A method for intelligent settlement of queue processes based on interval one-dimensional extreme value search analysis

Through the golden-divided interval one-dimensional extreme value search and analysis method, the procurement contract settlement process of the power company is automatically handled, solving the problem of extension of the settlement cycle caused by manual verification and review, and achieving efficient, accurate and automated flow of the settlement process.

CN114493813BActive Publication Date: 2024-12-06STATE GRID JIANGSU ELECTRIC POWER CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111541810.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-16
Publication Date
2024-12-06
Estimated Expiration
2041-12-16

AI Technical Summary

Technical Problem

During the settlement process of the power company's procurement contract, there is a lot of manual verification and review work, resulting in a long settlement cycle, the settlement conditions and materials at each step of the circulation process need to be reviewed multiple times, and there are human factors in the approval process that lead to delays.

Method used

Using the interval one-dimensional extreme value search analysis method based on golden segmentation, we can form a Fibonacci sequence by automatically collecting data, constructing interval monomer functions, filtering settlement applications that meet the conditions, realizing automatic circulation and online signature signatures and seals, and improving settlement efficiency .

Benefits of technology

The settlement process is automated, the manual workload is reduced, settlement efficiency and accuracy is improved, the impact of human factors on the circulation process is reduced, and the settlement process is efficient, non-blocking and high availability is ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114493813B_ABST
    Figure CN114493813B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for intelligent settlement of queue processes based on interval one-dimensional extreme value search and analysis of golden section, wherein the settlement documents are electronically entered into the system process, and the payment conditions of each section are automatically collected to form a Fibonacci sequence with the Hash value of the payment condition as an element; the obtained conditional sequence is subjected to interval one-dimensional extreme value search and analysis based on golden section to determine whether it fully meets the conditions for generating a settlement application form; the processes with conditions of each settlement node meeting the requirements are found to enter the settlement application queue of the corresponding type, and a settlement application form is generated; after the settlement application form is generated, online approval is initiated, and the settlement documents enter the online approval process of electronic signature and seal, and finally the settlement is completed. The present invention improves the fluency, integrity and accuracy of the settlement process nodes, and improves the settlement efficiency of each section.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to a method for queue process settlement based on interval one-dimensional extreme value search and analysis, and specifically to a method for intelligent settlement of queue processes based on interval one-dimensional extreme value search and analysis based on the golden section. Background Art

[0002] As a large enterprise unit, the power company circulates payment application forms and related documents such as purchase contracts in the settlement process of many purchase contracts in the form of documents. In the previous settlement process, a large amount of document material data needs to be manually checked and collected. In the settlement process, the settlement cycle is often long due to the lag in manual inspection and review or the heavy workload; and the settlement conditions and materials of each step in the settlement of each purchase contract in the circulation process need to be manually checked and even reviewed multiple times. At the same time, there are some uncertain human factors in the approval process of the settlement application form, which leads to the fact that the leaders of each step in the approval process of the settlement application form may not be able to approve it in time. With the implementation of electronic office, electronic government affairs, and the popularization of electronic documents; in order to ensure the automation of the settlement application process and the authenticity and reliability of electronic documents; it is necessary to introduce interval one-dimensional search analysis queue process settlement and electronic signature technology to solve it. Summary of the invention

[0003] In view of the problems existing in the existing settlement process, such as heavy audit workload, settlement process flow progress and settlement efficiency, the purpose of the present invention is to provide a queue process settlement method based on interval one-dimensional extreme value search analysis of the golden section, so as to realize node analysis and review of the contract settlement application form process, automatic classification and flow of the process, and online signature and sealing.

[0004] The purpose of the present invention is achieved through the following technical solutions:

[0005] A method for intelligent settlement of queue processes with interval one-dimensional extreme value search and analysis is characterized by: firstly, automatically collecting data, conditions and process nodes in the process of electronicization of documents, performing Hash conversion on the collected data to form a Fibonacci sequence; then constructing an interval univariate function and using the golden section characteristics to search for data conditions, screening out settlement applications that meet the conditions, entering the next settlement process queue to queue up and perform settlement circulation in turn; finally, initiating online approval to sign and seal the settlement, thereby improving the efficiency of settlement and releasing a large amount of settlement workload; comprising the following steps:

[0006] (1) Settlement documents are electronically entered into the system process, and the payment conditions of each section are automatically collected to form a Fibonacci sequence with the payment condition hash value as the element;

[0007] (2) Through the interval one-dimensional extreme value search analysis method, it is intelligently determined whether each settlement node of a settlement application process fully meets the conditions for generating a settlement application form;

[0008] (3) Filter out the processes that fully meet the requirements of each settlement node and enter them into the settlement application message queue, and generate a settlement application form;

[0009] (4) The queue is distributed and processed to initiate online approval. The settlement document enters the electronic signature and seal online approval process and finally completes the settlement.

[0010] Furthermore, in step (1), the offline paper documents of each stage are electronically operated and entered into the system to form informationized electronic documents. The system forms the Hash value Fibonacci sequence of the settlement conditions of each stage based on the input information and related system information, and automatically updates and improves the Hash value Fibonacci sequence of the conditions of each stage on a regular basis.

[0011] Step (2) specifically includes:

[0012] (3-1) In order to obtain the characteristic value of the search condition, before searching, the golden section characteristics are used to take the hash value of each node as the minimum value of the univariate function to form different objective functions for each node condition, and multiple node condition searches are assigned to univariate functions in different sections or intervals, and the interval one-dimensional search process is regarded as the process of solving the univariate function value. The univariate linear function obtains the minimum value through the simplex method; according to the different objective functions and the golden section characteristics, continuous trials are carried out, and at the same time, the minimum value is used as the "bottom", and the bottom is finally locked with a continuously shrinking range to obtain the minimum value; finally, the conditional characteristic value is obtained by screening and searching;

[0013] (3-2) When using interval-based one-dimensional search, we hope that the objective function is a convex function in the domain of definition or a convex function in a certain interval, so that the minimum value is unique; suppose a convex function f(x) on [a1, b1], and suppose the error coefficient e and the shortening ratio coefficient α (0<α<1), and take a symmetrical subinterval [λ1, μ1] in the initial interval [a1, b1]. The symmetry means that the distance from λ1 to the endpoint a1 is equal to the distance from μ1 to the endpoint b1. To satisfy the symmetry, construct λ1 and μ1 as follows: suppose the distance from λ1 to the endpoint b1 is α times the length of the interval [a1, b1], and the distance from μ1 to the endpoint a1 is also α times the length of the interval [a1, b1];

[0014] (3-3) After obtaining the positions of λ1 and μ1 and substituting λ1 and μ1 into the objective function, there are two cases:

[0015] ③f(λ1)<=f(μ1), then set the original endpoint b1 of the interval to μ1, α1 remains unchanged, and the new interval [a2, b2]=[α1, μ1] is obtained. At this time, the length of the new interval is shortened to α of the length of the original interval [a1, b1]; check the length of the new interval μ1-a1, if the length is less than e, exit the calculation, and the minimum point is (μ1+a1) / 2; if it is greater than the error coefficient e, then take a symmetrical subinterval [λ2, μ2] in the new interval [a1, μ1] again, and the rules for selecting λ2 and μ2 are the same as the initial ones, that is, the distance from λ2 to b2 is α times the length of the interval [a2, b2], and the distance from μ2 to the endpoint a2 is α times the length of the interval [a2, b2];

[0016] ④f(λ1)>f(μ1), then set the original endpoint a1 of the interval to λ1, and b1 remains unchanged. Then we get the new interval [a2, b2]=[λ1, b1]. Check the length b-λ1 of the new interval. If the length is less than e, exit the calculation. At this time, there is a minimum value (λ1+b1) / 2. If it is greater than the error coefficient e, take a symmetrical sub-interval [λ2, μ2] in the new interval [λ1, b1]. The interval setting rules of [λ2, μ2] are the same as those in case ①.

[0017] Step (3) After the interval one-dimensional extreme value search analysis, each payment process that meets the settlement application enters the corresponding queue, completes the generation of the payment application form and initiates the online approval process.

[0018] Step (4) The electronic signature system is used to select and process the signature and seal workflow to sign and seal the electronic document, complete the online approval process, and finally complete the settlement.

[0019] The present invention can utilize the characteristics of the Fibonacci sequence's leap-forward progress and perform interval-type one-dimensional extreme value search and analysis based on the golden section, the advantages of more efficient search, the high availability of message queues, the low probability of data loss, the advantages of decoupling, asynchrony, peak shaving, and the convenience, efficiency, and security of electronic signatures and seals, so as to more accurately search, calculate, and verify settlement condition information during the settlement process, and improve the fluency, completeness, and accuracy of settlement process nodes; at the same time, the high concurrency of the settlement process can be effectively solved during the settlement peak period, and the settlement efficiency of each section can be improved. The automatic search and analysis during the document circulation process and the automatic diversion settlement are performed through interval-type one-dimensional extreme value search and analysis to achieve the purpose of intelligent analysis and settlement and reduce the scope of influence of human factors on the settlement process circulation.

[0020] Compared with the existing settlement process, the present invention has the following significant advantages: through the method of the present invention, the efficiency and accuracy of settlement condition review can be greatly improved, and the transition from manual to automatic can be achieved; at the same time, the golden section search and classified message queue processing are used to allow multiple settlement processes to be carried out concurrently, thereby achieving a non-blocking, non-pause assembly line flow of the settlement process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart of the method of the present invention; DETAILED DESCRIPTION

[0022] This embodiment provides a method for intelligent settlement of queue processes based on interval one-dimensional extreme value search analysis of the golden section, such as Figure 1 As shown, including:

[0023] (1) When the settlement document is electronically entered into the system, the document data items and the node information of the current document are collected, and the relevant document data information of the associated system is obtained through the interface; the corresponding code is found according to the data and the hash value of the code is taken and stored in the sequence. During the conversion process, the corresponding Fibonacci sequence is formed for different types of data.

[0024] (1-1) The Fibonacci sequence has a number that, starting from the third term, each term is equal to the sum of the previous two terms; that is: n =a n-1 +a n-2 .

[0025] (1-2) When the number of elements in the ordered list is not a number in the Fibonacci sequence, it is necessary to pad the length of the number of elements in the ordered list to make it a value in the Fibonacci sequence. The general formula of the Fibonacci sequence is: Also known as the "Binet formula", it is an example of using irrational numbers to represent rational numbers. Note: In this case, a1 = 1, a2 = 1, a n =a n-1 +a n-2 ,

[0026] (2) Select an interval from the obtained sequence to construct the objective function, and perform a one-dimensional search analysis within the interval based on the golden section to determine whether the conditions for generating a settlement application form are fully met;

[0027] (2-1) We know that 0.618 is a golden ratio. When the shortening ratio α is 0.618, it is called the golden section. Suppose that the new interval [a k , b k ], and then generate an inner subinterval [λ k , μ k], according to the symmetry of the subinterval and the shortening ratio α, we can get the following two equations:

[0028] b k -λ k =μ k -a k

[0029] b k +1-a k +1=α(b k -a k ) (1)

[0030] This gives λ k , μ k The expression is:

[0031] λ k =α k +(1-α)(b k -a k )

[0032] μ k =a k +α(b k -a k ) (2)

[0033] ①At this time, if there is f(λ k )<=f(μ k ), according to the algorithm characteristics, we can know that the next new area construction time is:

[0034] [a k +1, b k +1]=[a k , μ k ]

[0035] At the same time, it is necessary to k +1, b k +1] to get the checkpoint [λ k +1, μ k +1], from equation group (2) we know:

[0036] μ k +1 = a k +1+α(b k +1-a k +1) = a k +α(μ k -a k )=a k +α 2 (b k -a k )

[0037] If there happens to be:

[0038] μ k +1 = a k +α 2 (b k -a k )=a k +(1-α)(b k -a k )=λ k (3)

[0039] That is μ k +1 = λ k When k +1 does not need to be calculated, just take the λ of the previous round k To satisfy equation (3), we need to satisfy the quadratic equation:

[0040] α 2 =1-α

[0041] Since α>0 is required, solving the above equation yields α=0.618

[0042] ②When f(λ k )>f(μ k ), we can also draw the following conclusion: when α=0.618, λ k +1 does not need to be calculated, just take the μ of the previous round k That's it.

[0043] It can be seen from this that if α=0.618 is selected, when obtaining the checkpoints λ and μ in each round of iteration, one of the parameters can be taken from the parameters of the previous round, which can save a lot of calculations.

[0044] (2-2) When the new subinterval [a k , b k ] When the distance is greater than the error e, it means that the range has not been narrowed down to a reasonable interval. At this time, the newly generated sub-interval needs to be further divided to obtain the interval [λ k , μ k ], and λ k , μ k Substitute the objective function into the calculation again, and then judge according to the above two situations until the distance is less than the error e and exit the calculation. k , b k] is constantly shrinking (0<α<1), and it is guaranteed to converge to the minimum point of the function in the end. On the other hand, in the algorithm implementation, the larger function value is selected to shorten the distance each time, just like the center of gravity of an object is biased towards the larger function value. Choosing the larger function value to shorten can ensure that the center of gravity gradually approaches the minimum point. The new interval range can always cover the bottom of the function. Each generated interval is a sub-nested interval. According to the closed interval nesting theorem, the limit of the continuously shrinking nested interval is a point, that is, the minimum point.

[0045] (3) All processes that fully meet the requirements enter the settlement application message queue and generate a settlement application form;

[0046] (3-1) Create a settlement process type queue, and search for settlement conditions for different objective functions in (2); if no settlement conditions are found, the current application is considered to be unqualified for entering the next settlement node and continues to enter the next round of condition screening. If the materials and settlement conditions required for the current settlement can be obtained through the search in (2) and their minimum value position in the objective function, they are thrown into the queue of the next settlement node and wait for automatic processing.

[0047] (4) Initiate online approval, enter the online approval process of electronic signature and seal, and finally complete the settlement.

[0048] (4-1) The settlement applications that enter the queue in (3) will be processed in sequence. For settlement applications with multiple steps and conditions, (2) and (3) will be analyzed, searched and screened in a loop, and finally a settlement application form will be generated. The online approval process will be initiated to sign and seal the settlement application online.

[0049] The above description is only a preferred embodiment of the present invention and cannot be used to limit the scope of rights of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for intelligent settlement of queue processes based on interval one-dimensional extreme value search and analysis, characterized in that: First, the data, conditions and process nodes in the electronic document process are automatically collected, and the collected data is hashed to form the Fibonacci sequence; then, an interval univariate function is constructed and the golden section characteristics are used to search for data conditions, and the qualified settlement applications are screened out, and the applications enter the next settlement process queue to queue up and proceed with the settlement flow in turn; finally, online approval is initiated for settlement signature and seal; the following steps are included: (1) Settlement documents are electronically entered into the system process, and the payment conditions of each section are automatically collected to form a Fibonacci sequence with the payment condition hash value as the element; (2) Through the interval one-dimensional extreme value search analysis method, it is intelligently determined whether each settlement node of a settlement application process meets the conditions for generating a settlement application form; (3) Filter out the processes that meet the requirements of each settlement node and enter them into the settlement application message queue, and generate a settlement application form; (4) The queue is distributed and processed to initiate online approval. The settlement application form enters the electronic signature and seal online approval process and finally completes the settlement; Wherein, step (2) specifically includes: Before searching, the golden section characteristics are used to take the hash value of each node as the minimum value of the univariate function to form different objective functions for each node condition. Multiple node condition searches are assigned to univariate functions in different sections or intervals, and the interval one-dimensional search process is regarded as the process of solving the univariate function value. The univariate linear function obtains the minimum value through the simplex method; according to the different objective functions and the golden section characteristics, the minimum value is used as the bottom of the valley, and the valley is locked with a continuously shrinking range to obtain the minimum value; finally, the conditional characteristic value is obtained by screening and searching; When using interval-based one-dimensional search, the objective function is a convex function in the domain or in a certain interval, and the minimum value is unique; suppose a convex function f(x) on [a1, b1], and set the error coefficient e and the shortening ratio coefficient α, 0<α<1, and take a symmetrical subinterval [λ1, μ1] in the initial interval [a1, b1]. The symmetry means that the distance from λ1 to the endpoint a1 is equal to the distance from μ1 to the endpoint b1. To satisfy the symmetry, construct λ1 and μ1: suppose the distance from λ1 to the endpoint b1 is α times the length of the interval [a1, b1], and the distance from μ1 to the endpoint a1 is also α times the length of the interval [a1, b1]; After obtaining the positions of λ1 and μ1 and bringing λ1 and μ1 into the objective function, there are two situations: ①f(λ1)<=f(μ1), set the original endpoint b1 of the interval to μ1, and keep a1 unchanged, and obtain a new interval [a2, b2]=[a1, μ1], and the length of the new interval is shortened to α times the length of the original interval [a1, b1]; check the length of the new interval μ1-a1, if the length is less than e, exit the calculation, and the minimum point is (μ1+a1) / 2; if it is greater than the error coefficient e, take a symmetrical subinterval λ2, μ2 in the new interval [a1, μ1], and the selection rule is the same as the initial one, that is, the distance from λ2 to b2 is α times the length of the interval [a2, b2], and the distance from μ2 to the endpoint a2 is α times the length of the interval [a2, b2]; ②f(λ1)>f(μ1), set the original endpoint a1 of the interval to λ1, b1 remains unchanged, and obtain the new interval [a2, b2]=[λ1, b1], check the length of the new interval b1-λ1, if the length is less than e, exit the calculation, the minimum value is (λ1+b1) / 2; if it is greater than the error coefficient e, take a symmetric sub-interval [λ2, μ2] in the new interval [λ1, b1], the interval setting rules of [λ2, μ2] are the same as situation ①.

2. The method for intelligent settlement of queue processes based on interval one-dimensional extreme value search analysis according to claim 1 is characterized in that: Step (1) electronically inputs the offline paper documents of each stage into the system to form an information-based electronic document. The system forms a Fibonacci sequence of hash values ​​of the settlement conditions of each stage based on the input information and related system information, and automatically updates and improves the Fibonacci sequence of hash values ​​of the conditions of each stage on a regular basis.

3. The method for intelligent settlement of queue processes based on interval one-dimensional extreme value search and analysis according to claim 1 is characterized in that: Step (3) After the interval one-dimensional extreme value search analysis, each payment process that meets the settlement application enters the corresponding queue, completes the generation of the payment application form and initiates the online approval process.

4. The method for intelligent settlement of queue processes based on interval one-dimensional extreme value search analysis according to claim 1 is characterized in that: Step (4) The electronic signature system is used to select and process the signature and seal workflow to perform signature and seal operations on the electronic document, complete the online approval process and finally complete the settlement.