Rainfall profile inversion method and device based on parallel processing and dynamic resource allocation, medium and product

By classifying and allocating meteorological data and combining it with parallel processing technology, the problem of timeliness in precipitation profile inversion under limited hardware resources has been solved, enabling the rapid generation of precipitation products in three-dimensional or multi-dimensional data.

CN119025784APending Publication Date: 2024-11-26BEIJING HUAYUN SHINETEK TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411098627.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

With limited hardware resources, existing technologies are unable to quickly generate precipitation products based on three-dimensional or multi-dimensional data in engineering business systems, thus failing to meet production timeliness requirements.

Method used

A precipitation profile inversion method based on parallel processing and dynamic resource allocation is adopted. By classifying meteorological data, calculating the time and space complexity of each profile, allocating memory resources, and using multi-process and multi-threaded parallel processing, the inversion is performed.

Benefits of technology

With limited hardware resources, the timeliness of precipitation profile inversion has been improved, meeting the timeliness requirements of engineering business systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025784A_ABST
    Figure CN119025784A_ABST
Patent Text Reader

Abstract

The invention discloses a precipitation profile inversion method and device based on parallel processing and dynamic resource allocation, a medium and a product. Relates to the technical field of meteorological inversion. According to the characteristics of meteorological data, inversion tasks are analyzed, the time complexity and the space complexity of each task are calculated, then reasonable division of memory resources is realized based on the time complexity and the space complexity, and meteorological inversion is rapidly completed by adopting a multi-process and multi-thread combined parallel processing mode. The operation timeliness of a complex target processing object can be improved, and hardware resources are reasonably utilized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological inversion, and in particular to a precipitation profile inversion method, equipment, medium and product based on parallel processing and dynamic resource allocation. Background Art

[0002] In the field of vertical detection precipitation product generation in engineering business systems, with the continuous development of hardware equipment, the time and space resources required to process three-dimensional data or even multi-dimensional data with large data volumes and extremely complex algorithms are gradually increasing. For engineering business products that require runtime efficiency, it is difficult to increase space and memory from the hardware side when resources are limited, and the runtime can no longer meet the production timeliness requirements. Summary of the invention

[0003] The purpose of the present invention is to provide a precipitation profile inversion method, device, medium and product based on parallel processing and dynamic resource allocation, which can improve the timeliness of precipitation profile inversion under the condition of limited hardware resources.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A precipitation profile inversion method based on parallel processing and dynamic resource allocation, the method comprising:

[0006] Obtain meteorological data required for precipitation profile inversion;

[0007] Classifying the types of profiles in the meteorological data to determine the types of profiles contained in the meteorological data and the number of each type of profiles; the types of profiles include one or more of profiles without precipitation, profiles with certain precipitation, and profiles with possible precipitation;

[0008] Calculate the time complexity and space complexity of the profile inversion of each target type; the target type of profile is a profile that is determined to have precipitation and a profile that may have precipitation;

[0009] According to the time complexity and space complexity of the profile inversion of each target type, the time resource proportion and memory resource proportion required for the profile inversion of each target type are calculated;

[0010] According to the time resource proportion and memory resource proportion required for the profile inversion of each target type, the memory resources are divided to obtain multiple data blocks;

[0011] Using the plurality of data blocks to perform parallel inversion on the meteorological data, to obtain an inversion result for each data block;

[0012] Merge the inversion results of multiple data blocks.

[0013] Optionally, calculate the time complexity and space complexity of the profile inversion for each target type, including:

[0014] The time complexity of profile inversion for each target type is calculated using the following formula;

[0015]

[0016] O i (h) = f i (x 1,h )*f i (x 2,h )*...*f i (x n,h );

[0017] Among them, t i is the time complexity of the profile inversion of the i-th target type, H is the number of vertical layers of meteorological data, O i (h) is the time complexity of inverting the profile of the i-th target type based on the meteorological data of the h-th layer, x 1,h 、x 2,h and x n,h are the first, second and nth variables related to profile inversion of the meteorological data of the hth layer, respectively, and f i () is the relationship model between the time complexity of inverting the profile of the i-th target type;

[0018] The spatial complexity of profile inversion for each target type is calculated using the following formula;

[0019] s i =O i,1 (x)+O i,n1 (x)+O i,n2 (x)+...+O i,nm (x);

[0020] Among them, s i is the spatial complexity of the profile inversion of the i-th target type, O i,1 (x) represents the spatial complexity function of the main loop of the profile inversion for the i-th target type, O i,n1 (x), O i,n2 (x) and O i,nm (x) represents the spatial complexity function of the first nested loop, the second nested loop and the mth nested loop of the profile inversion of the i-th target type, respectively, and x is the meteorological data.

[0021] Optionally, the time resource proportion and memory resource proportion required for the profile inversion of each target type are calculated according to the time complexity and space complexity of the profile inversion of each target type, specifically including:

[0022] According to the time complexity of profile inversion for each target type, the time resource ratio required for profile inversion for each target type is calculated using the following formula;

[0023] tw i =t i / T all ;

[0024] T all =t 1 +t 2 +...+t I ;

[0025] Among them, tw i is the time resource ratio required for the profile inversion of the i-th target type, t i is the time complexity of the profile inversion of the i-th target type, T all is the total time complexity, t 1 ,t 2 and t I are the time complexity of inversion of the profiles of the first, second and I target categories, respectively, and I is the number of profiles of the target category that needs to be inverted;

[0026] According to the spatial complexity of the profile inversion of each target type, the proportion of memory resources required for the profile inversion of each target type is calculated using the following formula;

[0027] sw i =s i / S all ;

[0028] S all =s 1 +s 2 +...+s I ;

[0029] Among them, sw i is the proportion of memory resources required for the profile inversion of the i-th target type, s i is the spatial complexity of the profile inversion of the i-th target type, S all is the total space complexity, s 1 、s 1 and I are the spatial complexity of profile inversion for the first, second and I target types respectively.

[0030] Optionally, the memory resources are divided according to the time resource proportion and memory resource proportion required for the profile inversion of each target type with the goal of ensuring that the memory space used by each data block in the memory resources and the required running time are balanced.

[0031] A computer device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above method.

[0032] A computer-readable storage medium stores a computer program, which implements the above method when executed by a processor.

[0033] A computer program product comprises a computer program, and the computer program implements the above method when executed by a processor.

[0034] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0035] The present invention provides a precipitation profile inversion method, device, medium and product based on parallel processing and dynamic resource allocation. The present invention analyzes the inversion task according to the characteristics of meteorological data, and calculates the time complexity and space complexity of each task, and then realizes the reasonable division of memory resources based on the time complexity and space complexity, adopts a parallel processing method combining multi-process and multi-threading, quickly completes meteorological inversion, can improve the runtime efficiency of complex target processing objects, and reasonably use hardware resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A schematic flow chart of a precipitation profile inversion method based on parallel processing and dynamic resource allocation provided in an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of a precipitation profile inversion method based on parallel processing and dynamic resource allocation provided by an embodiment of the present invention;

[0039] Figure 3 An internal structure diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] The purpose of the present invention is to provide a precipitation profile inversion method, device, medium and product based on parallel processing and dynamic resource allocation, which can improve the timeliness of precipitation profile inversion under the condition of limited hardware resources.

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] Example 1

[0044] like Figure 1 and Figure 2 As shown, the precipitation profile inversion method based on parallel processing and dynamic resource allocation in this embodiment includes:

[0045] Step 101, obtaining meteorological data required for precipitation profile inversion.

[0046] Specifically, the data required for generating precipitation profile products of meteorological elements are obtained, and the data characteristics for generating precipitation profiles are analyzed, including data dimension, data volume, correlation between each profile data, whether parallelism is required, parallel operability and other target characteristics.

[0047] In practical applications, the precipitation dimension characteristics observed by the Fengyun-3 03G batch are three-dimensional. The amount of data required to calculate the precipitation rate profile elements is large and the algorithm is complex. Since there is little correlation when inverting the precipitation for each profile, considering the huge amount of calculation and time requirements, parallel processing can be adopted.

[0048] Step 102, identifying the profiles in the meteorological data, determining the types of profiles contained in the meteorological data, and the number of each type of profiles; the types of profiles include one or more of profiles without precipitation, profiles with certain precipitation, and profiles with possible precipitation.

[0049] Exemplarily, the acquisition of the profiles in the meteorological data is mainly obtained through observations of the Fengyun-3 Batch 03G satellite precipitation measurement radar payload.

[0050] Step 103, calculating the time complexity and space complexity of the profile inversion of each target type; the profile of the target type is a profile where precipitation is confirmed to exist and a profile where precipitation may exist.

[0051] In practical applications, for each precipitation profile, the complexity of the inversion algorithm is calculated; combined with the precipitation inversion characteristics, the amount of calculation required to determine the profile without precipitation is very small and can be ignored. For the profiles that are determined to have precipitation or may have precipitation, the time and memory required for the inversion process are calculated. Of course, the specific time length and memory occupied by running 1 particle need to be tested on a computer. The specific calculation formula is as follows:

[0052] In precipitation inversion processing, there is a calculation method that requires multiple loop nesting and iterations. The time complexity calculation formula can be expressed as:

[0053] t(n)=O 1 ((1+n1)*n2*...*nm)+O 2 (h);

[0054] Among them, n1, n2, ..., nm are the number of multi-layer nested loops, O 1 It is a function of the time complexity of calculating the precipitation rate on a single profile. This function is related to various complex weather factors such as precipitation possibility, precipitation type, and rain top height, so the calculation formula of time complexity is also different. Taking a single profile of a precipitation star as an example, a single profile has 400 altitude layers. The precipitation possibility and precipitation type at each altitude layer are different. The rain top height of the precipitation point on each profile is also different. In addition, it is also related to other precipitation factors such as bright bands, phases, and raindrop sizes on the profile layer. Therefore, the calculation function of time complexity on each altitude layer can be expressed as:

[0055] O i (h) = f i (x 1,h )*f i (x 2,h )*...*f i (x n,h );

[0056] Among them, t i is the time complexity of the profile inversion of the i-th target type, H is the number of vertical layers of meteorological data, O i (h) is the time complexity of inverting the profile of the i-th target type based on the meteorological data of the h-th layer, x 1,h 、x 2,h and x n,h are the first, second and nth variables related to profile inversion of the meteorological data of the hth layer, respectively, and f i () is a relationship model of the time complexity of inverting the profile of the i-th target type and the variable. The relationship model in the embodiment of the present invention can be obtained by training a neural network model based on historical inversion process data, or can be expressed by a regression function, which will not be described in detail here.

[0057] Specifically, x 1,h 、x 2,h and x n,h are variables related to precipitation calculation, such as precipitation possibility, precipitation type, and rain top height. f is a calculation formula related to the three factors. If precipitation conditions exist, the effects of the corresponding factors on time are superimposed, that is, the more complex the precipitation conditions are, the longer it takes.

[0058] In addition, the vertical observation of Mercury has 400 layers of data. 2 This is the time complexity function of the vertical height layer. It can be expressed as: For each pixel point, the corresponding single precipitation profile, each layer of data must be judged and calculated accordingly. Therefore, to achieve the precipitation inversion of the target object, the amount of calculation is quite large and time-consuming. The complexity of the calculation of each layer of meteorological factors of the above single profile and the complexity of the number of layers at the vertical height of the single profile constitute the time complexity of precipitation inversion.

[0059] When processing a single precipitation inversion, in addition to the storage space occupied by the necessary input data, there is also temporary storage space in multiple nested loops. What we need to calculate is the total size of these temporary spaces and output spaces. Therefore, the spatial complexity calculation formula is:

[0060] s i =O i,1 (x)+O i,n1 (x)+O i,n2 (x)+...+O i,nm (x);

[0061] Among them, s i is the spatial complexity of the profile inversion of the i-th target type, O i,1 (x) represents the spatial complexity function of the main loop of the profile inversion for the i-th target type, O i,n1 (x), O i,n2 (x) and O i,nm (x) respectively represent the spatial complexity functions of the first nested loop, the second nested loop, and the mth nested loop of the profile inversion of the i-th target type, and x is the meteorological data. The spatial complexity function in the embodiment of the present application can be obtained by training the neural network model based on the historical inversion process data, or can be represented by a regression function, which will not be described in detail here.

[0062] O i,1 (x), O i,n1 (x), O i,n2 (x), O i,nm(x) is related to the data type and data dimension size of the precipitation inversion in each cycle. Similarly, only the precipitation inversion profile is listed here as the calculation formula for the occupied space.

[0063] Step 104 , calculating the time resource proportion and memory resource proportion required for the profile inversion of each target type according to the time complexity and space complexity of the profile inversion of each target type.

[0064] In practical applications, it is necessary to complete the inversion operation of the entire orbital data of the three-dimensional precipitation of the precipitation star. According to the characteristics of the precipitation star data observation, the number of pixels in the entire orbital target observation area is approximately 229,274 pixels, among which the profiles that are determined to have no precipitation are not inverted, the required time and memory can be ignored, and the weight can be assigned to 0. For the calculation of the profiles where precipitation is determined or may exist, the running time and required memory are calculated according to the inversion method. The time and memory resources required for the inversion of all profiles where precipitation may occur or is determined to occur are accumulated respectively to obtain the total running time and total memory resources. Based on the total running time and total memory resources, the time weight and spatial weight value required for the inversion of each profile where precipitation may occur or is determined to occur are calculated. The specific formula for the total time required for the inversion of all profiles where precipitation may occur or is determined to occur is as follows:

[0065] T all =t 1 +t 2 +...+t I ;

[0066] Among them, T all is the total time complexity, t 1 ,t 2 and t I are the time complexity of inversion of the profiles of the first, second and I target types respectively, and I is the number of profiles of the target type that need to be inverted.

[0067] Furthermore, the target type of profile is a profile where precipitation may occur or is determined to occur. In this case, the calculation formula for the total time complexity is:

[0068] T all =t p1 +t p2 +...+t pn +t mp1 +t mp2 +...+t mpk ;

[0069] Among them, t p1 ,t p2 ,t pnThe time complexity of the inversion of the profiles where precipitation occurs is determined for the first, second and nth items, t mp1 ,t mp2 and t mpk are the time complexity of the inversion of the first, second and kth profiles where precipitation may occur, respectively. n and k are the number of profiles where precipitation occurs and may occur, respectively.

[0070] The total memory resource required for inversion of all profiles where precipitation may occur or is determined to occur is as follows:

[0071] S all =s 1 +s 2 +...+s I ;

[0072] Among them, S all is the total space complexity, s 1 、s 1 and I are the spatial complexity of profile inversion for the first, second and I target types respectively, and I is the number of profiles of the target type that need to be inverted.

[0073] Furthermore, the above target type of profile is a profile where precipitation may occur or is determined to occur. In this case, the calculation formula of the total spatial complexity is:

[0074] S all =s p1 +s p2 +...+s pn +s mp1 +s mp2 +...+s mpk ;

[0075] Among them, sp 1 、sp 2 and pn The spatial complexity of the inversion of the profiles where precipitation occurs is determined for the first, second and nth items, s mp1 、s mp2 and mpk are the spatial complexity of the inversion of the first, second and kth profiles where precipitation may occur.

[0076] Determine the time resource ratio required for profile inversion of each target type as follows:

[0077] tw i =t i / T all ;

[0078] Among them, tw iis the time resource ratio required for the profile inversion of the i-th target type, t i is the time complexity of profile inversion for the i-th target type.

[0079] Determine the proportion of memory resources required for profile inversion of each target type as follows:

[0080] sw i =s i / S all ;

[0081] Among them, sw i is the proportion of memory resources required for the profile inversion of the i-th target type, s i is the spatial complexity of the profile inversion for the i-th target type.

[0082] Furthermore, the time proportion formula required for inversion of each profile that determines the occurrence of precipitation is as follows:

[0083]

[0084] in, For the jth 1 The proportion of time resources required for the inversion of the profiles that determine the occurrence of precipitation, For the jth 1 The time complexity of the inversion of the profiles where precipitation occurs is determined.

[0085] The formula for the time ratio required for inversion of each possible precipitation profile is as follows:

[0086]

[0087] in, For the jth 2 The proportion of time resources required to invert the profiles where precipitation may occur, For the jth 2 The time complexity of inverting the profiles where precipitation may occur.

[0088] The formula for the proportion of memory resources required for inversion of each profile that determines the occurrence of precipitation is as follows:

[0089]

[0090] For the jth 1 The proportion of spatial resources required for the inversion of the profiles that determine precipitation occurrence, For the jth 1 The spatial complexity of the inversion of the profiles where precipitation occurs is determined.

[0091] The formula for the proportion of memory resources required for inversion of each possible precipitation profile is as follows:

[0092]

[0093] in, For the jth 2 The proportion of memory resources required to invert the possible precipitation profiles, For the jth 2 The spatial complexity of inverting the profiles where precipitation may occur.

[0094] Step 105, dividing the memory resources according to the time resource proportion and memory resource proportion required for the profile inversion of each target type to obtain a plurality of data blocks;

[0095] In this step, based on the given memory resources, which are generally more abundant, the target data characteristics and weight values ​​are combined during the processing to divide the data blocks, reasonably allocate the resources of each data block, and give each data block a corresponding location identifier.

[0096] Specifically, under the constraints of memory resources and timeliness requirements, the target data required for inverting the precipitation of meteorological elements is reasonably divided into blocks based on the precipitation inversion characteristics, and the geographical location of each precipitation inversion block is saved.

[0097] In practical applications, due to the limitation of hardware conditions, when this embodiment processes three-dimensional precipitation inversion, the number of data blocks divided is upper bounded. This embodiment takes the maximum number of data blocks, combines the time resource proportion and memory resource proportion required for each profile inversion calculated, divides the entire target processing data into blocks, and ensures that the memory resources and running time required for each data block are balanced with the remaining data blocks during block division, prioritizes the time weight ratio of each data block, and saves the geographical location identification of each precipitation profile. For example, the computing resources allocated in the precipitation star engineering business system are 13 servers, each server is configured with 256GB, the maximum number of CPU threads is limited to 25, and the timeliness requirement is within 20 minutes. It is necessary to process 229,274 profile data of the target data, of which there are about 198,000 no precipitation profiles, about 30,000 precipitation profiles, and about 3,000 possible precipitation profiles. The weight of the no precipitation profile is 0, and the number of profiles processed by each data block is based on the time weight t mpk ,t pn and spatial weight s mpk 、s pnThe calculation shows that the division ratio of the 13 data blocks in the specific target data is: 7:17:14:10:5:2:22:6:13:1:1:1:1. Similarly, how to set the number of threads for each data block also depends on the time weight and space weight. Finally, we set the number of threads for processing the 13 data blocks according to the above method: 14:25:25:20:10:4:25:12:25:2:2:2:2. Since the 13 data blocks are processed in parallel, the total running time is about 10 minutes, which meets the timeliness requirements. It should be noted that the division of each data block can be regular or irregular, provided that the location identifier of each data can be retained to facilitate the synthesis of the three-dimensional precipitation data of the entire track in the later stage.

[0098] Step 106: Use the multiple data blocks to perform parallel inversion on the meteorological data to obtain an inversion result for each data block.

[0099] In this step, multi-process and multi-threaded parallel processing is used to accelerate target data processing without affecting its generation algorithm.

[0100] Specifically, according to the data blocks divided in step S4, a multi-process method is used to perform parallel operations on each precipitation data block, and each data block can be divided into small blocks according to weight distribution, and a thread is opened to process each small data block.

[0101] In practical applications, the number of processes is often limited by the memory resources themselves, that is, the server itself, while the number of threads depends on the data block itself. For data blocks with small data computation, the more threads there are, the more wasteful the running resources will be and the lower the time efficiency will be. Similarly, if the data computation is large and the number of threads is insufficient, the running time will be prolonged. Therefore, when choosing the number of threads, it is necessary to refer to relevant factors such as time weight and space weight. When memory resources allow, a reasonable number of threads can effectively allocate resources and improve running efficiency.

[0102] Step 107: merge the inversion results of multiple data blocks.

[0103] In this step, after all data blocks are processed, the data blocks are merged to finally generate complete target data.

[0104] Specifically, it is necessary to synthesize the small precipitation data blocks processed by each thread, and then combine the large precipitation data blocks processed by each process one by one according to the geographic location identifier, and finally generate the precipitation data of the entire track.

[0105] The method of this embodiment analyzes the characteristics of the target data, calculates data operation weights and performs data segmentation, dynamically allocates memory resources, and adopts a parallel processing method combining multi-process and multi-threading to quickly complete the processing of the target data. While reasonably using resources, it ensures the integrity and reliability of the target data processing results and meets the timeliness requirements.

[0106] Example 2

[0107] A computer device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method in Embodiment 1.

[0108] The computer device may be a database, and its internal structure diagram may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store pending transactions. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the method in Example 1 is implemented.

[0109] Example 3

[0110] A computer-readable storage medium stores a computer program, which implements the method in embodiment 1 when executed by a processor.

[0111] Example 4

[0112] A computer program product includes a computer program, and the computer program implements the method in embodiment 1 when executed by a processor.

[0113] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided by the present invention may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited thereto. The processor involved in each embodiment provided by the present invention may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, etc., but are not limited thereto.

[0114] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0115] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A precipitation profile inversion method based on parallel processing and dynamic resource allocation, characterized in that: The method comprises: Obtain meteorological data required for precipitation profile inversion; Identifying the profiles in the meteorological data, determining the types of profiles contained in the meteorological data, and the number of each type of profiles; the types of profiles include one or more of profiles without precipitation, profiles with certain precipitation, and profiles with possible precipitation; Calculate the time complexity and space complexity of the profile inversion of each target type; the target type of profile is a profile that is determined to have precipitation and a profile that may have precipitation; According to the time complexity and space complexity of the profile inversion of each target type, the time resource proportion and memory resource proportion required for the profile inversion of each target type are calculated; According to the time resource proportion and memory resource proportion required for the profile inversion of each target type, the memory resources are divided to obtain multiple data blocks; Using the plurality of data blocks to perform parallel inversion on the meteorological data, to obtain an inversion result for each data block; Merge the inversion results of multiple data blocks.

2. The precipitation profile inversion method based on parallel processing and dynamic resource allocation according to claim 1 is characterized in that: Calculate the time complexity and space complexity of the profile inversion for each target type, including: The time complexity of profile inversion for each target type is calculated using the following formula; O i (h)=f i (x 1,h )*f i (x 2,h )*...*f i (x n,h ); Among them, t i is the time complexity of the profile inversion of the i-th target type, H is the number of vertical layers of meteorological data, O i (h) is the time complexity of inverting the profile of the i-th target type based on the meteorological data of the h-th layer, x 1,h 、x 2,h and x n,h are the first, second and nth variables related to profile inversion of the meteorological data of the hth layer, respectively, and f i () is the relationship model between the time complexity of inverting the profile of the i-th target type; The spatial complexity of profile inversion for each target type is calculated using the following formula; s i =O i,1 (x)+O i,n1 (x)+O i,n2 (x)+...+O i,nm (x); Among them, s i is the spatial complexity of the profile inversion of the i-th target type, O i,1 (x) represents the spatial complexity function of the main loop of the profile inversion for the i-th target type, O i,n1 (x), O i,n2 (x) and O i,nm (x) represents the spatial complexity function of the first nested loop, the second nested loop and the mth nested loop of the profile inversion of the i-th target type, respectively, and x is the meteorological data.

3. The precipitation profile inversion method based on parallel processing and dynamic resource allocation according to claim 1, characterized in that: According to the time complexity and space complexity of the profile inversion of each target type, the time resource proportion and memory resource proportion required for the profile inversion of each target type are calculated, including: According to the time complexity of profile inversion for each target type, the time resource ratio required for profile inversion for each target type is calculated using the following formula; tw i =t i / T all ; T all =t1+t2+...+t I ; Among them, tw i is the time resource ratio required for the profile inversion of the i-th target type, t i is the time complexity of the profile inversion of the i-th target type, T all is the total time complexity, t1, t2 and t I are the time complexity of inversion of the profiles of the first, second and I target categories, respectively, and I is the number of profiles of the target category that needs to be inverted; According to the spatial complexity of the profile inversion of each target type, the proportion of memory resources required for the profile inversion of each target type is calculated using the following formula; sw i =s i / S all ; S all =s1+s2+...+s I ; Among them, sw i is the proportion of memory resources required for the profile inversion of the i-th target type, s i is the spatial complexity of the profile inversion of the i-th target type, S all is the total space complexity, s1, s1 and s I The spatial complexity of the profile inversion for the first, second and I target categories respectively.

4. The precipitation profile inversion method based on parallel processing and dynamic resource allocation according to claim 1, characterized in that: According to the time resource proportion and memory resource proportion required for the profile inversion of each target type, the goal of dividing the memory resources is to ensure that the memory space used by each data block in the memory resources and the required running time are balanced.

5. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.