CN114610475A - Training method of intelligent resource arrangement model - Google Patents

Abstract

The application is applicable to the technical field of computers, and provides a training method of an intelligent resource arrangement model, which comprises the following steps: the method comprises the steps that control equipment obtains first encryption information of a global resource arrangement model and sends the first encryption information to each computing cluster in a Federal learning alliance; receiving second encryption information of the updated local resource arrangement model returned by each computing cluster; and updating the global resource arrangement model according to all the second encryption information to obtain a new global resource arrangement model. According to the method, the distributed computing clusters are independently computed by adopting a federal learning method, the computing results are gathered to the control equipment, and a global resource arrangement model is obtained, and the global resource arrangement model can accurately and quickly determine the resource arrangement strategy, so that the capability of the edge computing clusters is fully exerted, the average completion time of tasks is minimized, the number of the completed tasks is maximized, and the resource utilization rate and the computing efficiency are improved.

Description

A training method for intelligent resource orchestration model

technical field

The present application belongs to the field of computer technology, and in particular relates to a training method of an intelligent resource arrangement model.

Background technique

With the rapid development of 5G and intelligent applications, the number of devices and the amount of data in the network has increased rapidly, and traditional centralized cloud computing cannot meet the needs of rapidly developing new applications. In order to provide low-latency, high-quality computing services and integrate the computing capabilities of the network edge, the existing technology will integrate the computing capabilities of the network edge, and reasonably allocate different computing clusters according to the performance and parameters of different computing clusters. resources, and determine a reasonable resource arrangement strategy.

However, traditional network resource management is mainly based on manual, optimization algorithms or heuristic methods in resource scheduling, which has high complexity and poor adaptive ability, making it difficult to meet the massive, dynamic and differentiated needs of edge computing users. Although some artificial intelligence-based resource orchestration can achieve better results, it requires global resource information, which is often difficult to achieve for decentralized and self-organizing edge computing clusters.

SUMMARY OF THE INVENTION

The embodiment of the present application provides a training method for an intelligent resource orchestration model, which can solve the problem that traditional network resource management is mainly based on manual, optimization algorithms or heuristic methods when performing resource orchestration, which has high complexity and poor self-adaptive ability, and is very difficult to achieve. It is difficult to meet the massive, dynamic and differentiated needs of edge computing users.

In a first aspect, an embodiment of the present application provides a training method for an intelligent resource orchestration model, which is applied to a control device, and the method includes:

acquiring first encrypted information of the global resource orchestration model, and sending the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update the local resource orchestration model of each of the computing clusters;

receiving the second encrypted information of the updated local resource orchestration model returned by each of the computing clusters;

The global resource arrangement model is updated according to all the second encrypted information to obtain a new global resource arrangement model.

Further, updating the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model, including:

Decrypt the second encrypted information according to a preset decryption rule to obtain update information corresponding to each of the computing clusters;

The global resource arrangement model is trained according to all the update information to obtain a new global resource arrangement model.

Further, the training method of the initial global resource orchestration model includes:

Obtain a first sample training set; the first sample training set includes historical task information and its corresponding resource scheduling strategy label;

The original resource arrangement model is trained according to the first sample training set to obtain an initial global resource arrangement model.

Further, the sample training set also includes preset special-case task information and its corresponding resource arrangement strategy label.

Further, after the global resource arrangement model is updated according to all the second encrypted information to obtain a new global resource arrangement model, the method further includes:

If the model information acquisition request sent by the computing cluster is received, the new global resource arrangement model and/or the third encrypted information corresponding to the new global resource arrangement model is sent to the computing cluster.

In a second aspect, an embodiment of the present application provides a training method for an intelligent resource orchestration model, which is applied to a computing cluster, and the method includes:

Obtain the first encrypted information sent by the control device;

updating the local resource orchestration model according to the first encrypted information to obtain the updated local resource orchestration model;

Sending second encrypted information of the updated local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller.

Further, the training method of the initial local resource orchestration model includes:

Obtain the original resource orchestration model and the second sample training set;

The initial intelligent resource arrangement model is trained according to the second sample training set to obtain an initial local resource arrangement model.

Further, updating the local resource orchestration model according to the first encrypted information to obtain the updated local resource orchestration model, including:

Iterative calculation is performed according to the gradient boosting tree algorithm and the first encrypted information, and a regression tree corresponding to the local resource arrangement model is constructed;

When the regression tree satisfies the preset condition, the current intelligent resource arrangement model is acquired as the updated local resource arrangement model.

In a third aspect, an embodiment of the present application provides a method for determining a resource arrangement strategy, including:

Get the task information of the task to be assigned;

Input the task information into a preset target intelligent resource arrangement model for processing, and obtain a resource arrangement strategy corresponding to the task information; wherein, the target intelligent resource arrangement model is arranged by the intelligent resource arrangement described in the first aspect. The training method of the model is obtained.

In a fourth aspect, an embodiment of the present application provides a control device, including:

a first processing unit, configured to obtain first encrypted information of the global resource orchestration model, and send the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update each of the computing The local resource orchestration model of the cluster;

a receiving unit, configured to receive the second encrypted information of the updated local resource orchestration model returned by each of the computing clusters;

A second processing unit, configured to update the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model.

Further, the second processing unit is specifically used for:

Decrypt the second encrypted information according to a preset decryption rule to obtain update information corresponding to each of the computing clusters;

The global resource arrangement model is trained according to all the update information to obtain a new global resource arrangement model.

Further, the control device also includes:

an obtaining unit, used for obtaining a first sample training set; the sample training set includes historical task information and its corresponding resource arrangement strategy label;

The third processing unit is configured to train the original resource arrangement model according to the first sample training set to obtain an initial global resource arrangement model.

Further, the sample training set also includes preset special-case task information and its corresponding resource arrangement strategy label.

Further, the control device also includes:

The fourth processing unit is configured to send the new global resource orchestration model and/or the third encrypted information corresponding to the new global resource orchestration model to the model information acquisition request sent by the computing cluster. the computing cluster.

In a fifth aspect, an embodiment of the present application provides a computing cluster, including:

a first obtaining unit, configured to obtain the first encrypted information sent by the control device;

an update unit, configured to update the local resource orchestration model according to the first encryption information to obtain an updated local resource orchestration model;

A sending unit, configured to send second encrypted information of the updated local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller.

Further, the computing cluster also includes:

The second acquisition unit is used to acquire the original resource arrangement model and the second sample training set;

A training unit, configured to train the initial intelligent resource arrangement model according to the second sample training set to obtain an initial local resource arrangement model.

Further, the update unit is specifically used for:

Iterative calculation is performed according to the gradient boosting tree algorithm and the first encrypted information, and a regression tree corresponding to the local resource arrangement model is constructed;

When the regression tree satisfies the preset condition, the current intelligent resource arrangement model is acquired as the updated local resource arrangement model.

In a sixth aspect, an embodiment of the present application provides an apparatus for determining a resource scheduling policy, including:

an acquisition unit for acquiring task information of the task to be assigned;

A processing unit, configured to input the task information into a preset target intelligent resource arrangement model for processing, and obtain a resource arrangement strategy corresponding to the task information; wherein, the target intelligent resource arrangement model is defined by the above-mentioned first aspect The training method of the intelligent resource orchestration model described above is obtained.

In a seventh aspect, an embodiment of the present application provides a control device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program The method for training an intelligent resource orchestration model as described in the first aspect above is implemented.

In an eighth aspect, an embodiment of the present application provides a computing cluster, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program The method for training an intelligent resource orchestration model according to the second aspect above is implemented.

In a ninth aspect, an embodiment of the present application provides a control device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, when the processor executes the computer program The method for determining a resource orchestration strategy as described in the third aspect above is implemented.

In a tenth aspect, an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, wherein, when the computer program is executed by a processor, the above-mentioned first aspect is implemented The training method of the intelligent resource orchestration model described above.

In an eleventh aspect, an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed by a processor, the above-mentioned second aspect is implemented The method for determining the resource arrangement strategy.

In a twelfth aspect, an embodiment of the present application provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed by a processor, the above third aspect is implemented The training method of the intelligent resource orchestration model.

In the embodiment of this application, the control device acquires the first encrypted information of the global resource orchestration model, and sends the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update each of the A local resource orchestration model of a computing cluster; receiving second encryption information of the updated local resource orchestration model returned by each of the computing clusters; updating the global resource orchestration model according to all the second encryption information to obtain a new The global resource orchestration model of . In the above method, when the number of users is large and the network conditions are complex, the federated learning method is used to independently calculate the scattered computing clusters, and the calculation results are aggregated to the control device to obtain a global resource arrangement model. The global resource arrangement model can accurately And quickly determine the resource scheduling strategy, improve resource utilization and computing efficiency.

On the other hand, the computing cluster obtains the first encrypted information sent by the controller; updates the local resource orchestration model according to the first encrypted information to obtain the updated local resource orchestration model; The second encrypted information is sent to the controller, and the second encrypted information is used to update the global resource orchestration model of the controller. In the above method, when the number of users is large and the network conditions are complex, the federated learning method is used to independently calculate the scattered computing clusters, and the calculation results are aggregated to the control device to obtain a global resource arrangement model. The global resource arrangement model can accurately And quickly determine the resource scheduling strategy, improve resource utilization and computing efficiency.

On the other hand, the device for determining the resource arrangement strategy obtains task information of the task to be assigned; inputs the task information into a preset target intelligent resource arrangement model for processing, and obtains the resource arrangement strategy corresponding to the task information. The target intelligent resource orchestration model used in the above method is obtained by means of federated learning, which independently calculates the distributed computing clusters, and aggregates the calculation results to the control device, which is then aggregated and calculated by the control device. The target intelligent resource scheduling model can accurately and quickly determine the resource scheduling strategy, which improves resource utilization and computing efficiency.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only for the present application. In some embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

1 is a schematic flowchart of a training method for an intelligent resource orchestration model provided by the first embodiment of the present application;

2 is a schematic flowchart of a training method for an intelligent resource orchestration model provided by a second embodiment of the present application;

3 is a schematic flowchart of a method for determining a resource scheduling strategy provided by a third embodiment of the present application;

4 is a schematic diagram of a control device provided by a fourth embodiment of the present application;

5 is a schematic diagram of a computing cluster provided by a fifth embodiment of the present application;

6 is a schematic diagram of an apparatus for determining a resource scheduling strategy provided by a sixth embodiment of the present application;

7 is a schematic diagram of a control device provided by a seventh embodiment of the present application;

8 is a schematic diagram of a computing cluster provided by an eighth embodiment of the present application;

FIG. 9 is a schematic diagram of a device for determining a resource scheduling policy provided by a ninth embodiment of the present application.

Detailed ways

In the following description, for the purpose of illustration rather than limitation, specific details such as a specific system structure and technology are set forth in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It is to be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described feature, integer, step, operation, element and/or component, but does not exclude one or more other The presence or addition of features, integers, steps, operations, elements, components and/or sets thereof.

It will also be understood that, as used in this specification and the appended claims, the term "and/or" refers to and including any and all possible combinations of one or more of the associated listed items.

As used in the specification of this application and the appended claims, the term "if" may be contextually interpreted as "when" or "once" or "in response to determining" or "in response to detecting ". Similarly, the phrases "if it is determined" or "if the [described condition or event] is detected" may be interpreted, depending on the context, to mean "once it is determined" or "in response to the determination" or "once the [described condition or event] is detected. ]" or "in response to detection of the [described condition or event]".

In addition, in the description of the specification of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description, and should not be construed as indicating or implying relative importance.

References in this specification to "one embodiment" or "some embodiments" and the like mean that a particular feature, structure or characteristic described in connection with the embodiment is included in one or more embodiments of the present application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in other embodiments," etc. in various places in this specification are not necessarily All refer to the same embodiment, but mean "one or more but not all embodiments" unless specifically emphasized otherwise. The terms "including", "including", "having" and their variants mean "including but not limited to" unless specifically emphasized otherwise.

With the vigorous development of 5G and intelligent technologies, many new delay-sensitive applications have emerged. For example, industrial control applications can remotely control equipment through the Internet; equipment monitoring applications can analyze equipment operating conditions through the Internet to achieve intelligent equipment management. ; Intelligent driving applications intelligently assist vehicles to make driving decisions by collecting a large amount of video, sensing and other data. Some remote industrial control applications require a delay of less than 20ms, intelligent driving scenarios require a delay of less than 10ms, and equipment security monitoring applications require a lower delay, the better.

Traditional centralized cloud computing cannot meet the needs of the above-mentioned new applications, because cloud computing centers usually have a certain distance from users and bring additional delays. For new intelligent scenarios, such as key industrial Internet applications, remote monitoring , intelligent driving, etc., there is a fatal impact, which may bring huge losses or safety hazards. Therefore, computing processing needs to be performed close to the user data source in order to provide users with a reliable experience.

Edge computing provides a new direction for solving the above problems. By deploying computing servers at the edge close to users, and providing information technology service environment and cloud computing capabilities at the edge of the network or close to users, communication delays can be greatly reduced, network performance optimized, and user experience improved.

However, in the actual application process, the existing edge computing technologies face the challenges of low resource utilization and computing efficiency: edge computing usually provides computing services for multiple users with different applications, and different applications have different communication resources and computing resource requirements. However, existing resource orchestration methods rely on user self-management or simple optimization algorithms to allocate computing units at a coarse granularity of physical machines or virtual machines, resulting in low resource utilization.

Therefore, in view of the above problems, the present application proposes a training method for an intelligent resource arrangement model, and a method for determining a resource arrangement strategy.

Please refer to FIG. 1. FIG. 1 is a schematic flowchart of a training method for an intelligent resource orchestration model provided by the first embodiment of the present application. The execution subject of the training method for an intelligent resource orchestration model in this embodiment is a control device. The training method of the intelligent resource orchestration model shown in Figure 1 may include:

S101: Acquire first encrypted information of a global resource orchestration model, and send the first encrypted information to each computing cluster in a federated learning alliance, where the first encrypted information is used to update the local resource orchestration of each of the computing clusters Model.

In this embodiment, the training of the intelligent resource orchestration model is based on federated learning. The adaptability of federated learning technology in ensuring information security and data privacy during data exchange enables it to show great advantages in carrying out efficient machine learning between distributed information units. The control device and the computing cluster together form a federated learning alliance. Each computing cluster can update the local model only based on local data without sharing global information. The local model information of the computing cluster is aggregated through the control device to obtain a new global resource arrangement. Model.

In the control device, the initial global resource arrangement model is stored, and the initial global resource arrangement model can be pre-trained by the control device, or can be transplanted to the control device after other devices have been trained.

Wherein, the control device can train the initial global resource arrangement model in the following manner: the control device obtains a first sample training set; the first sample training set includes historical task information and its corresponding resource arrangement strategy labels. The historical task information may include task-related parameters, such as the number of tasks, resources required for the task, time limit for task completion, task completion requirements, and the like. The historical task information and its corresponding resource scheduling strategy are labeled as the preferred resource scheduling strategy corresponding to the historical task information. The preferred resource scheduling strategy should be obtained after considering the resources, bandwidth, service quality and performance of each computing cluster. the optimal resource scheduling strategy. It is understandable that the computing clusters in the resource orchestration strategy should belong to the federated learning alliance mentioned above.

For the richness and authenticity of the samples, the greater the number of historical task information and its corresponding resource orchestration strategy labels included in the sample training set, the better.

After acquiring the first sample training set, the control device trains the original resource arrangement model according to the first sample training set to obtain an initial global resource arrangement model. During the training process, the input of the initial global resource orchestration model is historical task information, and the output of the initial global resource orchestration model is historical task information and its corresponding resource orchestration strategy.

Further, since the system runs in real time, it is possible to encounter some unexpected situations or situations that have not been encountered before, so some special cases can be set in the training samples, and the sample training set also includes presets. Special case task information and its corresponding resource orchestration policy label. Special case samples are used to simulate some network states that have not appeared in history for training, and obtain the optimal solution for this network state, and then when such a situation occurs in the future, it can respond adaptively and obtain optimized resources. Orchestration strategy.

After the control device obtains the initial global resource orchestration model, it will update the initial global resource orchestration model according to the information sent by the computing cluster, and after the update, send the encrypted information corresponding to the updated global resource orchestration model to each computing cluster . Each computing cluster updates its own local resource orchestration model according to the received first encryption information, and each computing cluster sends the encryption information corresponding to the updated local resource orchestration model to the control device, so as to control the device to continue to update the current global Resource Orchestration Model. In this way, multiple cycles of updating and learning are formed, and each time the control device obtains the encrypted information sent by the computing cluster, it updates the current global resource arrangement model, and finally obtains the final global resource arrangement model.

The encrypted information mentioned in this embodiment refers to an intermediate value in the construction process of the global resource arrangement model, which can be understood as a weight coefficient, and the construction of the global resource arrangement model can be determined by this intermediate value. Therefore, after the control device obtains the encrypted information of the computing cluster, it can adjust the current global resource arrangement model through the encrypted information of the computing cluster. After the computing cluster obtains the encrypted information of the control device, it can also use the encrypted information of the control device to Adjust the local resource orchestration model.

It should be noted that, because the federated learning method is adopted in this embodiment, information is not shared between each computing cluster and the control device, and the intermediate value mentioned above needs to be marked with a category in the calculation process Yes, they can be used for reconstruction to discover class information, so it is risky to send this intermediate value directly. To ensure security, each computing cluster and control device cannot directly send and access the intermediate value, and the control device first encrypts it before sending it to each computing cluster. When encrypting, the homomorphic encryption algorithm may be used to encrypt it to obtain the first encrypted information. For example, in homomorphic encryption, we define the expression of a number u in the additivity homomorphic encryption mode as <u>. According to the additivity principle of homomorphic encryption, for two arbitrary numbers u and v, we can There is <u>+<v>=<u+v>.

After receiving the first encrypted information, the computing cluster needs to use the intermediate value to update the local resource arrangement model without directly accessing the intermediate value.

In this embodiment, an update in the cyclic update is taken as an example to illustrate how the control device acquires the updated global resource arrangement model. The control device acquires the first encrypted information of the global resource orchestration model, and sends the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update the local resource orchestration model of each computing cluster.

S102: Receive second encrypted information of the updated local resource orchestration model returned by each of the computing clusters.

After the control device sends the first encrypted information to each computing cluster in the federated learning alliance, the computing cluster downloads the current global resource arrangement model from the control device as the local resource arrangement model, and the computing cluster performs the local resource arrangement model according to the first encrypted information. Update, obtain the updated local resource orchestration model, and obtain the second encrypted information in the update process. The computing cluster sends the second encrypted information to the control device, and the second encrypted information is used to update the current global resource arrangement model of the control device.

The end of the control device receives the updated second encrypted information of the local resource arrangement model returned by each computing cluster.

S103: Update the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model.

The control device updates the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model. In this embodiment, the method of using the second encrypted information to update the global resource arrangement model may be updated by a gradient boosting tree algorithm or by a neural network training method, which is not limited here.

In an implementation manner, a neural network training method can be used for updating to obtain a new global resource arrangement model. After acquiring the second encrypted information sent by the computing cluster, the control device decrypts the second encrypted information according to a preset decryption rule to obtain update information corresponding to each computing cluster.

The global resource arrangement model is trained according to all the update information to obtain a new global resource arrangement model. In this embodiment, the current global resource arrangement model is updated, and the current global resource arrangement model is trained by means of machine learning to obtain an updated global resource arrangement model. During the training process, the encrypted information can be used to adjust the training parameters.

In this training process, special case samples can be set to simulate some network states that have not appeared in history for training, and the optimal solution can be obtained according to this network state, and then when this situation occurs in the future, it can be adaptive React in a timely manner to obtain an optimized resource scheduling strategy.

After S103, it may further include: if a model information acquisition request sent by the computing cluster is received, sending the new global resource arrangement model and/or the third encrypted information corresponding to the new global resource arrangement model to the computing cluster. Here, after the control device obtains the new global resource orchestration model, it can obtain the intermediate value corresponding to the new global resource orchestration model, encrypt it, and obtain the third encrypted information. When the computing cluster sends the model information obtaining request, it will The new global resource orchestration model and/or the third encrypted information corresponding to the new global resource orchestration model is sent to the computing cluster.

In the embodiment of this application, the control device acquires the first encrypted information of the global resource orchestration model, and sends the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update each of the A local resource orchestration model of a computing cluster; receiving second encryption information of the updated local resource orchestration model returned by each of the computing clusters; updating the global resource orchestration model according to all the second encryption information to obtain a new The global resource orchestration model of . The above method, when the number of users is large and the network condition is complex, is realized by the method of federated learning, which independently calculates the distributed computing clusters, and aggregates the calculation results to the control device to obtain a global resource arrangement model. It can accurately and quickly determine the resource scheduling strategy, so that the capabilities of edge computing clusters can be fully utilized, providing high-quality computing support for delay-sensitive applications, dynamically and intelligently scheduling resources according to task characteristics, and minimizing the average completion time of tasks. Maximize the number of completed tasks, solve the problem of computing efficiency, and improve resource utilization and computing efficiency.

Please refer to FIG. 2. FIG. 2 is a schematic flowchart of a training method for an intelligent resource orchestration model provided by the second embodiment of the present application. The execution subject of the training method for an intelligent resource orchestration model in this embodiment is a computing cluster. The training method of the intelligent resource orchestration model shown in Figure 2 may include:

S201: Obtain the first encrypted information sent by the control device.

In this embodiment, the execution body is a computing cluster, and a computing cluster is a computer system, which is connected by a set of loosely integrated computer software and/or hardware to complete computing work in a highly close cooperation. In a sense, a computing cluster can be seen as a computer or a server. The computing cluster in this embodiment should belong to the federated learning alliance mentioned in the first embodiment.

The computing cluster can preset an initial local resource orchestration model. The initial local resource orchestration model can be pre-trained by the computing cluster, or it can be transplanted to the control device after other devices are trained.

Wherein, the training method of the initial local resource arrangement model includes: obtaining the original resource arrangement model and the second sample training set. Wherein, the second sample training set includes historical task information and its corresponding resource arrangement strategy label. Historical task information can include common feature data samples of all computing clusters in the federated learning alliance. Although the data of different computing clusters are independent of each other, due to some common features inherent in edge computing nodes, they can be included in the data of most clusters. Find some data samples that contain common features. These samples can be identified by their unique identifiers. At this time, the data across the computing cluster can be encrypted according to the existing privacy protection mechanism, and then the data samples containing the common features are selected.

Specifically, the common features are the basic attribute features of the computing cluster, such as CPU type, GPU type, memory size, storage space size, load situation, computing resource situation, etc.

The computing cluster trains the initial intelligent resource arrangement model according to the second sample training set to obtain an initial local resource arrangement model. During the training process, the input of the initial local resource scheduling model is historical task information, and the output of the initial local resource scheduling model is historical task information and its corresponding resource scheduling strategy.

Obtain the new global resource orchestration model from the control device as the local resource orchestration model. Wherein, each time the control device obtains a new global resource arrangement model, it can actively send the new global resource arrangement model to the computing cluster, or the computing cluster can send a request to the control device when the local resource arrangement model needs to be updated.

After acquiring the global resource arrangement model, the control device sends the first encrypted information corresponding to the global resource arrangement model to each computing cluster. The computing cluster acquires the first encrypted information sent by the control device, and the first encrypted information is used for the computing cluster to update the local resource arrangement model.

S202: Update the local resource arrangement model according to the first encrypted information to obtain an updated local resource arrangement model.

Wherein, the encrypted information mentioned in the first embodiment actually refers to an intermediate value in the construction process of the global resource arrangement model, which can be understood as a weight coefficient. When the first encrypted information is obtained by the control device through homomorphic encryption, after the computing cluster obtains the first encrypted information, the intermediate value is obtained by using the additivity principle of homomorphic encryption, and the local resource arrangement model is updated to obtain the updated information. Local resource orchestration model.

In one embodiment, a gradient boosting tree algorithm may be used to update the local resource arrangement model to obtain an updated local resource arrangement model. Iterative calculation is performed according to the gradient boosting tree algorithm and the first encrypted information, and a regression tree corresponding to the local resource arrangement model is constructed. When the regression tree satisfies the preset condition, the current intelligent resource orchestration model is acquired as the updated local resource orchestration model.

Specifically, the basic idea is as follows: Given a dataset X∈Rn ^×d with n samples and d-dimensional features, use the XGBoost algorithm to construct K regression trees to predict the output of the local resource orchestration model:

k个回归树的集成预测值。where, f _k ( _xi ): the output value of each of the k regression trees,

The ensemble predicted values of k regression trees.

To learn the regression tree model in the above formula, XGBoost adds a tree ft in _t iterations to minimize the following losses:

in,

回归树的预测损失函数，in,

The prediction loss function for the regression tree,

的一阶偏导与二阶偏导，g _i , h _i :

The first-order and second-order partial derivatives of ,

f _t : the modified regression tree used to reduce the calculation error and minimize the prediction loss,

φ ^(t) : the objective function of the t-th round of iteration, the iterative process is the process of minimizing φ ^(t) ,

γ: Split threshold.

When the model builds the regression tree at the t-th iteration, starting from the depth of 0, each time a split is added to a leaf node until the tree reaches the maximum depth. Further, the best split is determined using the following form:

In the above equation, _IL and IR are node samples of all _subtrees after classification. The classification that maximizes the split information is selected as the optimal split. When the model obtains an optimal tree structure, the optimal weight of leaf node j can be given by the following formula:

回归树编号为j的叶节点的最优权重值，I_j是叶子j的样本空间。in,

The optimal weight value of the leaf node of the regression tree number j, I _j is the sample space of leaf j.

Since the split candidate set and the optimal leaf node weight only depend on g _i and h _i , g _i and h _i are the intermediate values mentioned above, and the second encrypted information is obtained after homomorphic encryption.

S203: Send the second encrypted information of the updated local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller.

The computing cluster sends the updated second encrypted information of the local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller. The control device updates the global resource arrangement model according to the second encrypted information sent by all computer clusters to obtain a new global resource arrangement model.

In this embodiment, the computing cluster obtains the first encryption information sent by the controller; updates the local resource scheduling model according to the first encryption information to obtain the updated local resource scheduling model; arranges the updated local resources The second encrypted information of the model is sent to the controller, and the second encrypted information is used to update the global resource orchestration model of the controller. The above method, when the number of users is large and the network condition is complex, is realized by the method of federated learning, independent computing is performed on the scattered computing clusters, and the calculation results are aggregated to the control device to obtain a global resource orchestration model and a global resource orchestration model. It can accurately and quickly determine the resource scheduling strategy, so that the capabilities of the edge computing cluster can be fully utilized, providing high-quality computing support for delay-sensitive applications, dynamically and intelligently scheduling resources according to task characteristics, and minimizing the average completion time of tasks. Maximize the number of completed tasks, solve the problem of computing efficiency, and improve resource utilization and computing efficiency.

Please refer to FIG. 3 . FIG. 3 is a schematic flowchart of a method for determining a resource scheduling policy provided by a third embodiment of the present application. In this embodiment, a method for determining a resource arrangement strategy is executed by a device having a function of determining a resource arrangement strategy, such as a desktop computer, a server, and the like. The method for determining the resource orchestration strategy as shown in Figure 3 may include:

S301: Acquire task information of the task to be assigned.

The device obtains the task information of the task to be assigned, and the task information may include task-related parameters, such as the number of tasks, the resources required for the task, the completion time limit of the task, the task completion requirement, and so on.

S302: Input the task information into a preset target intelligent resource arrangement model for processing, and obtain a resource arrangement strategy corresponding to the task information; wherein, the target intelligent resource arrangement model is selected from any one of claims 1 to 5 The training method of the intelligent resource orchestration model described in item is obtained.

The target smart resource orchestration model is pre-stored in the device. For the acquisition method of the target smart resource orchestration model, reference may be made to the training methods of the smart resource orchestration model in the first and second embodiments, which will not be repeated here.

The task information is input into the preset target intelligent resource arrangement model for processing, and a resource arrangement strategy corresponding to the task information is obtained.

In this embodiment, the device for determining a resource arrangement strategy obtains task information of a task to be assigned; inputs the task information into a preset target intelligent resource arrangement model for processing, and obtains a resource arrangement strategy corresponding to the task information. The target intelligent resource orchestration model used in the above method is obtained by means of federated learning, which independently calculates the distributed computing clusters, and aggregates the calculation results to the control device, which is then aggregated and calculated by the control device. The target intelligent resource scheduling model can accurately and quickly determine the resource scheduling strategy, which improves resource utilization and computing efficiency.

Please refer to FIG. 4 , which is a schematic diagram of a control device provided by a fourth embodiment of the present application. The included units are used to execute the steps in the embodiment corresponding to FIG. 1 . For details, please refer to the relevant description in the embodiment corresponding to FIG. 1 . For convenience of explanation, only the parts related to this embodiment are shown. Referring to Figure 4, the control device 4 includes:

The first processing unit 410 is configured to obtain the first encrypted information of the global resource orchestration model, and send the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update each of the The local resource orchestration model of the computing cluster;

a receiving unit 420, configured to receive the second encrypted information of the updated local resource orchestration model returned by each of the computing clusters;

The second processing unit 430 is configured to update the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model.

Further, the second processing unit 430 is specifically used for:

Decrypt the second encrypted information according to a preset decryption rule to obtain update information corresponding to each of the computing clusters;

The global resource arrangement model is trained according to all the update information to obtain a new global resource arrangement model.

Further, the control device 4 also includes:

an obtaining unit, used for obtaining a first sample training set; the sample training set includes historical task information and its corresponding resource arrangement strategy label;

The third processing unit is configured to train the original resource arrangement model according to the first sample training set to obtain an initial global resource arrangement model.

Further, the sample training set also includes preset special-case task information and its corresponding resource arrangement strategy label.

Further, the control device 4 also includes:

The fourth processing unit is configured to send the new global resource orchestration model and/or the third encrypted information corresponding to the new global resource orchestration model to the model information acquisition request sent by the computing cluster. the computing cluster.

Please refer to FIG. 5 , which is a schematic diagram of a computing cluster provided by a fifth embodiment of the present application. The included units are used to execute the steps in the embodiment corresponding to FIG. 2 . For details, please refer to the relevant description in the embodiment corresponding to FIG. 2 . For convenience of explanation, only the parts related to this embodiment are shown. Referring to Figure 5, the computing cluster 5 includes:

a first obtaining unit 510, configured to obtain the first encrypted information sent by the control device;

an update unit 520, configured to update the local resource arrangement model according to the first encryption information, to obtain an updated local resource arrangement model;

The sending unit 530 is configured to send the second encrypted information of the updated local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller.

Further, the computing cluster 5 also includes:

The second acquisition unit is used to acquire the original resource arrangement model and the second sample training set;

A training unit, configured to train the initial intelligent resource arrangement model according to the second sample training set to obtain an initial local resource arrangement model.

Further, the updating unit 520 is specifically used for:

Iterative calculation is performed according to the gradient boosting tree algorithm and the first encrypted information, and a regression tree corresponding to the local resource arrangement model is constructed;

When the regression tree satisfies the preset condition, the current intelligent resource arrangement model is acquired as the updated local resource arrangement model.

Please refer to FIG. 6 , which is a schematic diagram of an apparatus for determining a resource scheduling policy provided by a sixth embodiment of the present application. The included units are used to execute the steps in the embodiment corresponding to FIG. 3 . For details, please refer to the relevant description in the embodiment corresponding to FIG. 3 . For convenience of explanation, only the parts related to this embodiment are shown. Referring to FIG. 6 , 6 of the device for determining a resource scheduling policy includes:

an obtaining unit 610, configured to obtain task information of the task to be assigned;

A processing unit 620, configured to input the task information into a preset target intelligent resource arrangement model for processing, and obtain a resource arrangement policy corresponding to the task information; wherein, the target intelligent resource arrangement model is determined by the above-mentioned first aspect The training method of the intelligent resource arrangement model is obtained.

FIG. 7 is a schematic diagram of a control device provided by a seventh embodiment of the present application. As shown in FIG. 7 , the control device 7 of this embodiment includes: a processor 70 , a memory 71 , and a computer program 72 stored in the memory 71 and executable on the processor 70 , for example, an intelligent resource orchestration model. training program. When the processor 70 executes the computer program 72, the steps in each of the above-mentioned embodiments of the training method for the smart resource orchestration model are implemented, for example, steps 101 to 103 shown in FIG. 1 . Alternatively, when the processor 70 executes the computer program 72, the functions of the modules/units in the foregoing apparatus embodiments, for example, the functions of the modules 410 to 430 shown in FIG. 4 are implemented.

Exemplarily, the computer program 72 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 71 and executed by the processor 70 to complete the this application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 72 in the control device 7 . For example, the computer program 72 can be divided into a first processing unit, a receiving unit, and a second processing unit, and the specific functions of each unit are as follows:

a first processing unit, configured to obtain first encrypted information of the global resource orchestration model, and send the first encrypted information to each computing cluster in the federated learning alliance, where the first encrypted information is used to update each of the computing The local resource orchestration model of the cluster;

a receiving unit, configured to receive the second encrypted information of the updated local resource orchestration model returned by each of the computing clusters;

A second processing unit, configured to update the global resource arrangement model according to all the second encrypted information to obtain a new global resource arrangement model.

The control device may include, but is not limited to, a processor 70 and a memory 71 . Those skilled in the art can understand that FIG. 7 is only an example of the control device 7, and does not constitute a limitation on the control device 7, and may include more or less components than the one shown, or combine some components, or different components For example, the control device may further include an input and output device, a network access device, a bus, and the like.

The so-called processor 70 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 71 may be an internal storage unit of the control device 7 , such as a hard disk or a memory of the control device 7 . The memory 71 may also be an external storage device of the control device 7, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) equipped on the control device 7. card, flash card (Flash Card) and so on. Further, the control device 7 may also include both an internal storage unit of the control device 7 and an external storage device. The memory 71 is used to store the computer program and other programs and data required by the control device. The memory 71 may also be used to temporarily store data that has been output or will be output.

FIG. 8 is a schematic diagram of a computing cluster provided by an eighth embodiment of the present application. As shown in FIG. 8 , the computing cluster 8 of this embodiment includes: a processor 80 , a memory 81 , and a computer program 82 stored in the memory 81 and executable on the processor 80 , such as an intelligent resource orchestration model training program. When the processor 80 executes the computer program 82 , the steps in each of the above-mentioned embodiments of the training method for the smart resource arrangement model are implemented, for example, steps 201 to 203 shown in FIG. 2 . Alternatively, when the processor 80 executes the computer program 82 , the functions of the modules/units in the above-mentioned device embodiments, for example, the functions of the modules 510 to 530 shown in FIG. 5 , are implemented.

Exemplarily, the computer program 82 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 81 and executed by the processor 80 to complete the this application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 82 in the computing cluster 8 . For example, the computer program 82 can be divided into a first acquisition unit, an update unit, a second processing unit, and a transmission unit, and the specific functions of each unit are as follows:

a first obtaining unit, configured to obtain the first encrypted information sent by the control device;

an update unit, configured to update the local resource orchestration model according to the first encryption information to obtain an updated local resource orchestration model;

A sending unit, configured to send second encrypted information of the updated local resource orchestration model to the controller, where the second encrypted information is used to update the global resource orchestration model of the controller.

The computing cluster may include, but is not limited to, the processor 80 and the memory 81 . Those skilled in the art can understand that FIG. 8 is only an example of the computing cluster 8 , and does not constitute a limitation on the computing cluster 8 , which may include more or less components than those shown in the figure, or combine some components, or different components For example, the computing cluster may further include input and output devices, network access devices, buses, and the like.

The so-called processor 80 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 81 may be an internal storage unit of the computing cluster 8 , such as a hard disk or a memory of the computing cluster 8 . The memory 81 may also be an external storage device of the computing cluster 8, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) equipped on the computing cluster 8. card, flash card (Flash Card) and so on. Further, the computing cluster 8 may also include both an internal storage unit of the computing cluster 8 and an external storage device. The memory 81 is used to store the computer program and other programs and data required by the computing cluster. The memory 81 can also be used to temporarily store data that has been output or will be output.

FIG. 9 is a schematic diagram of a device for determining a resource scheduling policy provided by a ninth embodiment of the present application. As shown in FIG. 9 , the device 9 for determining a resource scheduling policy in this embodiment includes: a processor 90 , a memory 91 , and a computer program 92 stored in the memory 91 and executable on the processor 90 , such as a resource Determining procedures for orchestration policies. When the processor 90 executes the computer program 92, the steps in each of the above-mentioned embodiments of the method for determining a resource scheduling policy are implemented, for example, steps 301 to 302 shown in FIG. 3 . Alternatively, when the processor 90 executes the computer program 92, the functions of the modules/units in each of the foregoing apparatus embodiments, such as the functions of the modules 610 to 620 shown in FIG. 6 , are implemented.

Exemplarily, the computer program 92 may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 91 and executed by the processor 90 to complete the this application. The one or more modules/units may be a series of computer program instruction segments capable of accomplishing specific functions, and the instruction segments are used to describe the execution process of the computer program 92 in the resource arrangement policy determining device 9 . For example, the computer program 92 can be divided into a receiving unit, a first processing unit, a second processing unit, a third processing unit, a fourth processing unit, and an execution unit, and the specific functions of each unit are as follows:

a receiving unit, configured to receive a timing request, where the timing request includes a timing duration and a processing task;

a first processing unit, configured to obtain a first increment time of a monotonically incrementing clock, and determine a first virtual alarm clock time according to the first increment time and the timing duration;

a second processing unit, configured to create a virtual timer according to the first virtual alarm clock time and the processing task, and determine a system alarm clock task according to the first virtual alarm clock time;

a third processing unit, configured to acquire the second increment time of the monotonically incrementing clock when the system clock triggers the system alarm clock task;

a fourth processing unit, configured to obtain the virtual alarm clock time corresponding to each of the virtual timers, and determine the virtual timer whose virtual alarm clock time is less than the second increment time as the target timer;

An execution unit, configured to execute the processing task corresponding to the target timer.

The device for determining the resource scheduling policy may include, but is not limited to, the processor 90 and the memory 91 . Those skilled in the art can understand that FIG. 9 is only an example of the device 9 for determining the resource orchestration strategy, and does not constitute a limitation on the device 9 for determining the resource orchestration strategy, and may include more or less components than those shown in the figure, or combinations thereof Some components, or different components, for example, the device for determining the resource scheduling policy may also include input and output devices, network access devices, buses, and the like.

The so-called processor 90 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 91 may be an internal storage unit of the device 9 for determining the resource arrangement strategy, such as a hard disk or memory of the device 9 for determining the resource arrangement strategy. The memory 91 may also be an external storage device of the device 9 for determining the resource scheduling strategy, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC) equipped on the device 9 for determining the resource scheduling strategy, Secure Digital (SD) card, flash card (Flash Card), etc. Further, the resource scheduling policy determining device 9 may further include both an internal storage unit of the resource scheduling policy determining device 9 and an external storage device. The memory 91 is used to store the computer program and other programs and data required by the device for determining the resource arrangement policy. The memory 91 can also be used to temporarily store data that has been output or will be output.

It should be noted that the information exchange, execution process and other contents between the above-mentioned devices/units are based on the same concept as the method embodiments of the present application. For specific functions and technical effects, please refer to the method embodiments section. It is not repeated here.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit, and the above-mentioned integrated units may adopt hardware. It can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present application. For the specific working processes of the units and modules in the above-mentioned system, reference may be made to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

Embodiments of the present application further provide a timing device for a virtual timer, where the timing device for a virtual timer includes: at least one processor, a memory, and a memory device that is stored in the memory and can run on the at least one processor. A computer program, when the processor executes the computer program, the steps in any of the foregoing method embodiments are implemented.

Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the steps in the foregoing method embodiments can be implemented.

The embodiments of the present application provide a computer program product, when the computer program product runs on a mobile terminal, the steps in the foregoing method embodiments can be implemented when the mobile terminal executes the computer program product.

The integrated unit, if implemented in the form of a software functional unit and sold or used as an independent product, may be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can be implemented by a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium, and the computer program When executed by the processor, the steps of the above-mentioned various method embodiments may be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, and the like. The computer-readable medium may include at least: any entity or device capable of carrying the computer program code to the photographing device/terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electrical carrier signal, telecommunication signal, and software distribution medium. For example, U disk, mobile hard disk, disk or CD, etc. In some jurisdictions, under legislation and patent practice, computer readable media may not be electrical carrier signals and telecommunications signals.

In the foregoing embodiments, the description of each embodiment has its own emphasis. For parts that are not described or described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of this application.

In the embodiments provided in this application, it should be understood that the disclosed apparatus/device and method may be implemented in other manners. For example, the apparatus/equipment embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or Components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

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

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the above-mentioned embodiments, those of ordinary skill in the art should understand that: it can still be used for the above-mentioned implementations. The technical solutions described in the examples are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions in the embodiments of the application, and should be included in the within the scope of protection of this application.

Claims (10)

1. A training method of an intelligent resource arrangement model is applied to a control device, and comprises the following steps:

acquiring first encryption information of a global resource arrangement model, and sending the first encryption information to each computing cluster in a Federal learning alliance, wherein the first encryption information is used for updating a local resource arrangement model of each computing cluster;

receiving second encryption information of the updated local resource orchestration model returned by each computing cluster;

and updating the global resource arrangement model according to all the second encryption information to obtain a new global resource arrangement model.

2. The method for training an intelligent resource arrangement model according to claim 1, wherein the updating the global resource arrangement model according to all the second encryption information to obtain a new global resource arrangement model comprises:

decrypting the second encrypted information according to a preset decryption rule to obtain update information corresponding to each computing cluster;

and training the global resource arrangement model according to all the updating information to obtain a new global resource arrangement model.

3. The method of training an intelligent resource orchestration model according to claim 1, wherein the method of training the initial global resource orchestration model comprises:

acquiring a first sample training set; the first sample training set comprises historical task information and corresponding resource arrangement strategy labels;

and training the original resource arrangement model according to the first sample training set to obtain an initial global resource arrangement model.

4. The training method of an intelligent resource orchestration model according to claim 3, wherein the sample training set further comprises preset special case task information and a resource orchestration strategy label corresponding thereto.

5. The method for training an intelligent resource arrangement model according to claim 1, wherein after the global resource arrangement model is updated according to all the second encryption information to obtain a new global resource arrangement model, the method further comprises:

and if a model information acquisition request sent by the computing cluster is received, sending the new global resource arrangement model and/or third encryption information corresponding to the new global resource arrangement model to the computing cluster.

6. A training method of an intelligent resource arrangement model is applied to a computing cluster, and comprises the following steps:

acquiring first encryption information sent by a controller;

updating the local resource arrangement model according to the first encryption information to obtain an updated local resource arrangement model;

and sending second encryption information of the updated local resource arrangement model to the controller, wherein the second encryption information is used for updating the global resource arrangement model of the controller.

7. The method of training an intelligent resource orchestration model according to claim 6, wherein the method of training the initial local resource orchestration model comprises:

acquiring an original resource arrangement model and a second sample training set;

and training the initial intelligent resource arrangement model according to the second sample training set to obtain an initial local resource arrangement model.

8. The method for training the intelligent resource arrangement model according to claim 6, wherein the updating the local resource arrangement model according to the first encryption information to obtain the updated local resource arrangement model includes:

performing iterative computation according to a gradient lifting tree algorithm and the first encryption information, and constructing a regression tree corresponding to the local resource arrangement model;

and when the regression tree meets the preset condition, acquiring the current intelligent resource arrangement model as the updated local resource arrangement model.

9. A method for determining a resource orchestration policy, comprising:

acquiring task information of a task to be distributed;

inputting the task information into a preset target intelligent resource arrangement model for processing to obtain a resource arrangement strategy corresponding to the task information; wherein the target intelligent resource arrangement model is obtained by the training method of the intelligent resource arrangement model of any one of claims 1 to 5.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 5, 6 to 8, or 9.

Images