RU2257611C2 - Method for automatic structuring of computer codes adequate for processed information - Google Patents

Abstract

FIELD: computer science.

SUBSTANCE: method includes performing a block of operations along N₁ channels, where N₁ is selected from 1 to 2²⁵⁶, wherein received information is separated on logically finished fragments, encoded on basis of preset algorithm, to produce a block of N-dimensional sets adequate for converted source information A_j with elements like {B_m, X₁, X₂,...,X_n}, where j - order number of set in range from 1 to 2²⁵⁶, B_m - identifier, X₁-X_n - coordinate of element from its coordinates center, m and n are selected from 1 to 2²⁵⁶; received block of sets is compared to already accumulated and/or newly produced sets from multiple channels, intersecting portions of sets are found and cut out; after that cut intersections and sets remaining after cutting are distributed among databases, placing each same set into database appropriate for it and each of sets different with some parameter to databases appropriate for them and identifiers of databases storing these sets are substituted in place of cut sets.

EFFECT: higher speed of operation, higher precision, lower costs, broader functional capabilities, higher efficiency.

9 dwg

Description

Technical field

The invention relates to the field of computer technology and can be used for automatic structuring of computer codes adequate to the processed information.

State of the art

A known method for the automatic structuring of computer codes adequate to the information being processed, including the allocation of part of the information received per unit time, which is selected depending on the type of information, for example, digital or textual, breaks the received information into logically complete fragments, encodes the resulting pieces of information and distributes them on the databases [Ailamazyan AK, Stas E.V. "Computer science and development theory." M., Science, 1989].

There is also known a method of automatically structuring computer codes that are adequate to the information being processed, including extracting from the volume of information received per unit time its part, which is selected depending on the type of information, for example, digital and / or textual and / or graphic, breaking the received information into logically completed fragments, encode the obtained fragments of information, converting them according to a given algorithm, the resulting set of transformed fragments is distributed across databases, placing azhdy of the fragments in the corresponding base [D.E.Bestens, V.-M.van den Berg, D.Vud. Neural networks and financial markets, M :, Publisher TVP, 1977].

The disadvantages of the known methods are the imperfection of processing computer codes that are adequate to the source information, and the relatively low rate of structuring of computer codes.

The problem to be solved and the technical result achieved

The objective of the invention is to improve methods for the automatic structuring of computer codes that are adequate to the information being processed, while achieving a technical result in relation, in particular, to increase the speed of structuring computer codes. Using the whole method of automatic structuring of computer codes adequate to the information being processed can significantly save the consumption of computing resources (memory, the number of processor operations, etc.), and increase the speed and accuracy of structuring computer codes.

Summary of the invention

As brief information revealing the essence of the invention, it should be noted that the achieved technical result is provided by the proposed method for automatically structuring computer codes adequate to the processed information, which includes a set of operations carried out on N ₁ channels, where N _{1 is} selected from 1 to 2 ²⁵⁶ . Moreover, in each of the N ₂ channels, selected in an amount satisfying the inequality 1 + 1 / N ₁ ≤ (N ₂ + α ₁ N ₁ ) / N ₁ ≤ 2.3, where α ₁ is the experimental coefficient, chosen depending on type of information and from the number of sources of its receipt in the range from 0.61≤ α ₁ ≤ 1.3, it is isolated from the volume V _{1 of information} received per unit time its part of the volume V ₂ , which is selected from the ratio within 0.56≤ (V ₂ + α ₂ V ₁ ) / V ₁ ≤ 2.8, where α ₂ is the experimental coefficient chosen depending on the type of information in the range from 0.56 ≤ α ₂ ≤ 1.8. The received information is divided into logically completed fragments, the volumes of V _{3i of} which are selected from the ratio within the range of 0.45≤ (V _3i + α ₃ V ₂ ) / V ₂ ≤ 2.7, where α ₃ is the experimental coefficient chosen depending on the particular their partitions in the range from 0.45 ≤ α ₃ ≤ 1.7, and i is selected in the range ₁ ≤ i ≤ 2 ²⁵⁶ .

Then, the obtained pieces of information are encoded, transforming them according to a predetermined algorithm to obtain a set of N-dimensional sets that are adequate to the source information being transformed. And _j elements of the form {B _m , X ₁ , X ₂ , ... X _n }, where j is the serial number of the set ranging from 1 to 2 ²⁵⁶ , B _m is the identifier, X ₁ -X _n is the coordinate of the element relative to the beginning element coordinates, am and n are selected in the range from 1 to 2 ²⁵⁶ . The resulting set of N-dimensional sets is compared with already accumulated sets and / or with newly arriving sets through different channels, intersecting parts of the sets are determined and cut out. After that, the cut-off intersections of the sets and the sets remaining after cutting are distributed across the databases, placing each of the same sets in its corresponding database and each of the sets that differ in some parameter in the corresponding sets of the database and put the identifiers storing in place of the cut-out sets these sets of databases.

List of drawings

Figure 1. The layout of computer codes adequate to the processed information. Figure 2. A simplified scheme of the method for one channel. Figure 3. A simplified diagram of the implementation of the method for two channels. Figure 4. A simplified diagram of the implementation of the method for ten channels. Figure 5. A detailed illustration of the claimed method. 6. A fragment of a detailed illustration of the claimed method.

Detailed Description and Embodiments

When presenting information confirming the possibility of carrying out the invention, it is advisable to describe in more detail the proposed method for automatically structuring computer codes that are adequate to the information being processed, when describing which it is inappropriate to dwell on the features of its operations known from published data. In this case, it is advisable to bring definitions of the concepts used for the convenience of perception of information.

Computer code is a machine-convertible representation of a piece of information (the original), in particular, in the form of electromagnetic signals.

A criterion is a factor by which recognition is performed, for example, a coordinate parameter, a logical condition, etc.

A system of criteria is a set of criteria interconnected by their target orientation, in particular, their suitability for use in the recognition process.

Memory is a functional unit of the means of implementing the method, fixing processed information for a certain period of time, with the possibility of its further reproduction, reading, and other uses.

A given algorithm is a set of rules by which the source information arriving on each channel is transformed.

A set is a combination of certain completely distinguishable objects into a single whole, which are called elements of the set formed by them [″ Handbook of Mathematics ″, I.N. Bronstein, K.A.Semendyaev, ed. ″ Science ″, Moscow, 1986. p. 380. p. 4.1.2.1].

The distinguished object is computer codes that are adequate to the corresponding pieces of information.

The intersection of two sets A _j and A _{j + k} is a set that consists of elements belonging to each of the sets A _j and A _{j + k} [″ Mathematics Reference ’, I.N. Bronstein, K.A.Semendyaev, ed. . ″ Science ”, Moscow, 1986, p. 381, clause 4.1.3.1], where j and j + k are the numbers of sets taken from the range from 1 to 2 ²⁵⁶ , or, by analogy, the intersection of any number structured according to the claimed method sets.

Further, in detail, it is advisable to dwell only on the distinguishing essential features of the operations of the proposed method, namely. that the method includes a set of operations carried out on N ₁ channels, where N ₁ choose from 1 to 2 ²⁵⁶ . The choice of the specified number of channels within the stated limits provides almost all currently existing needs for the implementation of the method. Moreover, in each of the N ₂ channels, selected in an amount satisfying the inequality 1 + 1 / N ₁ ≤ (N ₂ + α ₁ N ₁ ) / N ₁ ≤ 2.3, where α ₁ is the experimental coefficient, chosen depending on type of information and from the number of sources of its receipt in the range from 0.61≤ α ₁ ≤ 1.3, it is isolated from the volume V _{1 of information} received per unit time its part of the volume V ₂ , which is selected from the ratio within 0.56≤ (V ₂ + α ₂ V ₁ ) / V ₁ ≤ 2.8, where α ₂ is the experimental coefficient chosen depending on the type of information in the range from 0.56 ≤ α ₂ ≤ 1.8. In practice, these can be varieties of digital, textual, symbolic, graphic and mixed information. The received information is divided into logically completed fragments, the volumes of V _{3i of} which are selected from the ratio within the range of 0.45≤ (V _3i + α ₃ V ₂ ) / V ₂ ≤ 2.7, where α ₃ is the experimental coefficient selected depending on the particular their partitions in the range from 0.45 ≤ α ₃ ≤ 1.7, and i is selected in the range ₁ ≤ i ≤ 2 ²⁵⁶ . Among the simplest features of dividing the received information into logically complete fragments, it is possible to indicate, for example, equal volumes of fragments or not equal fractions.

Then, the obtained pieces of information are encoded, transforming them according to a predetermined algorithm to obtain a set of N-dimensional sets adequate to the transformed initial information, A _j elements (and / or frames) of the form {B _m , X ₁ , X ₂ , ... X _n ) where j is the sequence number of the set in the range from 1 to 2 ²⁵⁶ , In _m is the identifier, X ₁ -X _n is the coordinate of the element relative to the origin of the element, am and n are selected in the range from 1 to 2 ²⁵⁶ . It is worth noting here that these elements are often called frames and each element has its own coordinate system. The resulting set of N-dimensional sets is compared with already accumulated sets and / or with newly arriving sets through different channels, intersecting parts of the sets are determined and cut out. Comparison with already accumulated sets and / or with new entrants is carried out simultaneously or sequentially in time. After that, the cut intersections of the sets and the sets remaining after cutting are distributed over the databases, placing each of the same sets in its corresponding database and each of them differing in some parameter of the set in the corresponding sets of the database sets and put the identifiers storing in place of the cut sets these sets of databases.

Given the use of a number of analytical relationships, it should be noted that for practical implementation, from sets that satisfy the analytical relationships of the parameter values within the stated limits, they are chosen for the parameters, for example, N, N ₁ or N ₂ are positive integer positive integers, and for other parameters - these are real numbers, excluding irrational, transcendental, complex, negative and other technically incorrect or practically not applicable, or not reproducible values of interrelated parameters moat.

Industrial applicability and technical result achievement

Compliance with the criterion of industrial applicability of the claimed object is proved both by the wide receipt and use of the described techniques of computer technology on a massive scale, and the absence of any practically difficult features in the claimed claims. Achievable technical result, as shown by experimental data, can be realized only by an interconnected set of all the essential features of the claimed method, reflected in the claims. For example, when searching for the word ″ Example ″ in an unsorted, non-indexed database of 100 million records or 100 billion records, the number of operations will be approximately the same, i.e. search speed will be the highest compared to known methods under adequate conditions for their use, it depends only on the length of the request and very weakly depends on the size of the database.

The lower and upper values of the declared limits were obtained on the basis of statistical processing of the results of experimental studies, analysis and generalization of them and known from published data sources, as well as inventive intuition, based on the conditions for achieving the specified technical result. The differences indicated in the claims make it possible to conclude that this technical solution is new, and the set of claimed claims is about its inventive step, which is proved both by the above detailed description of the claimed objects and the following description of examples of practical implementation of the claimed method. To this end, it is advisable to give the main functional diagrams of the method for automatically structuring computer codes adequate to the processed information when creating databases.

The conversion of computer codes adequate to the information being processed by the proposed method can partially be explained using Fig. 1A), which reflects the case when the first set arrives, schematically displayed as oval 1. Since the general database is still empty, the whole set is placed in new database database number 1 data. Fig.1B) reflects the case when the next two sets 2 and 4, containing matching elements. The intersection of sets 3 is cut out and placed in the corresponding database No. 3, in its place in sets 2 and 4 the identifier of database No. 3 is inserted, the modified sets 2 and 4 are placed in the corresponding database of database No. 3 and database No. 4.

Figv) reflects the case when the next set N-1 arrives that does not contain any matching element with the previous sets, therefore, the set N-1 is placed in a new database DB-NN-1. Fig.1G) reflects the case when the next set N arrives containing completely matching elements with the set N-1 stored in the database database No. N-1. The intersection of the sets N is cut out and the corresponding database of the database No. N is inserted, in its place in the set N-1 is inserted the identifier of the database of the database No. N, the database of the database No. N-1 is accordingly changed. Thus, in each separate database of the database, sets are stored that do not have matching elements, i.e. none of the sets intersect.

Simplified diagrams of the method in series for one, two and 10 channels are shown in FIG. 2, FIG. 3 and FIG. 4, respectively. Figure 2 schematically shows a structured database of text information designed to store large arrays of text information with the selection of one channel of text information.

Figure 3 schematically shows a structured database of text and graphic information, designed to store large arrays of text and graphic information with a choice of two channels - one for text information, the second for graphic information.

It is advisable to characterize in detail a structured database of text, graphic and audio information, intended for storing large arrays of text, graphic and audio information arriving simultaneously through 10 channels (Figure 4), for example 6 text, 2 graphic and 2 audio. The set of operations of the method implemented in the device of Figure 4 can be characterized as follows: N _{1 is} chosen equal to 10, N ₂ from 2 to 6 with α ₁ = 1 for text information, α ₁ = 0.9 for graphic information, α ₁ = 1.1 for audio information. Allocated from the volume V ₁ = 10 000 Kb / s of information received per unit time, its part of the volume V ₂ = 1000 Kb / s with α ₂ = 1 for textual information, α ₂ = 0.8 for graphic information, α ₂ = 1, 2 for audio information. The received information is divided into logically completed fragments, the volumes of which are selected from the ratio 1.4 ≤ (V _3i + α ₃ V ₂ ) / V ₂ ≤ 1.6, where α ₃ = 1 for textual information, α ₃ = 0.5 for graphic information, α ₃ = 1.7 for audio information, and i is selected within 1≤ i≤ 2 ²⁰ .

Encode the obtained pieces of information, transforming them according to a given algorithm to obtain a set of N-dimensional sets A _{j of} elements of the form {Bm, X ₁ , X ₂ , ... Xn}, where B _m is the identifier, X ₁ -X _n is the coordinate of the element , a _j is the serial number of the set in the range from 1 to 2 ⁶³ , relative to the origin of the element (each element has its own coordinate system), adequate to the transformed initial information, where m and n are selected in the range from 1 to 2 ⁶³ . The resulting set of N-dimensional sets is compared with already accumulated sets and / or with newly arriving (simultaneously or sequentially in time) sets in different channels, the intersecting parts of the sets are extracted and cut out, after which the cut out intersections of the sets and the sets remaining after cutting are distributed over the databases , placing each of the same sets in its corresponding database and each of the sets differing in some parameter in the sets of the database corresponding to these types and in eschayut the place of cut sets identifiers storing these multiple databases.

In this case, an experimentally obtained rate of automatic structuring of computer codes adequate to the processed information, equal to 930.55 Kb / s on a computing device with four Pentium III 800MHz processors, and 1024 MB of RAM.

To illustrate the claimed method in the example of Figure 5, a structured database of textual information was used. This database is designed to store large amounts of text information. In the process of implementing the inventive method, choose (Figure 5, the operation is selected, shown by arrow 1) the number of channels (item 2) N ₁ equal to 1, N ₂ = 1 with α ₁ = 1, which transmit the volume V ₁ (Figure 5 , position 3) information. Isolate (Figure 5, the operation is highlighted, indicated by an arrow, is indicated by 4) from the volume V ₁ = 100 Kb / s of the information received per unit of time, its part (Figure 5, position 5) of volume V ₁ = 8 b / s, t .e. in this case, V ₂ is a piece of textual information, for example, the word ″ argument ″, with α ₂ = 0.9.

Then, the received information is divided (FIG. 5, POS. 6) into logically completed fragments (FIG. 5, POS. 7), the volumes of which are selected within 1.3 ≤ (V _{3 (i)} + α ₃ V ₂ ) / V ₂ ≤ 1.5, where α ₃ = 1, ai is chosen within the range of 1≤ i≤ 2 ³ . As a result, the following specific values determined for the volumes of fragments 7 determined using the above analytical relation are obtained: V _{3 (1)} = 1 b / s, V _{3 (2)} = 1 b / s, V _{3 (3)} = 1 b / s , V _{3 (4)} = 1 b / s, V _{3 (5)} = 1 b / s, V _{3 (6)} = 1 b / s, V _{3 (7)} = 1 b / s, V _{3 (8)} = 1 b / s. The selected specific values of the logically complete fragments 7 are encoded (Figure 5, item 8), transforming them according to a given algorithm, in particular for this example, y = f (x), where f (x) = V _{3 (i)} -191, to get the set of N-dimensional sets A _j (pos. 9) of elements a _k (pos. 10) of the form {B _m , X ₁ , X ₂ , ... X _n }, where B _m is the identifier, X ₁ -X _n is the coordinate of the element, where j is the sequence number of the set in the range from 1 to 2 ³ , 1≤ k≤ 2 ³ , relative to the origin of the element (each element a _k has its own coordinate system), adequate to the transformed initial information, where m and n choose before Lah from 1 to 2 ³¹ .

As a result, transformed elements are obtained in this particular case of the form a ₁ = {1, 1}, and ₂ = {2, 2}, and ₃ = {3, 3}, a ₄ = {4, 4}, and ₅ = { 5, 5}, and ₆ = {2, 6}, and ₇ = {6, 7}, and ₈ = {7, 8}. In this case, individual identifiers (first digits in curly brackets) of the elements of the sets can coincide if the results of converting pieces of information according to a given algorithm coincide. Further, the set of operations is performed as a single unit (Figure 5, item 11). The resulting set of N-dimensional sets 9 are compared (Fig. 5, position 12) with the already accumulated sets (for example, database 13 (DB No. 1)) and / or with the new sets (simultaneously or sequentially in time) of sets 14 and 15 each different channels. The intersecting parts of the sets 17 and 18 are determined and cut out (Fig. 6, position 16). After that, the intersected sets of the sets (Fig. 6, position 19) and the sets remaining after cutting out with the replaced parts are distributed (Fig. 5, position 19) into the bases data, placing each of the same sets in its corresponding database and each of the sets differing in some parameter in the sets corresponding to these types of database sets and put (Figure 5, position 20 or Figure 6, position 20) in place of the cut-out sets identifiers storing these sets ba data.

The result is (Fig. 5, position 21) a common database consisting of hierarchically related databases - each child database is represented by its identifier in the respective parent databases.

In this case, an experimentally obtained rate of automatic structuring of computer codes adequate to the processed information, equal to 96 Kb / s on a computing device with one Pentium III 800 MHz processor, 256 MB of RAM.

In addition to the above technical result, the practical implementation of the claimed object allows you to significantly expand the possibilities of its use as applied, for example, to various types and volumes of incoming information, provides an additional opportunity for highly efficient hardware implementation (price / quality / speed) using a wide variety of element base types, as well as automatic adaptation to solving problems of various types, unlike neural networks that require a predetermined topology for each specific the exact task.

Claims (1)

A method for automatically structuring computer codes in the form of electromagnetic signals adequate to the processed information, including a set of operations automatically performed on N ₁ channels, where N _{1 is} automatically selected from 1 to 2 ²⁵⁶ , so that in each of N ₂ channels are automatically selected in an amount satisfying the inequality 1 + 1 / N ₁ ≤ (N ₂ + α ₁ N ₁ ) / N ₁ ≤ 2,3, where α ₁ is the experimental coefficient, automatically selected depending on the type of information presented in the form of computer codes, adequate processing Pipeline information on the number of sources and its receipt within 0,61≤ α ₁ ≤ 1,3, automatically isolated from volume V ₁ supplied per unit of time information represented in the form of computer code, adequate information processed, its part of the volume V _2, which is automatically selected from a ratio in the range of 0.56≤ (V ₂ + α ₂ V ₁ ) / V ₁ ≤ 2.8, where α ₂ is the experimental coefficient, automatically selected depending on the type of information presented in the form of computer codes adequate processed information, within 0.56≤ α ₂ ≤ 1.8, the received information, presented in the form of computer codes adequate to the processed information, is automatically divided into logically completed fragments, the volumes of V _{3i of} which are automatically selected from the ratio within 0.45≤ (V _3i + α ₃ V ₂ ) / V ₂ ≤ 2.7, where α ₃ is the experimental coefficient, automatically selected depending on the features of their partition within 0.45 ≤ α ₃ ≤ 1.7, and i are automatically selected within 1 ≤ i ≤ 2 ²⁵⁶ , the resulting fragments of information presented in the form of computer codes, hell cotton processed information, automatically converting them by a given algorithm until a plurality of N-dimensional sets containing computer code, transformed adequate initial information, A _j elements of the form {B _m, X _1, X _2, ... X _n}, where j - the serial number of the set in the range from 1 to 2 ²⁵⁶ , B _m is the identifier, X ₁ -X _n is the coordinate of the element relative to the origin of the element, and m and n are automatically selected in the range from 1 to 2 ²⁵⁶ , the resulting set of N-dimensional sets automatically correlate with already accumulated sets and / or with newly automatically arriving sets through different channels, the intersecting parts of the sets are automatically detected and automatically cut, after which the cut-off intersections of the sets and the remaining sets after cutting are automatically distributed to the databases, automatically placing each of the same sets into its corresponding database and each of them differing by some parameter sets in the corresponding to these types of sets of databases, and automatically put in place of the cut sets identifier ry storing these sets of databases containing computer codes adequate to the processed information.

Images