radio electronic facilities of the complex, for which the vector of frequency assignments is set; determining the highest value of the objective function for the generated population and the corresponding frequency assignment vector; the target function calculation unit is connected to the frequency usage feature vector generating unit and is configured to, if the target function reaches the required value or performs a predetermined number of iterations, outputs a control signal to the frequency usage feature vector generation unit to generate the frequency usage feature vector; the frequency use feature vector generation unit is connected to the frequency on/off control unit, and the frequency on/off control unit is connected to the electronic equipment unit and is configured to control the on/off frequencies of the radio electronic equipment of the complex.

Images