Super-heterogeneous terminal intelligent engine cross-end cooperation implementation method

Abstract

The invention relates to the technical field of computer intelligent terminals, in particular to a super-heterogeneous terminal intelligent engine cross-terminal collaboration implementation method which specifically comprises the following steps: step 1, equipment identification and description: in a super-heterogeneous environment, identifying and describing terminal equipment participating in collaboration; task analysis and decomposition: when a user initiates a cross-end collaborative task, deeply analyzing and decomposing the task; step 3, resource evaluation and modeling: in a super-heterogeneous environment, evaluating the resource condition of each terminal device, and for each device, evaluating the computing resource, the storage resource and the network resource of the device; according to the method, the execution efficiency is improved through accurate task allocation and resource scheduling, allocation of computing resources and task scheduling are simplified, the development cost is reduced, heterogeneous barriers are broken, seamless collaboration is achieved, and the difficulty brought by cross-end collaboration is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of angle iron drawing equipment, and in particular to a method for realizing cross-terminal collaboration of a super heterogeneous terminal intelligent engine. Background Art

[0002] With the rapid development of information technology, a large number of various intelligent terminal devices such as smart phones, tablet computers, smart watches, smart home devices, and various edge computing devices have emerged. These devices show a high degree of heterogeneity in terms of hardware architecture, computing power, storage capacity, network connection ability, and the operating systems and application programs they run. The collaborative work between different terminal devices has become increasingly important. For example, users may hope to start a task on their mobile phone and then continue to complete it on a tablet computer, or use smart home devices to interact with mobile terminals to achieve a more convenient life experience.

[0003] Traditional terminal collaboration methods face many challenges in the face of such a super heterogeneous environment. First, the hardware differences between different devices make the allocation of computing resources and task scheduling complex. Different operating systems have different requirements for the interfaces and running environments of application programs, which increases the workload and cost of application developers. In addition, the heterogeneity of the network environment, such as different network bandwidths, latencies, and stabilities, also brings difficulties to cross-terminal collaboration. Summary of the Invention

[0004] In view of the technical problems raised in the background art, the present invention provides a method for realizing cross-terminal collaboration of a super heterogeneous terminal intelligent engine.

[0005] The technical solution adopted by the present invention is: a method for realizing cross-terminal collaboration of a super heterogeneous terminal intelligent engine, specifically including the following steps:

[0006] Step 1, device identification and description: In a super heterogeneous environment, identify and describe each terminal device participating in the collaboration:

[0007] Step 2, task analysis and decomposition: When a user initiates a cross-terminal collaboration task, deeply analyze and decompose the task;

[0008] Step 3, resource evaluation and modeling: In a super heterogeneous environment, evaluate the resource status of each terminal device, and for each device, evaluate its computing resources, storage resources, and network resources;

[0009] Step 4, task allocation strategy formulation: Based on the results of device identification, task analysis, and resource evaluation, formulate a task allocation strategy;

[0010] Step 5, data transmission and synchronization: During the cross-terminal collaboration process, data transmission and synchronization operations are required between different devices;

[0011] Step 6, Collaborative Execution and Monitoring: After the task assignment is completed and the data transmission and synchronization are ready, each device officially starts to collaboratively execute subtasks;

[0012] Step 7, Result Integration and Feedback: After all subtasks are executed, integrating the execution results on each device is to achieve cross-device collaboration.

[0013] In one embodiment, in Step 1, Device Identification and Description: In a hyperheterogeneous environment, the specific method for identifying and describing each terminal device participating in collaboration is as follows:

[0014] Each device is assigned a globally unique device identifier, denoted as D i , where i represents the i-th device;

[0015] Introduce the device feature vector F i =(f i1 , f i2 , …, f in ), n represents the number of device features, and f ij represents the j-th feature value of the i-th device, and f i1 represents the CPU model of the device; f i2 corresponds to the memory size of the device; f i3 represents the storage capacity of the device; f ij may also include the GPU model, screen resolution, and battery capacity of the device.

[0016] In one embodiment, in Step 2, Task Analysis and Decomposition: When a user initiates a cross-device collaboration task, the method for in-depth analysis and decomposition of the task is as follows;

[0017] The task is represented by a directed acyclic graph G=(V, E), where V is the set of vertices, and the vertex v k represents a subtask in the task, and E is the set of edges; the edge (v i , v j ) ∈ E indicates that there is a dependency relationship between subtasks v i and v j , that is, subtask v j can only be started and executed after subtask v i is completed.

[0018] In one embodiment, in Step 3, Resource Evaluation and Modeling: In a hyperheterogeneous environment, evaluate the resource status of each terminal device. For each device, the specific method for evaluating its computing resources, storage resources, and network resources is as follows;

[0019] Computing resources: Measured by the CPU performance metrics of the device. A common metric is the number of instructions executed per second;

[0020] Let device D i have a CPU performance of

[0021] Storage resources: represented by the available storage space size of the device to indicate,

[0022] Network resources: described by the current network bandwidth of the device and network latency to describe;

[0023] Construct a resource model: The resource model is presented in the form of a tuple to present.

[0024] In one embodiment, in step four, task allocation strategy formulation: Based on the results of device identification, task analysis, and resource evaluation, the specific method for formulating the task allocation strategy is as follows:

[0025] Introduce the task allocation function A(v k , D i ). When the value of this function is 1, it means that the subtask v k is allocated to device D i to execute. When the value is 0, it means it is not allocated;

[0026] The total task execution time T can be calculated by the following formula:

[0027]

[0028] where m represents the number of devices, covering all different types of terminal devices participating in the collaboration; n represents the number of subtasks, that is, the total number of subtasks obtained through task analysis and decomposition, and t(v k , D i ) represents the execution time of the subtask v k on device D i to execute.

[0029] In one embodiment, in step five, data transmission and synchronization: In the process of cross-terminal collaboration, the specific method for data transmission and synchronization operations between different devices is as follows:

[0030] Introduce a data transmission protocol. Let the data transmission volume be D size , the network bandwidth be B network , the transmission delay be L network , and the data transmission time T transfer can be calculated by the following formula:

[0031]

[0032] In one embodiment, in step six, collaborative execution and monitoring: when the task assignment is completed and the data transmission and synchronization are ready, the specific method for each device to officially start collaborative execution of subtasks is as follows:

[0033] For each subtask v k Define an execution status variable S(v k ), whose value is "not started", "executing", "completed";

[0034] By real-time monitoring the value of S(v k ), the system can understand the overall execution progress of the task.

[0035] In one embodiment, in step seven, result integration and feedback: when all subtasks are executed, the specific method to integrate the execution results on each device to achieve cross-terminal collaboration is as follows:

[0036] Let the result obtained by executing the subtask on device D i be The final integrated result R final can be obtained by merging each R resulti through a merging algorithm;

[0037] The specific merging algorithm depends on the type of task and the characteristics of the data;

[0038] Feed back the integrated result to the user, and optimize and improve the entire cross-terminal collaboration process according to the user's feedback information.

[0039] The beneficial effects of the present invention are: compared with the prior art, in the present invention, in step one, device identification and description, a unique ID is assigned to the device and a feature vector is constructed, laying a foundation for subsequent steps; in step two, the task is represented as a directed acyclic graph and decomposed to quantify the complexity of subtasks; in step three, a model is constructed by evaluating the computing, storage, and network resources of the device; in step four, a task assignment strategy is formulated based on the previous results and solved using an optimization algorithm; in step five, a protocol is introduced to ensure efficient and reliable data transmission and achieve synchronization; in step six, the execution status of the task and the device resources are monitored in real time; in step seven, the results are integrated and optimized according to the user's feedback. It solves the problem of improving the execution efficiency through precise task assignment and resource scheduling, simplifies the allocation of computing resources and task scheduling, reduces the development cost, breaks the heterogeneous barrier to achieve seamless collaboration, enhances the user experience, adapts to diverse network environments, improves the system scalability, and solves the difficulties brought by cross-terminal collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "front", "upper", "lower", "left", "right", "vertical", "horizontal", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention.

[0042] To solve the problems existing in the background technology, the present application proposes the following technical solutions: A method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine, specifically including the following steps:

[0043] Step 1, device identification and description: In a hyperheterogeneous environment, identify and describe each terminal device participating in the collaboration:

[0044] Step 2, task analysis and decomposition: When the user initiates a cross-terminal collaboration task, deeply analyze the task and decompose it;

[0045] Step 3, resource evaluation and modeling: In a hyperheterogeneous environment, evaluate the resource status of each terminal device. For each device, evaluate its computing resources, storage resources, and network resources;

[0046] Step 4, task allocation strategy formulation: Based on the results of device identification, task analysis, and resource evaluation, formulate a task allocation strategy;

[0047] Step 5, data transmission and synchronization: During the cross-terminal collaboration process, data transmission and synchronization operations are required between different devices;

[0048] Step 6, collaborative execution and monitoring: When the task allocation is completed and the data transmission and synchronization are ready, each device officially starts to collaboratively execute subtasks;

[0049] Step 7, result integration and feedback: When all subtasks are executed, integrating the execution results on each device is to achieve cross-terminal collaboration

[0050] The above technical solutions are explained in sequence as follows:

[0051] In Step 1, device identification and description: In a hyperheterogeneous environment, the specific method for identifying and describing each terminal device participating in the collaboration is as follows:

[0052] Each device is assigned a globally unique device identifier, denoted as D i , where i represents the i-th device;

[0053] This identifier is like the "ID card" of the device in the entire hyperheterogeneous network, ensuring that the system can quickly and accurately locate the specific device);

[0054] To comprehensively and deeply characterize the device features, a device feature vector F is introduced i =(f i1 , f i2 , …, f in ), where n represents the number of device features and is determined comprehensively based on various factors such as the device's hardware, software, and operating environment.

[0055] f ij represents the j-th feature value of the i-th device, and f i1 represents the CPU model of the device. CPUs of different models have significant differences in architecture, core count, main frequency, instruction set, etc. For example, Intel's Core series CPUs are very different from mobile phone chips based on the ARM architecture in terms of performance and power consumption management, and these differences directly affect the device's ability and efficiency in processing various tasks.

[0056] f i2 corresponds to the memory size of the device. As the temporary storage and processing space for data during device operation, the memory capacity determines the amount of data that the device can process simultaneously and the smoothness of multitasking parallel processing.

[0057] For example, when running large games or operating multiple applications simultaneously, devices with large memory can better avoid lagging.

[0058] f i3 represents the storage capacity of the device, which determines the device's ability to store data long-term. Whether it is a large number of multimedia files such as photos and videos, or complex system files and applications, all rely on sufficient storage capacity to achieve persistent storage. For example, devices used by professional photographers often require a large storage capacity to save high-resolution photo and video materials.

[0059] f ij may also include the GPU model of the device (which affects the graphics processing ability and is crucial for applications such as games and video editing), screen resolution (which is related to the display effect and affects the user's visual experience), and battery capacity (which determines the device's battery life and is particularly critical in mobile devices).

[0060] Constructing the device feature vector is not just a simple listing of the device's hardware parameters, but rather a quantification and integration of the device's various characteristics through a mathematical model, providing a solid data foundation for subsequent task allocation and resource scheduling. For example, when performing complex video editing tasks, the system can quickly screen out devices with high-performance CPUs, large memory, and powerful GPUs by analyzing the device feature vector and preferentially allocate relevant tasks to ensure efficient task execution.

[0061] In a further design, in step two, task analysis and decomposition: when a user initiates a cross-terminal collaboration task, the method of deeply analyzing and decomposing the task is as follows;

[0062] The task is represented by a directed acyclic graph G=(V, E), where V is the set of vertices, and vertex v k represents a subtask in the task, and E is the set of edges; edge (v i , v j ) ∈ E indicates that there is a dependency relationship between subtasks v i and v j , that is, subtask v j can only be started and executed after subtask v i is completed;

[0063] For example, in a video production task, after the subtask v1 of video material collection is completed, the subtask v2 of material editing can be carried out, which forms an edge pointing from v1 to v2.

[0064] To more precisely quantify the complexity of subtasks, a task complexity index C k is defined for each subtask v k . This index comprehensively considers various factors such as the computing resources required by the subtask, the amount of data processing, and the amount of data transmission. The specific calculation method can adopt the method of weighted summation, such as where represents the CPU computing resources required by subtask v k (which can be measured by the estimated CPU operation time or the number of instruction executions), represents the amount of data that the subtask needs to process (such as the number of bytes of the data), T transferk represents the amount of data transmission during the execution of the subtask (also calculated in bytes), and a, b, c are weight coefficients determined according to the actual task characteristics, and a + b + c = 1.

[0065] For example, for a subtask mainly focused on data processing, such as big data analysis, the weight of b may be relatively large; while for a subtask that needs to interact frequently with other devices, such as real-time video transmission, the weight of c will be higher.

[0066] By decomposing complex cross-terminal collaboration tasks into a series of interrelated subtasks and clarifying the task complexity index of each subtask, the system can, according to the characteristics of the device, allocate appropriate subtasks to the most matching device for execution, thereby improving the overall task execution efficiency. For example, for subtasks with high computational complexity, they can be allocated to devices with powerful computing performance; for subtasks with a large amount of data transmission, they are preferentially allocated to devices with superior network performance.

[0067] In Step 3, resource evaluation and modeling: In the hyperheterogeneous environment, evaluate the resource status of each terminal device. For each device, the specific methods for evaluating its computing resources, storage resources, and network resources are as follows;

[0068] Computing resources: Measured by the CPU performance metrics of the device. A common metric is the number of instructions executed per second;

[0069] Let device D i have a CPU performance of This metric reflects the number of instructions that the device can process per unit time and is an important basis for measuring the computing power of the device. For example, the CPU performance of a high-end server may reach billions of instructions per second, while that of an ordinary mobile phone is around hundreds of millions of instructions per second. In addition, factors such as the number of CPU cores and the main frequency can also be considered to make corrections to more accurately reflect the actual computing power of the device.

[0070] Storage resources: Represented by the available storage space size of the device This is directly related to the capacity of data that the device can store. With the continuous growth of data volume, the importance of storage resources has become increasingly prominent. For example, enterprise-level storage devices may have several petabytes of available storage space for storing massive amounts of business data; while the available storage space of a personal mobile phone may be between dozens of gigabytes and hundreds of gigabytes, mainly used to store user photos, videos, application programs, etc. data.

[0071] Network resources: Described by the current network bandwidth and network latency of the device. Network bandwidth determines the amount of data that the device can transmit per unit time. For example, the theoretical peak bandwidth of a 5G network can reach several gigabits per second, while the bandwidth of a home broadband network may be between dozens of megabits per second and hundreds of megabits per second. Network latency reflects the time required for data to travel from the sender to the receiver. Low latency is crucial for applications with high real-time requirements, such as online games and video conferencing. To better manage and schedule resources,

[0072] Construct a resource model: The resource model is presented in the form of a tuple . Through this resource model, the system can intuitively and comprehensively understand the resource status of each device, providing accurate and detailed basis for subsequent task allocation. For example, when performing a real-time video stream processing task, the system can, according to the resource model, preferentially select devices with high network bandwidth and low latency for video data transmission and reception, and at the same time allocate the video processing task to devices with sufficient computing resources to ensure the smoothness of video playback and the efficiency of processing.

[0073] In Step 4, task assignment strategy formulation: Based on the results of device identification, task analysis, and resource evaluation, the specific method for formulating the task assignment strategy is as follows:

[0074] The core goal of task assignment is to accurately assign subtasks to the most suitable devices on the premise of satisfying task dependencies, so as to minimize the total task execution time or maximize the overall system performance.

[0075] Introduce the task assignment function A(v k ,D i ). When the value of this function is 1, it means that the subtask v k is assigned to the device D i for execution. When the value is 0, it means it is not assigned;

[0076] The total task execution time T can be calculated by the following formula:

[0077]

[0078] Among them, m represents the number of devices, covering all different types of terminal devices participating in the collaboration; n represents the number of subtasks, that is, the total number of subtasks obtained through task analysis and decomposition. t(v k ,D i ) represents the execution time of the subtask v k on the device D i . Its calculation method is relatively complex and needs to comprehensively consider the task complexity index C k of the subtask and the computing resources i of the device D and other factors;

[0079] Among them, the startup time of the task on the device is related to the operating system characteristics of the device, the application loading mechanism, etc. For example, some devices may require a long initialization time when starting a large application. The data transmission time is closely related to the data transmission volume, network bandwidth, and network latency. As mentioned before, the data transmission time T transfer can be calculated by the formula . Here, the data transmission volume Ds i ze is the data volume that needs to be transmitted during the execution of the subtask v k ,D i on the device D k . i

[0080] The formulation of the task assignment strategy is essentially a typical optimization problem, that is, to find a set of optimal A(v k ,D i ​) value to minimize T. To solve this optimization problem, various advanced optimization algorithms can be used, such as genetic algorithms, particle swarm optimization algorithms, etc. Taking the genetic algorithm as an example, this algorithm simulates the genetic, mutation, and selection mechanisms in the process of biological evolution, encodes the task allocation scheme as a chromosome, and through continuous iteration, gradually optimizes the chromosome, that is, searches for a better task allocation scheme, and finally obtains the optimal solution that minimizes the total task execution time. In practical applications, through these optimization algorithms, in a complex heterogeneous environment, an approximately optimal task allocation strategy can be quickly and effectively found, improving the overall performance of the system.

[0081] In step five, data transmission and synchronization: In the process of cross-device collaboration, the specific methods for data transmission and synchronization operations between different devices are as follows:

[0082] However, due to the diverse network environments in which the devices are located and the different sizes of the data volumes, the efficiency and reliability of data transmission have become the key factors restricting the effect of cross-device collaboration.

[0083] To ensure the high efficiency of data transmission,

[0084] introduce a data transmission protocol. Let the data transmission volume be D size , the network bandwidth be B network , the transmission delay be L network , and the data transmission time be T transfer It can be calculated by the following formula:

[0085]

[0086] It can be seen from the formula that the larger the network bandwidth, the shorter the transmission time when the data transmission volume is certain; the lower the transmission delay, the shorter the data transmission time. For example, in a 5G network environment, compared with a 4G network, a higher network bandwidth can significantly shorten the transmission time of large files. To improve the reliability of data transmission, a data verification and retransmission mechanism is adopted. At the data sending end, a specific verification algorithm (such as the CRC cyclic redundancy check algorithm) is used to perform a verification calculation on the data to be transmitted to generate a verification code. At the receiving end, the same verification calculation is performed on the received data, and the verification code obtained from the calculation is compared with the verification code transmitted from the sending end. If the two are inconsistent, it means that an error may have occurred during data transmission, and the receiving end immediately requests the sending end to retransmit the data. This mechanism effectively guarantees the accuracy of data transmission and avoids task failures or result deviations caused by data errors. At the same time, to achieve data synchronization, a version control mechanism is introduced. Each data object is assigned a version number, and when the data is updated, the version number is automatically incremented. Devices determine whether data synchronization is required by comparing the version numbers of the data.

[0087] In Step 6, collaborative execution and monitoring: When the task assignment is completed and data transmission and synchronization are ready, the specific method for each device to officially start collaborative execution of subtasks is as follows:

[0088] During the execution process, it is crucial to monitor the execution status of the task in real time. This helps to promptly detect problems and take corresponding adjustment measures to ensure the smooth completion of the task.

[0089] For each subtask v k Define an execution status variable S(v k ), whose values are "not started", "executing", and "completed";

[0090] By monitoring the value of S(v k ) in real time, the system can understand the overall execution progress of the task. For example, in a cross - device collaborative development task of a large - scale project, by monitoring the execution status of each subtask, the project leader can intuitively see which modules have been developed, which are in progress, and which have not been started, thus reasonably arranging subsequent work. At the same time, closely monitor the resource usage of the device. Take the real - time monitoring of the CPU usage rate memory usage rate of the device as an example. The CPU usage rate reflects the busy degree of the device's CPU over a period of time. The memory usage rate reflects the occupancy of the device's memory. If it is found that the resource usage rate of a certain device is too high, such as the CPU usage rate continuously exceeding 80% and the memory usage rate approaching 100%, this may lead to a decline in device performance and thus affect the execution efficiency of the task. At this time, the system can dynamically adjust the task assignment strategy and migrate some tasks to other devices with more idle resources.

[0091] In Step 7, result integration and feedback: When all subtasks are executed, the specific method to integrate the execution results on each device to achieve cross - device collaboration is as follows:

[0092] The result integration process needs to fully consider many issues such as data format and consistency.

[0093] Let the result obtained from executing the subtask on device D i be The final integrated result R final can be obtained by merging each R resulti through a merging algorithm;

[0094] The specific merging algorithm depends on the type of task and the characteristics of the data;

[0095] Feed back the integrated result to the user and optimize and improve the entire cross - device collaboration process based on the user's feedback information.

[0096] For example, if the results of the subtasks are data segments and these segments have a specific sequential relationship, then the complete data can be obtained by concatenating them in sequence; if the results of the subtasks are different processing results of the same data, such as different devices performing filter processing of different styles on a picture, then a fusion algorithm may be required to fuse these results according to the user's preferences or preset rules.

[0097] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for realizing cross-terminal collaboration of a hyper heterogeneous terminal intelligent engine, characterized in that, Specifically, it includes the following steps: Step 1, Device Identification and Description: In a hyperheterogeneous environment, identify and describe each terminal device participating in the collaboration: Step 2, Task Analysis and Decomposition: When a user initiates a cross-terminal collaboration task, analyze and decompose the task; Step 3, Resource Evaluation and Modeling: In a hyperheterogeneous environment, evaluate the resource status of each terminal device. For each device, evaluate its computing resources, storage resources, and network resources; Step 4, Task Allocation Strategy Formulation: Based on the results of device identification, task analysis, and resource evaluation, formulate a task allocation strategy; Step 5, Data Transmission and Synchronization: During the cross-terminal collaboration process, data transmission and synchronization operations are required between different devices; Step 6, Collaborative Execution and Monitoring: When the task allocation is completed and the data transmission and synchronization are ready, each device officially starts to collaboratively execute subtasks; Step 7, Result Integration and Feedback: When all subtasks are executed, integrating the execution results on each device realizes cross-terminal collaboration.

2. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 1, Device Identification and Description: In a hyperheterogeneous environment, the specific method for identifying and describing each terminal device participating in the collaboration is as follows: Each device is assigned a globally unique device identifier, denoted as D i , where i represents the i-th device; Introduce the device feature vector F i =(f i1 , f i2 , …, f in ), where n represents the number of device features, and f ij represents the j-th feature value of the i-th device. f i1 represents the CPU model of the device; f i2 corresponds to the memory size of the device; f i3 represents the storage capacity of the device; f ij may also include the GPU model, screen resolution, and battery capacity of the device.

3. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 2, Task Analysis and Decomposition: When a user initiates a cross-terminal collaboration task, the method for deeply analyzing and decomposing the task is as follows; The task is represented by a directed acyclic graph G = (V, E), where V is the set of vertices, and vertex v k represents a subtask in the task, and E is the set of edges; the edge (v i , v j ) ∈ E indicates that there is a dependency relationship between subtasks v i and v j , that is, subtask v j can be started and executed only after subtask v i is completed.

4. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 3, Resource Evaluation and Modeling: In a hyperheterogeneous environment, the specific method for evaluating the resource status of each terminal device. For each device, evaluating its computing resources, storage resources, and network resources is as follows; Computing resources: Measured by the CPU performance index of the device, and a common index is the number of instructions executed per second; Let device D i have a CPU performance of Storage resources: represented by the available storage space size of the device to indicate Network resources: Described by the current network bandwidth of the device and network latency to describe; Construct a resource model: The resource model is presented in the form of a tuple ​ 5. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 4, Task Allocation Strategy Formulation: Based on the results of device identification, task analysis, and resource evaluation, the specific method for formulating a task allocation strategy is as follows: Introduce the task assignment function A(v k , D i ). When the value of this function is 1, it means that the subtask v k is assigned to device D i . When the value is 0, it means it is not assigned; The total task execution time T can be calculated by the following formula: ； Among them, m represents the number of devices, covering all different types of terminal devices participating in collaboration; n represents the number of subtasks, that is, the total number of subtasks obtained through task analysis and decomposition, and t(v k ,D i ) represents the execution time of subtask v k on device D i .

6. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 5, Data Transmission and Synchronization: During the cross-terminal collaboration process, the specific method for data transmission and synchronization operations required between different devices is as follows: Introduce a data transmission protocol, and set the data transmission volume as D size , the network bandwidth is B network , the transmission delay is L network , the data transmission time T transfer It can be calculated by the following formula: 。 7. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In Step 6, Collaborative Execution and Monitoring: When the task allocation is completed and the data transmission and synchronization are ready, the specific method for each device to officially start to collaboratively execute subtasks is as follows: For each subtask v k define an execution status variable S(v k ), whose values are "not started", "executing", "completed"; By monitoring the value of S(v k ) in real time, the system can understand the overall execution progress of the task.

8. The method for realizing cross-terminal collaboration of a hyperheterogeneous terminal intelligent engine according to claim 1, wherein In step 7, result integration and feedback: When all subtasks are executed, the specific method of achieving cross-device collaboration by integrating the execution results on each device is as follows: Set device D i The result obtained by executing the subtask on it is The final integration result R final Each R can be merged through a merging algorithm resulti to obtain by merging; The specific merging algorithm depends on the type of task and the characteristics of the data; Feed the integrated result back to the user, and optimize and improve the entire cross-device collaboration process based on the user's feedback information.

Images