CN117112245B

CN117112245B - Multithreading synchronous creation method and system based on equipment linkage

Info

Publication number: CN117112245B
Application number: CN202311376134.9A
Authority: CN
Inventors: 谢莎; 郑炜南; 陈柳园; 刘剑明
Original assignee: Shenzhen Jitong Intelligent Technology Co ltd
Current assignee: Shenzhen Jitong Intelligent Technology Co ltd
Priority date: 2023-10-23
Filing date: 2023-10-23
Publication date: 2023-12-26
Anticipated expiration: 2043-10-23
Also published as: CN117112245A

Abstract

The invention relates to the technical field of thread synchronization, and discloses a multithreading synchronization creation method based on equipment linkage, which comprises the following steps: splitting, cleaning and thread synchronous carding are carried out on the pre-acquired linkage data of the historical equipment, so as to obtain a linkage thread state set; respectively extracting an equipment thread state set and a linkage thread feature set from the linkage thread state set; training the thread time sequence model into a thread analysis model by utilizing the linkage thread characteristic set and the equipment thread state set; analyzing the real-time equipment thread state corresponding to the real-time equipment linkage data by using a thread analysis model; and synchronously matching the thread state of the real-time equipment according to a thread synchronization mechanism of the linkage data of the real-time equipment to obtain a real-time standard thread state, and carrying out multithreading synchronous creation according to the real-time standard thread state. The invention also provides a multithreading synchronous creation system based on equipment linkage. The invention can improve the efficiency based on multi-thread synchronization.

Description

Multithreading synchronous creation method and system based on equipment linkage

Technical Field

The invention relates to the technical field of thread synchronization, in particular to a multithreading synchronization creation method and system based on equipment linkage.

Background

The equipment linkage refers to that different equipment are connected together through a network, a protocol or other modes, data exchange, cooperation or coordination work can be carried out, and with the establishment of an information society, the equipment linkage plays a role in more and more fields, and in the equipment linkage process, the joint task of multiple equipment is required to be executed by utilizing a multi-thread synchronous establishment mode.

The existing multithread synchronous creation method is mostly a thread synchronous method based on task demand carding, namely threads required by each device are determined according to task demands, the creation sequence and creation time of each thread of each device are determined according to a synchronous mechanism among the threads, and in practical application, the thread synchronous method based on task demand carding may require a large amount of logic analysis and thread debugging by personnel, which may result in lower efficiency in multithread synchronization.

Disclosure of Invention

The invention provides a multithreading synchronization creation method and system based on equipment linkage, and mainly aims to solve the problem of low efficiency in multithreading synchronization.

In order to achieve the above object, the present invention provides a multithreading synchronization creation method based on device linkage, including:

Splitting and cleaning the pre-acquired historical equipment linkage data to obtain a historical linkage data set, and carrying out thread synchronous carding on the historical linkage data set to obtain a linkage thread state set;

extracting a linkage equipment set, a linkage thread set and an equipment thread state set from the linkage thread state set respectively, generating a linkage thread feature set according to the linkage equipment set and the linkage thread set, wherein the generating the linkage thread feature set according to the linkage equipment set and the linkage thread set comprises the following steps: selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups; sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features; extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method; and calculating the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature by using the following shared weight algorithm: Wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the thread name is specialFeature total dimension of the feature, and feature total dimension of the thread name feature, the thread resource feature and the thread communication feature are equal, +.>Is a hyperbolic tangent function, ">Refers to the +.>Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++>Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics; performing full-connection operation on the shared thread features and the target equipment features to obtain linkage thread features, and integrating all linkage thread features into a linkage thread feature set;

analyzing a predicted equipment thread state set corresponding to the linkage thread feature set by using a preset thread time sequence model, and training the thread time sequence model into a thread analysis model by using the equipment thread state set and the predicted equipment thread state set;

acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, and analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using the thread analysis model;

And extracting a thread synchronization mechanism from the real-time equipment linkage data, performing synchronization matching on the thread state of the real-time equipment according to the thread synchronization mechanism to obtain a real-time standard thread state, and performing multi-thread synchronization creation according to the real-time standard thread state.

Optionally, the splitting and cleaning the pre-acquired linkage data of the historical equipment to obtain a historical linkage data set includes:

splitting the historical equipment linkage data into a periodic linkage data set according to a preset linkage period;

extracting a linkage name data set from the periodic linkage data set, and performing disorder code detection on the linkage name data set to obtain a disorder code name data set;

performing data denoising on the periodic linkage data set according to the messy code name data set to obtain a denoising linkage data set;

screening the messy code name data set from the linkage name data set to obtain a denoising name data set, and hashing the denoising name data set into a hash name data set;

and carrying out data deduplication on the denoising linkage data set by using the hash name data set to obtain a historical linkage data set.

Optionally, the performing thread synchronization carding on the historical linkage data set to obtain a linkage thread state group set includes:

Selecting the historical linkage data in the historical linkage data set one by one as target linkage data, and extracting a target equipment name group from the target linkage data;

splitting the target linkage data into a target equipment data set according to the target equipment name set;

extracting a target thread state group from the target equipment data group;

and carrying out time sequence sequencing on each thread state in the target thread state group to obtain a target linkage thread state group, and collecting all the target linkage thread state groups into a linkage thread state group set.

Optionally, the extracting the linkage device group set, the linkage thread group set and the device thread state set from the linkage thread state group set respectively includes:

selecting the linkage thread state groups in the linkage thread state group set one by one as target linkage thread state groups, and searching the device names of the target linkage thread state groups to obtain linkage device groups;

performing thread retrieval on the target linkage thread state group to obtain a linkage thread group;

performing equipment marking on the linkage thread group by utilizing the linkage equipment group to obtain a primary equipment thread;

Performing time stamp marking on the equipment thread mark by utilizing the linkage thread state group set to obtain a secondary equipment thread;

sequencing threads in the secondary equipment threads according to a time sequence to obtain equipment thread states;

all linkage equipment groups are gathered into a linkage equipment group set, all linkage thread groups are gathered into a linkage thread group set, and all equipment thread states are gathered into an equipment thread state set.

Optionally, the method for extracting the thread name feature, the thread resource feature and the thread communication feature from the target linkage thread group respectively includes:

extracting a thread name group, a thread resource group and a thread communication group from the target linkage thread group respectively;

sequentially carrying out text vectorization and feature joint mapping operation on the thread name group to obtain primary thread name features;

vectorizing the thread resource group into a resource name feature group, and carrying out feature aggregation on the resource name feature group to obtain an aggregate resource feature class group;

performing feature mapping operation on the resource name feature group by using a clustering center of the aggregate resource feature group to obtain a standard resource feature group, and jointly mapping the standard resource feature group into primary thread resource features;

Vectorizing the thread communication group into a communication name feature group, and carrying out feature aggregation on the communication name feature group to obtain an aggregate communication feature group;

performing feature mapping operation on the communication name feature group by using a clustering center of the aggregate communication feature group to obtain a standard communication feature group, and jointly mapping the standard communication feature group into primary thread communication features;

and performing overall dimension reduction on the primary thread name feature to form a thread name feature, performing overall dimension reduction on the primary thread resource feature to form a thread resource feature, and performing overall dimension reduction on the primary thread communication feature to form a thread communication feature.

Optionally, the analyzing, by using a preset thread timing model, a prediction device thread state set corresponding to the linked thread feature set includes:

selecting linkage thread features in the linkage thread feature set one by one as target linkage thread features, and respectively extracting thread long-term features and thread short-term features from the target linkage thread features by using a preset thread time sequence model;

and carrying out feature fusion on the long-term features of the threads and the short-term features of the threads by using the following long-short fusion formula to obtain the time sequence features of the threads: Wherein (1)>Is +.>Characteristic value of time>Is the time sequence number->Is the weight factor preset by the long and short fusion formula,/->Means +.>Characteristic value of time>Is the number of hidden units of the jump layer of the thread timing model, < >>、/>Is the unit number->Is the +.>Characteristic weight of time of day->Means +.>Characteristic value of time>Is the number of hidden units of the recursive level of the thread timing model,/for the thread timing model>Is the +.o in the thread short-term feature>Characteristic weight of time of day->Means +.>Characteristic value of time>Is a bias factor preset by the long and short fusion formula;

and performing feature decoding on the thread time sequence features to obtain the thread state of the prediction equipment, and collecting all the thread states of the prediction equipment into a thread state set of the prediction equipment.

Optionally, the extracting the long-term thread feature and the short-term thread feature from the target linkage thread feature by using a preset thread timing model includes:

performing frequency domain transformation on the target linkage thread characteristics by using a preset thread time sequence model to obtain target thread frequency domain characteristics;

Carrying out peak period analysis on the frequency domain characteristics of the target thread to obtain thread period characteristics;

utilizing the jump layer of the thread time sequence model and the thread cycle characteristic to carry out jump characteristic extraction on the target linkage thread characteristic to obtain a thread long-term characteristic;

and carrying out recursive feature extraction on the target linkage thread features by using a recursive layer of the thread time sequence model to obtain thread short-term features.

Optionally, the training the thread timing model into a thread analysis model using the device thread state set and the predictive device thread state set includes:

the overall pooling layer of the thread time sequence model is utilized to dimension the equipment thread state set into a thread state feature set, and dimension the prediction equipment thread state set into a prediction state feature set;

calculating a state loss value of the thread timing model from the thread state feature set and the predicted state feature set using a state loss function as follows:wherein (1)>Means the state loss value, +.>Refers to the total number of features of the thread state feature set, and the total number of features of the thread state feature set is equal to the total number of features of the prediction state feature set,/ >Is a characteristic sequence number->Is the +.>Thread state feature->Is the +.>The number of predicted state characteristics is determined,as a covariance function;

judging whether the state loss value is larger than a preset state loss threshold value or not;

if not, carrying out iterative updating on the model parameters of the thread time sequence model according to the state loss value, and returning to the step of analyzing the prediction equipment thread state set corresponding to the linkage thread feature set by using the preset thread time sequence model;

if yes, the updated thread time sequence model is used as a thread analysis model.

Optionally, the step of synchronously matching the thread state of the real-time device according to the thread synchronization mechanism to obtain a real-time standard thread state includes:

sequentially extracting a mutual exclusion lock and a condition variable from the thread synchronization mechanism;

extracting a synchronization point from the thread state of the real-time equipment according to the mutual exclusion lock and the condition variable;

performing primary sequencing on the thread state of the real-time equipment according to the synchronization point to obtain a real-time primary thread state;

and carrying out deadlock sequencing on the real-time primary thread state to obtain a real-time standard thread state.

In order to solve the above problems, the present invention further provides a multithreading synchronization creation system based on device linkage, the system comprising:

the synchronous carding module is used for splitting and cleaning the pre-acquired linkage data of the historical equipment to obtain a historical linkage data set, and carrying out thread synchronous carding on the historical linkage data set to obtain a linkage thread state group set;

the feature fusion module is configured to extract a linkage device group set, a linkage thread group set and a device thread state set from the linkage thread state set, and generate a linkage thread feature set according to the linkage device group set and the linkage thread group set, where the generating a linkage thread feature set according to the linkage device group set and the linkage thread group set includes: selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups; sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features; extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method; using the following sharing weights The recalculation method calculates the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature:wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the feature total dimension of the thread name feature is equal to the feature total dimension of the thread name feature, the thread resource feature and the thread communication feature, < >>Is a hyperbolic tangent function, ">Refers to the +.>Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++>Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics; performing full-connection operation on the shared thread features and the target equipment features to obtain linkage thread features, and integrating all linkage thread features into a linkage thread feature set;

the model training module is used for analyzing a predicted equipment thread state set corresponding to the linkage thread feature set by using a preset thread time sequence model, and training the thread time sequence model into a thread analysis model by using the equipment thread state set and the predicted equipment thread state set;

The model analysis module is used for acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, and analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using the thread analysis model;

and the synchronous matching module is used for extracting a thread synchronous mechanism from the real-time equipment linkage data, synchronously matching the thread state of the real-time equipment according to the thread synchronous mechanism to obtain a real-time standard thread state, and carrying out multi-thread synchronous creation according to the real-time standard thread state.

According to the invention, the historical linkage data acquired in advance is split and cleaned to obtain the historical linkage data set, so that data grouping can be realized, subsequent model training is facilitated, meanwhile, error and repeated data are screened out from the data, the accuracy of the model training is improved, the linkage thread state set is obtained through carrying out thread synchronous carding on the historical linkage data set, formatting processing can be carried out on the thread state according to time sequence, further subsequent feature extraction is facilitated, the linkage thread state set, the linkage thread set and the equipment thread state set are respectively extracted from the linkage thread state set, the linkage thread feature set is generated according to the linkage thread set and the linkage thread set, the independent variables of equipment and threads during equipment linkage can be normalized, the feature quantity is reduced, further the efficiency of the subsequent model training is improved, the thread state set of the prediction equipment corresponding to the linkage thread feature set is analyzed by utilizing a preset thread time sequence model, the thread state set of the prediction equipment is trained into an analysis model by utilizing the equipment thread state set and the prediction equipment thread state set, and the thread synchronous state relation between equipment and the equipment can be conveniently established by utilizing the analysis of the thread state of the prediction equipment, and the thread synchronous state is convenient to establish the thread synchronous state.

The method comprises the steps of acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using a thread analysis model, extracting real-time equipment thread states corresponding to the real-time equipment linkage data by using threads, equipment and thread synchronization creation states in the thread analysis model, thereby facilitating subsequent multithread synchronization creation, improving multithread synchronization efficiency, synchronously matching the real-time equipment thread states according to a thread synchronization mechanism by extracting the thread synchronization mechanism from the real-time equipment linkage data to obtain real-time standard thread states, and carrying out multithread synchronization creation according to the real-time standard thread states, wherein the sequence and time of multithread synchronization creation during equipment linkage can be determined according to thread names and equipment names during real-time multi-equipment linkage, deadlock and thread logic errors are avoided, and multithread synchronization creation efficiency is improved. Therefore, the multithreading synchronization creation method and system based on equipment linkage can solve the problem of low efficiency in multithreading synchronization.

Drawings

FIG. 1 is a schematic flow chart of a multithreading synchronization creation method based on device linkage according to an embodiment of the present invention;

FIG. 2 is a flowchart illustrating a method for extracting a linked thread state set according to an embodiment of the present invention;

FIG. 3 is a flowchart illustrating a process for extracting long-term features and short-term features of threads according to an embodiment of the present invention;

FIG. 4 is a functional block diagram of a device linkage based multi-threaded synchronization creation system according to an embodiment of the present invention;

the achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.

Detailed Description

It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

The embodiment of the application provides a multithreading synchronous creation method based on equipment linkage. The execution subject of the device linkage-based multithreading synchronization creation method includes, but is not limited to, at least one of a server, a terminal, and the like, which can be configured to execute the method provided by the embodiment of the application. In other words, the device linkage-based multithreading synchronization creation method may be performed by software or hardware installed in a terminal device or a server device, where the software may be a blockchain platform. The service end includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like. The server may be an independent server, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

Referring to fig. 1, a flow chart of a multithreading synchronization creation method based on device linkage according to an embodiment of the present invention is shown. In this embodiment, the method for creating multithreading synchronization based on device linkage includes:

s1, splitting and cleaning pre-acquired linkage data of historical equipment to obtain a historical linkage data set, and carrying out thread synchronous carding on the historical linkage data set to obtain a linkage thread state group set.

In the embodiment of the invention, the historical equipment linkage data refer to data such as time stamps and execution calls created by each thread of each equipment in the process of equipment linkage of multiple equipment in the past time period, and each historical linkage data in the historical linkage data set comprises the data such as time stamps and thread execution calls created by each thread of multiple equipment in the past sequential equipment linkage.

In detail, the device linkage (Device Interconnectivity) refers to that different devices are connected together through a network, a protocol or other modes to perform data exchange, cooperation or coordination work, for example, devices in a smart home can exchange information and instructions mutually through internet connection.

In the embodiment of the present invention, splitting and cleaning the previously acquired linkage data of the historical equipment to obtain a historical linkage data set includes:

In detail, the linkage period refers to a time period for coordination and cooperation between a plurality of devices each time, namely, a time period from the beginning of device linkage to the end of device linkage, and each period linkage data in the period linkage data set comprises related records created by line synchronization in device linkage between a plurality of devices in one linkage period time.

Specifically, the linkage name data set is a data set composed of a device name, a thread name and a corresponding start-stop timestamp, which are included in each period linkage data in the period linkage data set, the linkage name data set can be extracted from the period linkage data set by means of regular expressions, keyword matching and the like, disorder code detection is carried out on the linkage name data set, disorder code name data set is obtained by collecting data with abnormal data types and disorder code name data set, wherein the disorder code name data set can be obtained by carrying out disorder code detection on the linkage name data set by using a charset library in Python or detection tools such as Z-score and IQR, and the like.

In detail, the step of performing data denoising on the periodic linkage data set according to the messy code name data set to obtain a denoising linkage data set refers to taking data corresponding to the messy code name data set in the periodic linkage data set as a messy code linkage data set, and screening out the messy code linkage data set from the periodic linkage data set to obtain the denoising linkage data set.

Specifically, the denoising name data set may be hashed into a hash name data set by using an MD5 algorithm, a SHA-1 algorithm or a BLAKE2 algorithm, and the performing data deduplication on the denoising linkage data set by using the hash name data set to obtain a history linkage data set refers to taking repeated data of the hash name data in the hash name data set as a repeated name data set, taking linkage data corresponding to the repeated name data set in the denoising linkage data set as a repeated linkage data set, and screening out the repeated linkage data set in the denoising linkage data set to obtain the history linkage data set.

In detail, referring to fig. 2, the process of performing thread synchronous carding on the historical linkage data set to obtain a linkage thread state group set includes:

s21, selecting the historical linkage data in the historical linkage data set one by one as target linkage data, and extracting a target equipment name group from the target linkage data;

s22, splitting the target linkage data into a target equipment data group according to the target equipment name group;

s23, extracting a target thread state group from the target equipment data group;

s24, performing time sequence sequencing on each thread state in the target thread state group to obtain a target linkage thread state group, and collecting all the target linkage thread state groups into a linkage thread state group set.

Specifically, the target device name group is a data group formed by device names of devices in the target linkage data, and the target device name group can be extracted from the target linkage data by using a regular expression or keyword matching and other modes.

In detail, splitting the target linkage data into the target device data set according to the target device name set refers to collecting all device data into the device data set by taking data corresponding to each device name in the target linkage data as device data.

Specifically, the extracting the target thread state group from the target device data group refers to collecting thread states in each target device data in the target device data group into a target thread state group, wherein the thread states refer to each thread created by one target device data in one linkage period and corresponding timestamp and other data, the sequentially ordering each thread state in the target thread state group to obtain a target linkage thread state group refers to ordering each thread in each thread state in the target thread state group according to the time sequence of creating the timestamp of the thread to obtain a linkage thread state, collecting all linkage thread states corresponding to the target thread state group into a linkage thread state group, and collecting all linkage thread state groups into a linkage thread state group set.

In the embodiment of the invention, the historical linkage data set is obtained by splitting and cleaning the pre-acquired historical equipment linkage data, so that data grouping can be realized, subsequent model training is convenient, meanwhile, error and repeated data are screened out from the data, the accuracy of model training is improved, and the linkage thread state set is obtained by carrying out thread synchronous carding on the historical linkage data set, and further, the formatting treatment can be carried out on the thread state according to time sequence, so that subsequent feature extraction is convenient.

S2, respectively extracting a linkage equipment group set, a linkage thread group set and an equipment thread state set from the linkage thread state set, and generating a linkage thread feature set according to the linkage equipment group set and the linkage thread group set.

In the embodiment of the invention, the linkage equipment group set is a set formed by a plurality of linkage equipment groups, each linkage equipment group comprises equipment names of all linkage thread states in the corresponding linkage thread state group in the linkage thread state group set, each linkage thread group in the linkage thread group set comprises threads of all linkage threads in the corresponding linkage thread state group in the linkage thread state group set, and each equipment thread state in the equipment thread state group corresponds to the states of all threads of all equipment in time sequence arrangement in each linkage thread state group in the linkage thread state group set.

In the embodiment of the present invention, the extracting the linkage equipment group set, the linkage thread group set and the equipment thread state set from the linkage thread state set respectively includes:

In detail, the device name, such as a Printer (Printer), a Monitor (Monitor), or a Mouse (Mouse), and the thread name, such as explorer. Exe, wiword. Exe, mysql. Exe, or python. Exe, may perform device retrieval on the target linkage thread state group by using an algorithm such as a regular expression or keyword matching, to obtain a linkage device group, and perform thread name retrieval on the target linkage thread state group to obtain a linkage thread group.

In detail, the device labeling of the linkage thread group by using the linkage device group, obtaining a primary device thread, namely, labeling each thread in the linkage thread group by using a device name corresponding to each linkage device in the linkage device group, and performing time stamp labeling on the device thread label by using the linkage thread state group set, namely, performing time stamp labeling on each thread in the device thread label by using a time stamp of each thread in the linkage thread state group, so as to obtain a secondary device thread.

In detail, the generating the linkage thread feature set according to the linkage equipment group set and the linkage thread group set includes:

selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups;

sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features;

extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method;

and calculating the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature by using the following shared weight algorithm:wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the feature total dimension of the thread name feature is equal to the feature total dimension of the thread name feature, the thread resource feature and the thread communication feature, < >>Is a hyperbolic tangent function, ">Refers to the +. >Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++>Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics;

and performing full-connection operation on the shared thread feature and the target equipment feature to obtain a linkage thread feature, and integrating all linkage thread features into a linkage thread feature set.

Specifically, the shared thread characteristics corresponding to the thread name characteristics, the thread resource characteristics and the thread communication characteristics are calculated by utilizing the shared weight algorithm, so that the method can be better adapted to the data distribution of complex modes, the effect of better feature fusion is achieved, and the model training efficiency is improved.

In detail, the sequentially performing text vectorization and feature joint mapping operations on the target linkage equipment set to obtain target equipment features refers to performing text vectorization operations on each equipment name in the target linkage equipment set after being arranged according to a fixed sequence to obtain equipment name feature sets, and performing feature mapping operations on the equipment name feature sets to obtain target equipment features, wherein Word2Vec or Doc2Vec algorithm can be used for text vectorization on the target linkage equipment set.

Specifically, the method for extracting thread name features, thread resource features and thread communication features from the target linkage thread group respectively by using feature aggregation includes:

In detail, each thread name in the thread name group corresponds to a thread name of a thread in each linkage thread in the target linkage thread group, each thread resource in the thread resource group corresponds to a resource name of a thread in each linkage thread in the target linkage thread group, wherein the resource name can be a name of a resource such as a stack space, a local storage and the like, and each thread communication in the thread communication group corresponds to a communication name of a thread in each linkage thread in the target linkage thread group, wherein the communication name can be a name of a signal quantity, a mutual exclusion lock and message passing.

Specifically, the method for sequentially performing text vectorization and feature joint mapping operations on the thread name group to obtain the primary thread name feature is consistent with the method for sequentially performing text vectorization and feature joint mapping operations on the target linkage equipment group in the step S2 to obtain the target equipment feature, which is not described herein.

In detail, the resource name feature group can be subjected to feature aggregation by using a k-means clustering algorithm, a hierarchical clustering algorithm or a DBSCAN clustering algorithm to obtain an aggregate resource feature group, the primary thread name feature can be subjected to overall dimension reduction into thread name feature by using a full connection layer, the primary thread resource feature can be subjected to overall dimension reduction into thread resource feature, and the primary thread communication feature can be subjected to overall dimension reduction into thread communication feature.

In the embodiment of the invention, the linkage equipment group set, the linkage thread group set and the equipment thread state set are respectively extracted from the linkage thread state set, and the linkage thread feature set is generated according to the linkage equipment group set and the linkage thread group set, so that independent variables of equipment and threads during equipment linkage can be normalized, the feature quantity is reduced, and further the efficiency of subsequent model training is improved.

S3, analyzing a predicted equipment thread state set corresponding to the linkage thread feature set by using a preset thread time sequence model, and training the thread time sequence model into a thread analysis model by using the equipment thread state set and the predicted equipment thread state set.

In the embodiment of the invention, the thread timing model may be a cyclic neural network (Recurrent Neural Network, abbreviated as RNN), a Long Short-Term Memory (LSTM), a Transformer (transducer), or the like.

In the embodiment of the present invention, the analyzing, by using a preset thread timing model, a thread state set of a prediction device corresponding to the linked thread feature set includes:

and carrying out feature fusion on the long-term features of the threads and the short-term features of the threads by using the following long-short fusion formula to obtain the time sequence features of the threads:wherein (1)>Is +.>Characteristic value of time>Is the time sequence number->Is the weight factor preset by the long and short fusion formula,/->Means +.>Characteristic value of time>Is the number of hidden units of the jump layer of the thread timing model, < >>、/>Is the unit number->Is the +.>Characteristic weight of time of day->Means +.>Characteristic value of time>Is the number of hidden units of the recursive level of the thread timing model,/for the thread timing model>Is the +.o in the thread short-term feature>Characteristic weight of time of day- >Means the thread short term feature +.>Characteristic value of time in->Is a bias factor preset by the long and short fusion formula;

Specifically, the long-term characteristics of the threads and the short-term characteristics of the threads are subjected to characteristic fusion by utilizing the long-short fusion formula to obtain the time sequence characteristics of the threads, so that the model can consider more time scales and characteristics, and further, the complex mode in the time sequence data can be captured better.

Specifically, referring to fig. 3, the extracting, by using a preset thread timing model, a long-term thread feature and a short-term thread feature from the target linked thread feature includes:

s31, performing frequency domain transformation on the target linkage thread characteristics by using a preset thread time sequence model to obtain target thread frequency domain characteristics;

s32, carrying out peak period analysis on the frequency domain characteristics of the target thread to obtain thread period characteristics;

s33, utilizing the jump layer of the thread time sequence model and the thread cycle characteristic to carry out jump characteristic extraction on the target linkage thread characteristic to obtain a thread long-term characteristic;

S34, performing recursion feature extraction on the target linkage thread features by using a recursion layer of the thread time sequence model to obtain thread short-term features.

In detail, the frequency domain transformation of the target linkage thread feature can be performed by using a fast fourier transform formula or a wavelet transform formula in a preset thread time sequence model to obtain a target thread frequency domain feature, and the peak period analysis is performed on the target thread frequency domain feature to obtain a thread period feature, namely, observing a peak frequency domain with periodicity in a plurality of target thread frequency domain features, and taking a corresponding time sequence period as the thread period feature.

Specifically, the jump layer and the recursion layer are network layers composed of an input gate, a forgetting gate and an output gate in a Long Short-Term Memory network (LSTM), and a Decoder (transcoder) of a Transformer model can be used for performing feature decoding on the thread time sequence features to obtain the thread state of the prediction device.

In detail, the training the thread timing model into a thread analysis model using the device thread state set and the predictive device thread state set includes:

Calculating a state loss value of the thread timing model from the thread state feature set and the predicted state feature set using a state loss function as follows:wherein (1)>Means the state loss value, +.>Refers to the total number of features of the thread state feature set, and the lineThe total number of features of the program state feature set is equal to the total number of features of the prediction state feature set,/>Is a characteristic sequence number->Is the +.>Thread state feature->Is the +.>Predicted status features->As a covariance function;

In detail, by calculating the state loss value of the thread timing model according to the thread state feature set and the predicted state feature set by using the state loss function, the model parameters of the thread timing model can be iteratively updated according to the state loss value by using a fast gradient descent algorithm or a random gradient descent algorithm according to the similarity measure between vectors, the similarity measure of covariance matrix and the difference between euclidean distances more comprehensively measured data.

In the embodiment of the invention, the thread state set of the prediction equipment corresponding to the linkage thread feature set is analyzed by utilizing the preset thread time sequence model, the thread time sequence model is trained into the thread analysis model by utilizing the equipment thread state set and the prediction equipment thread state set, and the relation among threads, equipment and thread synchronous creation states can be represented by utilizing the thread analysis model, so that the control of the subsequent thread synchronization is convenient.

S4, acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, and analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using the thread analysis model.

In the embodiment of the present invention, the real-time device linkage data refers to thread information corresponding to each device and each device in a device linkage process performed by multiple devices, where the method for extracting real-time linkage thread features from the real-time device linkage data is consistent with the method for generating a linkage thread feature set according to the linkage device set and the linkage thread set in the step S2, and is not described herein again.

In detail, the method for analyzing the real-time device thread state corresponding to the real-time linked thread feature by using the thread analysis model is consistent with the method for analyzing the predicted device thread state set corresponding to the linked thread feature set by using the preset thread time sequence model in the step S3, and will not be described herein.

In the embodiment of the invention, the real-time equipment linkage data is acquired, the real-time linkage thread characteristics are extracted from the real-time equipment linkage data, the thread analysis model is utilized to analyze the real-time equipment thread state corresponding to the real-time linkage thread characteristics, and the thread, equipment and thread synchronous creation state in the thread analysis model can be utilized to extract the real-time equipment thread state corresponding to the real-time equipment linkage data, so that the subsequent multi-thread synchronous creation is convenient, and the multi-thread synchronous efficiency is improved.

S5, extracting a thread synchronization mechanism from the real-time equipment linkage data, performing synchronization matching on the thread state of the real-time equipment according to the thread synchronization mechanism to obtain a real-time standard thread state, and performing multi-thread synchronization creation according to the real-time standard thread state.

In the embodiment of the invention, the thread synchronization mechanism refers to a mode for controlling multiple threads to access the shared resource when the multiple threads are executed concurrently, so as to avoid the problem of inconsistent competing conditions and data.

In the embodiment of the present invention, the step of synchronously matching the thread state of the real-time device according to the thread synchronization mechanism to obtain a real-time standard thread state includes:

Specifically, the Mutex Lock (Mutex Lock) is a synchronization mechanism for protecting critical sections from multiple threads accessing shared resources at the same time, thereby avoiding occurrence of race conditions, the condition variable (Condition Variable) is a synchronization mechanism for signaling between threads to achieve waiting and waking of threads, and the synchronization point is a specific location in a multi-thread or concurrent program, where threads need to wait or synchronize to ensure that specific conditions are satisfied, and then continue execution.

In detail, extracting the synchronization point from the real-time device thread state according to the mutual exclusion lock and the condition variable refers to extracting a waiting condition from the condition variable, setting the mutual exclusion lock, waiting for the waiting condition until the condition is met, releasing the mutual exclusion lock, and determining the synchronization point.

Specifically, performing primary sequencing on the thread state of the real-time device according to the synchronization point to obtain a real-time primary thread state refers to sequencing threads in the thread state of the real-time device according to a thread condition required by the synchronization point, performing deadlock sequencing on the real-time primary thread state, and obtaining a real-time standard thread state refers to performing deadlock detection on the real-time primary thread state and sequencing according to a detection result.

According to the embodiment of the invention, the thread synchronization mechanism is extracted from the real-time equipment linkage data, the thread state of the real-time equipment is synchronously matched according to the thread synchronization mechanism, the real-time standard thread state is obtained, the multi-thread synchronous creation is carried out according to the real-time standard thread state, the sequence and time of the multi-thread synchronous creation during equipment linkage can be determined according to the thread names and equipment names during the real-time multi-equipment linkage, the deadlock and the thread logic error are avoided, and the efficiency of the multi-thread synchronous creation is further improved.

The method comprises the steps of acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using a thread analysis model, extracting real-time equipment thread states corresponding to the real-time equipment linkage data by using threads, equipment and thread synchronization creation states in the thread analysis model, thereby facilitating subsequent multithread synchronization creation, improving multithread synchronization efficiency, synchronously matching the real-time equipment thread states according to a thread synchronization mechanism by extracting the thread synchronization mechanism from the real-time equipment linkage data to obtain real-time standard thread states, and carrying out multithread synchronization creation according to the real-time standard thread states, wherein the sequence and time of multithread synchronization creation during equipment linkage can be determined according to thread names and equipment names during real-time multi-equipment linkage, deadlock and thread logic errors are avoided, and multithread synchronization creation efficiency is improved. Therefore, the multithreading synchronization creation method based on equipment linkage can solve the problem of lower efficiency in multithreading synchronization.

Fig. 4 is a functional block diagram of a device linkage-based multithreading synchronization creation system according to an embodiment of the present invention.

The device linkage-based multithreading synchronization creation system 100 of the present invention may be installed in an electronic device. Depending on the functions implemented, the device linkage-based multithreading synchronization creation system 100 may include a synchronization carding module 101, a feature fusion module 102, a model training module 103, a model analysis module 104, and a synchronization matching module 105. The module of the invention, which may also be referred to as a unit, refers to a series of computer program segments, which are stored in the memory of the electronic device, capable of being executed by the processor of the electronic device and of performing a fixed function.

In the present embodiment, the functions concerning the respective modules/units are as follows:

the synchronous carding module 101 is configured to split and clean pre-acquired linkage data of a historical device to obtain a historical linkage data set, and perform thread synchronous carding on the historical linkage data set to obtain a linkage thread state set;

the feature fusion module 102 is configured to extract a linkage device group set, a linkage thread group set, and a device thread state set from the linkage thread state set, and generate a linkage thread feature set according to the linkage device group set and the linkage thread group set, where the generating a linkage thread feature set according to the linkage device group set and the linkage thread group set includes: selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups; sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features; extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method; and calculating the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature by using the following shared weight algorithm: Wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the feature total dimension of the thread name feature is equal to the feature total dimension of the thread name feature, the thread resource feature and the thread communication feature, < >>Is a hyperbolic tangent function, ">Refers to the +.>Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++>Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics; performing full-connection operation on the shared thread features and the target equipment features to obtain linkage thread features, and integrating all linkage thread features into a linkage thread feature set;

the model training module 103 is configured to analyze a predicted device thread state set corresponding to the linked thread feature set by using a preset thread time sequence model, and train the thread time sequence model into a thread analysis model by using the device thread state set and the predicted device thread state set;

the model analysis module 104 is configured to obtain real-time device linkage data, extract real-time linkage thread features from the real-time device linkage data, and analyze a real-time device thread state corresponding to the real-time linkage thread features by using the thread analysis model;

The synchronization matching module 105 is configured to extract a thread synchronization mechanism from the real-time device linkage data, perform synchronization matching on the real-time device thread state according to the thread synchronization mechanism, obtain a real-time standard thread state, and perform multi-thread synchronization creation according to the real-time standard thread state.

In detail, each module in the device linkage-based multithreading synchronization creation system 100 in the embodiment of the present invention adopts the same technical means as the device linkage-based multithreading synchronization creation method described in fig. 1 to 3, and can produce the same technical effects, which are not described herein.

In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus, system and method may be implemented in other manners. For example, the system embodiments described above are merely illustrative, e.g., the division of the modules is merely a logical function division, and other manners of division may be implemented in practice.

The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.

It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof.

The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.

The embodiment of the application can acquire and process the related data based on the artificial intelligence technology. Among these, artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a digital computer-controlled machine to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use knowledge to obtain optimal results.

Furthermore, it is evident that the word "comprising" does not exclude other elements or steps, and that the singular does not exclude a plurality. Multiple units or systems set forth in the system embodiments may also be implemented by one unit or system in software or hardware. The terms first, second, etc. are used to denote a name, but not any particular order.

Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for multi-threaded synchronization creation based on device linkage, the method comprising:

s1: splitting and cleaning the pre-acquired historical equipment linkage data to obtain a historical linkage data set, and carrying out thread synchronous carding on the historical linkage data set to obtain a linkage thread state set;

s2: extracting a linkage equipment set, a linkage thread set and an equipment thread state set from the linkage thread state set respectively, generating a linkage thread feature set according to the linkage equipment set and the linkage thread set, wherein the generating the linkage thread feature set according to the linkage equipment set and the linkage thread set comprises the following steps:

S21: selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups;

s22: sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features;

s23: extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method;

s24: and calculating the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature by using the following shared weight algorithm:wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the thread name feature, the thread resource feature and the thread communication feature are equal in feature total dimension,is a hyperbolic tangent function, ">Refers to the +.>Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++ >Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics;

s25: performing full-connection operation on the shared thread features and the target equipment features to obtain linkage thread features, and integrating all linkage thread features into a linkage thread feature set;

s3: analyzing a predicted equipment thread state set corresponding to the linkage thread feature set by using a preset thread time sequence model, and training the thread time sequence model into a thread analysis model by using the equipment thread state set and the predicted equipment thread state set;

s4: acquiring real-time equipment linkage data, extracting real-time linkage thread characteristics from the real-time equipment linkage data, and analyzing real-time equipment thread states corresponding to the real-time linkage thread characteristics by using the thread analysis model;

s5: and extracting a thread synchronization mechanism from the real-time equipment linkage data, performing synchronization matching on the thread state of the real-time equipment according to the thread synchronization mechanism to obtain a real-time standard thread state, and performing multi-thread synchronization creation according to the real-time standard thread state.

2. The method for creating the multithreading synchronization based on the equipment linkage according to claim 1, wherein the splitting and cleaning the pre-acquired historical equipment linkage data to obtain the historical linkage data set comprises the following steps:

3. The method for creating the multithreading synchronization based on the equipment linkage according to claim 1, wherein the step of performing thread synchronization carding on the historical linkage data set to obtain a linkage thread state group set comprises the following steps:

Extracting a target thread state group from the target equipment data group;

4. The device linkage-based multithreading synchronization creation method of claim 1, wherein the extracting the linkage device group set, the linkage thread group set, and the device thread state set from the linkage thread state group set, respectively, comprises:

5. The method for creating multithreading synchronization based on device linkage according to claim 1, wherein the method for extracting thread name features, thread resource features and thread communication features from the target linkage thread group by using feature aggregation comprises:

6. The method for creating multithreading synchronization based on equipment linkage according to claim 1, wherein the analyzing the predicted equipment thread state set corresponding to the linked thread feature set by using a preset thread timing model comprises:

and carrying out feature fusion on the long-term features of the threads and the short-term features of the threads by using the following long-short fusion formula to obtain the time sequence features of the threads:wherein (1)>Is +. >Characteristic value of time>Is the time sequence number->Is the weight factor preset by the long and short fusion formula,/->Means +.>Characteristic value of time>Is the number of hidden units of the jump layer of the thread timing model, < >>、/>Is the unit number->Is the +.>The characteristic weight of the moment in time,means +.>Characteristic value of time>Is the number of hidden units of the recursive level of the thread timing model,/for the thread timing model>Is the +.o in the thread short-term feature>Characteristic weight of time of day->Means +.>Characteristic value of time>Is a bias factor preset by the long and short fusion formula;

7. The method for creating multi-thread synchronization based on equipment linkage according to claim 6, wherein the extracting the long-term thread feature and the short-term thread feature from the target linkage thread feature by using a preset thread timing model respectively comprises:

8. The device linkage-based multithreading synchronization creation method of claim 1, wherein training the thread timing model into a thread analysis model using the set of device thread states and the set of predicted device thread states comprises:

calculating a state loss value of the thread timing model from the thread state feature set and the predicted state feature set using a state loss function as follows:wherein (1)>Means the state loss value, +.>Refers to the total number of features of the thread state feature set, and the total number of features of the thread state feature set is equal to the total number of features of the predicted state feature set ，/>Is a characteristic sequence number->Is the +.>Thread state feature->Is the +.>The number of predicted state characteristics is determined,as a covariance function;

9. The method for creating multithreading synchronization based on device linkage according to claim 1, wherein the step of synchronously matching the thread state of the real-time device according to the thread synchronization mechanism to obtain a real-time standard thread state comprises the steps of:

10. A device linkage-based multithreading synchronization creation system, the system comprising:

the feature fusion module is configured to extract a linkage device group set, a linkage thread group set and a device thread state set from the linkage thread state set, and generate a linkage thread feature set according to the linkage device group set and the linkage thread group set, where the generating a linkage thread feature set according to the linkage device group set and the linkage thread group set includes: selecting linkage equipment groups in the linkage equipment groups one by one as target linkage equipment groups, and selecting linkage thread groups corresponding to the target linkage equipment groups from the linkage thread groups to serve as target linkage thread groups; sequentially carrying out text vectorization and feature joint mapping operation on the target linkage equipment group to obtain target equipment features; extracting thread name characteristics, thread resource characteristics and thread communication characteristics from the target linkage thread group by using a characteristic aggregation method; and calculating the thread name feature, the thread resource feature and the shared thread feature corresponding to the thread communication feature by using the following shared weight algorithm: Wherein (1)>Means the linked thread feature, +.>Refers to the feature dimension sequence number, & lt>Means that the thread name feature, the thread resource feature and the thread communication feature are equal in feature total dimension,is a hyperbolic tangent function, ">Refers to the +.>Vitamin characteristics (I)>、/>、/>Is a weight matrix of the shared weight algorithm, < ->Is a transposed symbol->Is a feature dimension function, ++>Refers to the +.>Vitamin characteristics (I)>Refers to the>Dimensional characteristics; performing full-connection operation on the shared thread features and the target equipment features to obtain linkage thread features, and integrating all linkage thread features into a linkage thread feature set;