Resource adaptive scheduling method and device based on transaction characteristics

Abstract

This application relates to the field of artificial intelligence technology, and provides a resource adaptive scheduling method and apparatus based on transaction characteristics. The method includes: acquiring historical transaction characteristics of a transaction system within a historical period, wherein the historical transaction characteristics include: historical transaction quantity characteristics, periodic characteristics, and trend characteristics of various transaction types; constructing a transaction volume prediction model based on a recurrent neural network according to the historical transaction characteristics, and using the transaction volume prediction model to predict the transaction volume of the transaction system in a future time period, obtaining a predicted transaction volume sequence for the future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each transaction type in the future time period; and scheduling the computing resources of the transaction system according to the predicted transaction volume sequence. This solves the problem of high system response latency and increased customer waiting time in resource scheduling during peak business periods in existing technologies.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and more specifically, to a resource adaptive scheduling method, apparatus, computer-readable storage medium, and computer program product based on transaction characteristics. Background Technology

[0002] In the daily operations of banking systems, the management of business traffic and resource scheduling are crucial to ensuring the system's efficient and stable operation. Currently, banking systems typically use a fixed resource pool to process various business instructions, including complex batch transactions such as payroll disbursements and payment settlements. While this method simplifies resource management to some extent, it has significant limitations and problems in practical applications.

[0003] During peak business periods, especially when processing complex business instructions, the system needs extra time to process backlogged business due to the fixed allocation of resources, resulting in increased customer waiting time and a decline in user experience.

[0004] In summary, existing technologies suffer from high system response latency during peak business periods, leading to increased customer waiting times. Summary of the Invention

[0005] The main objective of this application is to provide a resource adaptive scheduling method, apparatus, computer-readable storage medium, and computer program product based on transaction characteristics, which addresses the problem that existing technologies suffer from high system response latency and increased customer waiting time during peak business periods in resource scheduling.

[0006] To achieve the above objectives, according to one aspect of this application, a resource adaptive scheduling method based on transaction characteristics is provided, comprising: acquiring historical transaction characteristics of a transaction system within a historical period, wherein the historical transaction characteristics include: historical transaction quantity characteristics, periodic characteristics, and trend characteristics of various transaction types; constructing a transaction volume prediction model based on a recurrent neural network according to the historical transaction characteristics, and using the transaction volume prediction model to predict the transaction volume of the transaction system in a future time period, thereby obtaining a predicted transaction volume sequence for the future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the transaction types in the future time period; and scheduling the computing resources of the transaction system according to the predicted transaction volume sequence.

[0007] Optionally, scheduling the computing resources of the trading system according to the predicted transaction volume sequence includes: determining the total predicted transaction volume at a target time point based on the predicted transaction volume sequence, wherein the total predicted transaction volume includes the predicted transaction volume of each of the transaction types; detecting whether the total predicted transaction volume is less than or equal to a transaction volume threshold, wherein the transaction volume threshold is determined based on the computing resources of the trading system; if the total predicted transaction volume is less than or equal to the transaction volume threshold, determining the proportion of the predicted transaction volume of each of the transaction types at the target time point to the total predicted transaction volume; and scheduling the computing resources of the trading system at the target time point based on the proportion.

[0008] Optionally, after determining whether the total predicted transaction volume is less than or equal to the transaction volume threshold, the method further includes: if the total predicted transaction volume is greater than the transaction volume threshold, elastically scaling up the platform service container according to the total predicted transaction volume to increase the computing resources of the transaction system.

[0009] Optionally, after the total predicted transaction volume is less than or equal to the transaction volume threshold, the method further includes: detecting whether the predicted transaction volume for each of the transaction types is greater than the transaction volume threshold corresponding to that transaction type; and if the predicted transaction volume for any transaction type is greater than the transaction volume threshold corresponding to that transaction type, increasing the proportion of computing resources used to process that transaction type according to the predicted transaction volume of that transaction type.

[0010] Optionally, based on a recurrent neural network, a trading volume prediction model is constructed according to the historical trading features, including: processing the historical trading features into a historical trading feature sequence according to time granularity; training the recurrent neural network with the historical trading feature sequence so that the recurrent neural network can identify the historical trading quantity features, the periodic features and the trend features of each of the trading types, thereby obtaining the trading volume prediction model.

[0011] Optionally, training the recurrent neural network using the historical transaction feature sequence includes: standardizing the historical transaction feature sequence to obtain a standardized historical transaction feature sequence; dividing the standardized historical transaction feature sequence to obtain a training window sequence and a prediction window sequence; and training the recurrent neural network by using the training window sequence as the input of the recurrent neural network and the prediction window sequence as the output of the recurrent neural network.

[0012] Optionally, during the training of the recurrent neural network using the historical transaction feature sequence, the method further includes: calculating the prediction accuracy of the recurrent neural network using a mean squared error or mean absolute error loss function.

[0013] According to another aspect of this application, a resource adaptive scheduling device based on transaction characteristics is provided, comprising: an acquisition unit, configured to acquire historical transaction characteristics of a transaction system within a historical period, wherein the historical transaction characteristics include: historical transaction quantity characteristics, periodic characteristics, and trend characteristics of various transaction types; a prediction unit, configured to construct a transaction volume prediction model based on the historical transaction characteristics using a recurrent neural network, and use the transaction volume prediction model to predict the transaction volume of the transaction system in a future time period, thereby obtaining a predicted transaction volume sequence for the future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the transaction types in the future time period; and a scheduling processing unit, configured to perform scheduling processing on the computing resources of the transaction system according to the predicted transaction volume sequence.

[0014] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform any of the described transaction-based adaptive resource scheduling methods.

[0015] According to another aspect of this application, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of any of the transaction-feature-based resource adaptive scheduling methods.

[0016] By applying the technical solution of this application, deep learning of historical transaction data using recurrent neural networks effectively predicts future fluctuations in transaction volume. By extracting historical transaction features, including transaction quantity, periodicity, and trend characteristics, a transaction volume prediction model is constructed, capable of accurately predicting the future transaction volume of various transaction types. The prediction results guide the dynamic adjustment of computing resources, especially for high-transaction-volume business types, enabling the pre-allocation of more resources and significantly improving processing efficiency and response speed. This addresses the problem of existing technologies exhibiting high system response latency during peak business periods, leading to increased customer waiting times. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1A hardware structure block diagram of a mobile terminal that performs a resource adaptive scheduling method based on transaction characteristics, according to an embodiment of this application, is shown.

[0019] Figure 2 A flowchart of a resource adaptive scheduling method based on transaction characteristics provided according to an embodiment of this application is shown;

[0020] Figure 3 A structural block diagram of a resource adaptive scheduling apparatus based on transaction characteristics provided according to an embodiment of this application is shown. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] As described in the background section, existing computing resource scheduling schemes suffer from inefficient resource allocation, high response latency, and poor scheduling flexibility when processing a large number of business instructions in banking systems, leading to increased customer waiting time. To address the issue of high system response latency causing increased customer waiting time during peak business periods, embodiments of this application provide a resource adaptive scheduling method, apparatus, computer-readable storage medium, and computer program product based on transaction characteristics.

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] The methods and embodiments provided in this application can be executed on a mobile terminal, a computer terminal, or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a resource adaptive scheduling method based on transaction characteristics, according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0027] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the resource adaptive scheduling method based on transaction characteristics in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0028] This embodiment provides a transaction-based adaptive resource scheduling method that runs on a mobile terminal, computer terminal, or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0029] Figure 2 This is a flowchart of a resource adaptive scheduling method based on transaction characteristics according to an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0030] Step S201: Obtain the historical trading characteristics of the trading system within the historical period, wherein the aforementioned historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types;

[0031] Step S202: Based on the recurrent neural network, a transaction volume prediction model is constructed according to the above-mentioned historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the above-mentioned transaction system in the future time period, so as to obtain the predicted transaction volume sequence of the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0032] Among them, the recurrent neural network is LSTM (Long Short-Term Memory). LSTM is an improved recurrent neural network (RNN) that effectively captures long-term dependencies in time series data by introducing memory units and gating mechanisms (input gate, forget gate, output gate). It can handle non-linear relationships: suitable for modeling complex time series patterns (such as periodic fluctuations and trend changes); it supports multivariate inputs: capable of simultaneously processing time series data and features such as transaction types; and it has dynamic prediction capabilities: by training on historical data, it predicts future transaction volumes and outputs resource allocation suggestions.

[0033] Step S203: Schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0034] In this embodiment, by applying steps S201, S202, and S203 above, a recurrent neural network is used to perform deep learning on historical transaction data, effectively predicting future fluctuations in transaction volume. By extracting historical transaction features, including transaction quantity, periodicity, and trend characteristics, a transaction volume prediction model is constructed, which can accurately predict the future transaction volume of various transaction types. The prediction results guide the dynamic adjustment of computing resources, especially for high-transaction-volume business types, enabling the allocation of more resources in advance, significantly improving processing efficiency and response speed. Therefore, existing technologies suffer from high system response latency during peak business periods, leading to increased customer waiting times.

[0035] Furthermore, by monitoring system performance metrics such as CPU utilization and processing latency, the PaaS container elastic scaling mechanism is triggered, ensuring that the system can automatically expand its capacity when resources are scarce, further optimizing resource utilization and overall system performance. Therefore, this embodiment achieves precise and efficient resource scheduling through dynamic resource allocation and intelligent monitoring and compensation mechanisms, significantly improving the operational efficiency and user experience of the transaction system. PaaS (Platform as a Service) is a cloud computing service model. In typical cloud computing layers, the Platform as a Service layer lies between Software as a Service (SaaS) and Infrastructure as a Service (IaaS). Its goal is to address the software environment upon which applications depend, including middleware, application frameworks, etc. It features lightweight containers and their packaging technology, allows for unified visual configuration and monitoring, and possesses elastic scaling capabilities.

[0036] In some embodiments, scheduling the computing resources of the trading system according to the predicted transaction volume sequence includes: determining the total predicted transaction volume at a target time point according to the predicted transaction volume sequence, wherein the total predicted transaction volume includes the predicted transaction volume of each of the aforementioned transaction types; detecting whether the total predicted transaction volume is less than or equal to a transaction volume threshold, wherein the transaction volume threshold is determined based on the computing resources of the trading system; if the total predicted transaction volume is less than or equal to the transaction volume threshold, determining the proportion of the predicted transaction volume of each of the aforementioned transaction types to the total predicted transaction volume at the target time point; and scheduling the computing resources of the trading system at the target time point according to the proportion.

[0037] In this embodiment, the computing resources of the trading system are scheduled based on the predicted transaction volume sequence. Specifically, this includes: First, threshold detection is performed on the total predicted transaction volume at the target time point and the subdivided predicted transaction volumes for each transaction type, based on the transaction volume prediction model. If the total predicted transaction volume is less than or equal to the transaction volume threshold, the proportion of the predicted transaction volume for each transaction type to the total predicted transaction volume is further calculated. Then, based on the calculated proportions, the allocation of computing resources in the trading system is dynamically adjusted at the target time point to achieve precise resource allocation for different types of transactions. This dynamic adjustment mechanism effectively improves resource utilization, reduces the average latency of transaction processing, and ensures the stable operation and response speed of the trading system.

[0038] In some embodiments, after determining whether the total predicted transaction volume is less than or equal to the transaction volume threshold, the method further includes: if the total predicted transaction volume is greater than the transaction volume threshold, elastically scaling up the platform service container according to the total predicted transaction volume to increase the computing resources of the transaction system.

[0039] In this embodiment, when the total predicted transaction volume exceeds a transaction volume threshold, the system will further dynamically and elastically scale the platform service container based on the prediction results, adjusting the computing resources of the transaction system in real time. By monitoring transaction volume in real time and using a transaction volume prediction model to predict future transaction volume, once the predicted total transaction volume exceeds a preset threshold, the PaaS container will be automatically expanded, thereby increasing the amount of computing resources that the system can handle. This design can effectively cope with the resource demands during peak transaction periods, avoid system bottlenecks, and improve transaction processing efficiency and response speed. In addition, the duration and frequency of PaaS container expansion will be dynamically adjusted according to changes in transaction volume, ensuring efficient utilization of system resources while maintaining stable system operation.

[0040] In some embodiments, after the total predicted transaction volume is less than or equal to the transaction volume threshold, the method further includes: detecting whether the predicted transaction volume for each of the above transaction types is greater than the transaction volume threshold corresponding to that transaction type; and if the predicted transaction volume for any transaction type is greater than the transaction volume threshold corresponding to that transaction type, increasing the proportion of the computing resources used to process that transaction type according to the predicted transaction volume of that transaction type.

[0041] In this embodiment, when the total predicted transaction volume is less than or equal to a predetermined transaction volume threshold, the system includes a step of individually detecting the predicted transaction volume for each transaction type. If the predicted transaction volume for any transaction type exceeds the specific threshold corresponding to that transaction type, the system will automatically adjust the proportion of computing resources used to process that type of transaction based on this predicted transaction volume. This ensures that even with a stable overall transaction volume, a surge in a specific transaction type can be responded to promptly. By dynamically adjusting resource allocation, the system's processing efficiency and response speed for specific transaction types are improved, avoiding processing delays or performance bottlenecks caused by uneven resource allocation. Simultaneously, by setting different transaction type thresholds, resources can be adjusted more precisely according to the real-time needs of different transaction types, thereby improving resource utilization and transaction processing flexibility while ensuring overall system stability. After any transaction type's predicted transaction volume triggers a threshold adjustment, the system will continuously monitor the adjustment effect to ensure the rationality and efficiency of resource allocation until the peak transaction period ends or the transaction volume returns to normal levels.

[0042] In some embodiments, a transaction volume prediction model is constructed based on a recurrent neural network according to the aforementioned historical transaction features, including: processing the aforementioned historical transaction features into a historical transaction feature sequence according to time granularity; training the aforementioned recurrent neural network with the aforementioned historical transaction feature sequence so that the recurrent neural network can identify the aforementioned historical transaction quantity features, the aforementioned periodic features and the aforementioned trend features of each of the aforementioned transaction types, thereby obtaining the aforementioned transaction volume prediction model.

[0043] In this embodiment, a transaction volume prediction model based on a recurrent neural network is constructed. Historical transaction features are processed into a historical transaction feature sequence at the time granularity. This sequence not only includes the specific transaction quantity but also incorporates key information such as transaction type and periodic event markers. Subsequently, the recurrent neural network is trained using the constructed historical transaction feature sequence, focusing on enabling the network to identify and learn the historical transaction quantity characteristics, periodic characteristics, and trend characteristics of transaction types. Through this training process, the resulting transaction volume prediction model can accurately predict the future transaction volume of each transaction type when receiving new transaction data, thereby guiding the dynamic allocation of system resources and improving resource utilization efficiency and system processing speed.

[0044] In some embodiments, training the recurrent neural network using the aforementioned historical transaction feature sequence includes: standardizing the aforementioned historical transaction feature sequence to obtain a standardized historical transaction feature sequence; dividing the standardized historical transaction feature sequence to obtain a training window sequence and a prediction window sequence; and training the recurrent neural network by using the training window sequence as the input of the recurrent neural network and the prediction window sequence as the output of the recurrent neural network.

[0045] The standardization process involves scaling the data range to a specific interval, typically [0, 1] or [-1, 1], using a feature scaling algorithm applied during the data preprocessing stage. This method ensures that each feature has the same scale, thus avoiding the disproportionate impact of features with large numerical values ​​on the results in the machine learning model.

[0046] In this embodiment, a recurrent neural network (RNN) is trained using historical transaction feature sequences. These sequences are then standardized to obtain a standardized historical transaction feature sequence, ensuring data comparability and predictive accuracy within the algorithm. The standardized historical transaction feature sequence is then divided into a training window sequence and a prediction window sequence. This division helps the model learn the correlation between the past and the future. The training window sequence is used as the input to the RNN, and the prediction window sequence is used as the output. By training the RNN in this way, the model can learn and predict future transaction volumes based on past transaction data, providing a predictive basis for resource scheduling and enabling dynamic adjustment and optimized allocation of resources. During implementation, standardization ensures the consistency of data input, window sequence division helps the model understand the dynamic characteristics of time series, and RNN training enables the model to capture and predict the fluctuation patterns of transaction volumes, thereby improving the accuracy and efficiency of resource scheduling.

[0047] In some embodiments, during the training of the recurrent neural network using the aforementioned historical transaction feature sequence, the method further includes: calculating the prediction accuracy of the recurrent neural network using a mean squared error or mean absolute error loss function.

[0048] In this embodiment, when training the recurrent neural network using historical transaction feature sequences, the mean squared error or mean absolute error loss function is further used to calculate the prediction accuracy of the recurrent neural network. This ensures that the model can accurately capture and predict changes in transaction volume in the time series, thereby achieving accurate estimation of resource demand. By optimizing the loss function, the model training process can more accurately adjust the weights, improving predictive ability. This embodiment improves the accuracy of resource scheduling, enabling timely adjustment of resource allocation when transaction volume peaks are predicted, avoiding system bottlenecks, and ensuring the efficiency of transaction processing and the continuity of user services.

[0049] The information collected is information and data authorized by the user or fully authorized by all parties. The collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse.

[0050] To enable those skilled in the art to better understand the technical solution of this application, the implementation process of the resource adaptive scheduling method based on transaction characteristics of this application will be described in detail below with reference to specific embodiments.

[0051] This embodiment relates to a specific resource adaptive scheduling method based on transaction characteristics, including the following steps:

[0052] Step S1: Obtain complex business instructions submitted by customers (such as bulk payment or payment settlement), generate a unique instruction ID based on the complex business instructions, set the status to "pending processing", and return a "submission successful" response;

[0053] Step S2: The background asynchronous callback resource scheduling module pre-allocates computing resources based on the instruction type and time characteristics of complex business instructions using the LSTM algorithm, queues up and waits for system resources to process specific business logic, and monitors the average processing time and CPU performance indicators of the system at the same time.

[0054] Step S3: Construct a trading volume prediction model using the LSTM algorithm, including: determining the time series characteristics of trading volume: organizing historical trading volume into a time series according to time granularity (such as hour, day, week), extracting the markers of periodic events such as payday and settlement day (such as whether it is a payday or the end-of-month settlement day), constructing time series feature input data, which can be combined with macroeconomic indicators (such as GDP, CPI) or corporate financial announcements as auxiliary features.

[0055] Step S4: Standardize the transaction volume, divide the time series into training and prediction windows, predict the transaction volume in the short term, and use the mean squared error (MSE) or mean absolute error (MAE) loss function to optimize the prediction accuracy of the prediction model. If the predicted transaction volume of a certain type of transaction is greater than a predetermined threshold, the proportion of computing resources for the corresponding transaction will be dynamically increased to improve processing efficiency. If the total predicted transaction volume is greater than a predetermined threshold, the PaaS container will be elastically scaled up to increase the total number of schedulable resources.

[0056] Step S5: Integrate the monitoring capabilities of the PaaS platform, collect system performance indicators (CPU utilization, memory usage, request latency, etc.) in real time, comprehensively monitor command response time and PaaS container CPU performance indicators, and if the average response time or performance indicator exceeds the threshold, issue an alarm and elastically scale up the PaaS container to increase the total number of schedulable resources.

[0057] Step S6: Based on the predicted transaction volume, schedule computing resources to process the transaction business: execute business logic (such as payroll verification, payment settlement routing), update the instruction status to "completed", and the customer can obtain the final status of the instruction through the query interface (such as RESTAPI).

[0058] This embodiment sets the total resource amount to be dynamically scaled by analyzing system performance indicators in real time, ensuring system availability and overall processing speed; and uses the LSTM algorithm to predict and adjust the system resource allocation rate, increasing system resource utilization and average system processing speed.

[0059] This application also provides a resource adaptive scheduling device based on transaction characteristics. It should be noted that this resource adaptive scheduling device based on transaction characteristics can be used to execute the resource adaptive scheduling method based on transaction characteristics provided in this application. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0060] The following describes the resource adaptive scheduling device based on transaction characteristics provided in the embodiments of this application.

[0061] Figure 3 This is a structural block diagram of a resource adaptive scheduling device based on transaction characteristics according to an embodiment of this application. Figure 3 As shown, the device includes:

[0062] The acquisition unit 31 is used to acquire the historical trading characteristics of the trading system within a historical period, wherein the aforementioned historical trading characteristics include: historical trading quantity characteristics, periodic characteristics, and trend characteristics of various trading types;

[0063] The prediction unit 32 is used to construct a transaction volume prediction model based on the above-mentioned historical transaction characteristics using a recurrent neural network, and to use the transaction volume prediction model to predict the transaction volume of the above-mentioned transaction system in a future time period, thereby obtaining a predicted transaction volume sequence for the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0064] The scheduling processing unit 33 is used to schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0065] In this embodiment, the acquisition unit is used to acquire historical transaction characteristics of the transaction system within a historical period. These historical transaction characteristics include: historical transaction quantity characteristics, periodic characteristics, and trend characteristics for various transaction types. The prediction unit is used to construct a transaction volume prediction model based on the historical transaction characteristics using a recurrent neural network, and then use this model to predict the transaction volume of the transaction system in a future time period, obtaining a predicted transaction volume sequence for the future time period. This predicted transaction volume sequence includes the predicted transaction volume for each transaction type in the future time period. The scheduling and processing unit is used to schedule and process the computing resources of the transaction system according to the predicted transaction volume sequence. By using a recurrent neural network to perform deep learning on historical transaction data, the fluctuations in future transaction volume are effectively predicted. By extracting historical transaction characteristics, including transaction quantity, periodicity, and trend characteristics, a transaction volume prediction model is constructed, which can accurately predict the future transaction volume of each transaction type. The prediction results guide the dynamic adjustment of computing resources, especially for high-transaction-volume business types, allowing for the pre-allocation of more resources, significantly improving processing efficiency and response speed. Therefore, existing technologies suffer from high system response latency during peak business periods, leading to increased customer waiting times.

[0066] As an optional scheme, the scheduling processing unit includes a first determining module, a first detecting module, a second determining module, and a scheduling processing module; the first determining module is used to determine the total predicted transaction volume at the target time point based on the predicted transaction volume sequence, wherein the total predicted transaction volume includes the predicted transaction volume of each of the above-mentioned transaction types; the first detecting module is used to detect whether the total predicted transaction volume is less than or equal to a transaction volume threshold, wherein the transaction volume threshold is determined based on the computing resources of the trading system; the second determining module is used to determine the proportion of the predicted transaction volume of each of the above-mentioned transaction types to the total predicted transaction volume at the target time point if the total predicted transaction volume is less than or equal to the transaction volume threshold; the scheduling processing module is used to schedule the computing resources of the trading system at the target time point based on the proportions.

[0067] In one optional scheme, the scheduling processing unit further includes a scaling module, which, after detecting whether the total predicted transaction volume is less than or equal to the transaction volume threshold, expands the platform service container elastically according to the total predicted transaction volume if the total predicted transaction volume is greater than the transaction volume threshold, so as to increase the computing resources of the transaction system.

[0068] In one optional scheme, the scheduling processing unit further includes a second detection module and a first processing module; the second detection module is used to detect whether the predicted transaction volume of each of the above transaction types is greater than the transaction volume threshold corresponding to that transaction type after the total predicted transaction volume is less than or equal to the above transaction volume threshold; the first processing module is used to increase the proportion of the computing resources used to process that transaction type according to the predicted transaction volume of that transaction type when the predicted transaction volume of any transaction type is greater than the above transaction volume threshold corresponding to that transaction type.

[0069] In one optional scheme, the prediction unit includes a second processing module and a training module; the second processing module is used to process the above-mentioned historical transaction features into a historical transaction feature sequence according to the time granularity; the training module is used to train the above-mentioned recurrent neural network using the above-mentioned historical transaction feature sequence, so that the recurrent neural network can identify the above-mentioned historical transaction quantity features, the above-mentioned periodic features and the above-mentioned trend features of each of the above-mentioned transaction types, and obtain the above-mentioned transaction volume prediction model.

[0070] In one optional scheme, the training module includes a standardization processing submodule, a partitioning processing submodule, and a training submodule. The standardization processing submodule is used to standardize the aforementioned historical transaction feature sequence to obtain a standardized historical transaction feature sequence. The partitioning processing submodule is used to partition the standardized historical transaction feature sequence to obtain a training window sequence and a prediction window sequence. The training submodule is used to train the recurrent neural network by using the training window sequence as the input and the prediction window sequence as the output of the recurrent neural network.

[0071] In one alternative approach, the training module includes a computation submodule, used to calculate the prediction accuracy of the recurrent neural network by employing a mean squared error or mean absolute error loss function during the training process of the recurrent neural network using the aforementioned historical transaction feature sequence.

[0072] The aforementioned resource adaptive scheduling device based on transaction characteristics includes a processor and a memory. All the aforementioned units are stored as program units in the memory, and the processor executes these program units to achieve the corresponding functions. All the aforementioned modules are located in the same processor; alternatively, the aforementioned modules may be located in different processors in any combination.

[0073] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters can address the issue of high system response latency and increased customer wait times caused by resource scheduling issues during peak business periods in existing technologies.

[0074] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0075] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to execute the transaction-based adaptive resource scheduling method.

[0076] Specifically, resource adaptive scheduling methods based on transaction characteristics include:

[0077] Step S201: Obtain the historical trading characteristics of the trading system within the historical period, wherein the aforementioned historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types;

[0078] Step S202: Based on the recurrent neural network, a transaction volume prediction model is constructed according to the above-mentioned historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the above-mentioned transaction system in the future time period, so as to obtain the predicted transaction volume sequence of the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0079] Step S203: Schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0080] This invention provides a processor for running a program, wherein the program executes the resource adaptive scheduling method based on transaction characteristics during runtime.

[0081] Specifically, resource adaptive scheduling methods based on transaction characteristics include:

[0082] Step S201: Obtain the historical trading characteristics of the trading system within the historical period, wherein the aforementioned historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types;

[0083] Step S202: Based on the recurrent neural network, a transaction volume prediction model is constructed according to the above-mentioned historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the above-mentioned transaction system in the future time period, so as to obtain the predicted transaction volume sequence of the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0084] Step S203: Schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0085] This invention provides a device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0086] Step S201: Obtain the historical trading characteristics of the trading system within the historical period, wherein the aforementioned historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types;

[0087] Step S202: Based on the recurrent neural network, a transaction volume prediction model is constructed according to the above-mentioned historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the above-mentioned transaction system in the future time period, so as to obtain the predicted transaction volume sequence of the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0088] Step S203: Schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0089] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0090] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps:

[0091] Step S201: Obtain the historical trading characteristics of the trading system within the historical period, wherein the aforementioned historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types;

[0092] Step S202: Based on the recurrent neural network, a transaction volume prediction model is constructed according to the above-mentioned historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the above-mentioned transaction system in the future time period, so as to obtain the predicted transaction volume sequence of the above-mentioned future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the above-mentioned transaction types in the above-mentioned future time period.

[0093] Step S203: Schedule the computing resources of the above-mentioned trading system according to the above-mentioned predicted trading volume sequence.

[0094] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0095] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0099] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0100] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0101] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0102] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0103] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0104] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A resource adaptive scheduling method based on transaction characteristics, characterized in that, include: Obtain historical trading characteristics of the trading system within a historical period, wherein the historical trading characteristics include: historical trading volume characteristics, periodic characteristics, and trend characteristics of various trading types; Based on a recurrent neural network, a transaction volume prediction model is constructed according to the historical transaction characteristics, and the transaction volume prediction model is used to predict the transaction volume of the transaction system in a future time period, thereby obtaining a predicted transaction volume sequence for the future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each transaction type in the future time period. The computing resources of the trading system are scheduled and processed based on the predicted trading volume sequence.

2. The method according to claim 1, characterized in that, include: Scheduling the computing resources of the trading system based on the predicted trading volume sequence includes: The total predicted transaction volume at the target time point is determined based on the predicted transaction volume sequence, wherein the total predicted transaction volume includes the predicted transaction volume for each of the transaction types. Detect whether the total predicted transaction volume is less than or equal to a transaction volume threshold, wherein the transaction volume threshold is determined based on the computing resources of the transaction system; If the total predicted transaction volume is less than or equal to the transaction volume threshold, determine the proportion of the predicted transaction volume of each transaction type at the target time point to the total predicted transaction volume; Based on the stated proportion, the computing resources of the transaction system are scheduled and processed at the target time point.

3. The method according to claim 2, characterized in that, After detecting whether the total predicted trading volume is less than or equal to the trading volume threshold, the method further includes: If the total predicted transaction volume exceeds the transaction volume threshold, the platform service container is elastically scaled up and down according to the total predicted transaction volume to increase the computing resources of the transaction system.

4. The method according to claim 2, characterized in that, After the total predicted trading volume is less than or equal to the trading volume threshold, the method further includes: Detect whether the predicted transaction volume for each of the transaction types is greater than the transaction volume threshold corresponding to that transaction type; If the predicted transaction volume for any transaction type is greater than the transaction volume threshold corresponding to that transaction type, the proportion of computing resources used to process that transaction type is increased according to the predicted transaction volume for that transaction type.

5. The method according to claim 1, characterized in that, A transaction volume prediction model is constructed based on a recurrent neural network and the historical transaction characteristics, including: The historical transaction features are processed into a historical transaction feature sequence according to time granularity; The recurrent neural network is trained using the historical transaction feature sequence to enable it to identify the historical transaction quantity features, periodic features, and trend features of each transaction type, thereby obtaining the transaction volume prediction model.

6. The method according to claim 5, characterized in that, include: Training the recurrent neural network using the historical transaction feature sequence includes: The historical transaction feature sequence is standardized to obtain the standardized historical transaction feature sequence. The standardized historical transaction feature sequence is divided and processed to obtain a training window sequence and a prediction window sequence; The recurrent neural network is trained by using the training window sequence as input and the prediction window sequence as output.

7. The method according to claim 5, characterized in that, In the process of training the recurrent neural network using the historical transaction feature sequence, the method further includes: The prediction accuracy of the recurrent neural network is calculated using the mean squared error or mean absolute error loss function.

8. A resource adaptive scheduling device based on transaction characteristics, characterized in that, include: The acquisition unit is used to acquire the historical trading characteristics of the trading system within a historical period, wherein the historical trading characteristics include: historical trading quantity characteristics, periodic characteristics, and trend characteristics of various trading types; The prediction unit is used to construct a transaction volume prediction model based on the historical transaction characteristics using a recurrent neural network, and to use the transaction volume prediction model to predict the transaction volume of the transaction system in a future time period, thereby obtaining a predicted transaction volume sequence for the future time period, wherein the predicted transaction volume sequence includes the predicted transaction volume of each of the transaction types in the future time period. The scheduling processing unit is used to schedule the computing resources of the trading system according to the predicted transaction volume sequence.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the resource adaptive scheduling method based on transaction characteristics as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the resource adaptive scheduling method based on transaction characteristics as described in any one of claims 1 to 7.

Images