Multi-application scene-oriented window layout cross-screen collaborative nonlinear reduction system

Abstract

The invention provides a multi-application scene-oriented window layout cross-screen collaborative nonlinear restoration system, which belongs to the technical field of computers and comprises a system data acquisition module, an application scene generation module, an intelligent layout strategy module and a scene layout restoration module. The system data acquisition module is responsible for acquiring application-related data; the application scene generation module is in communication connection with the system data acquisition module and the intelligent layout strategy module to generate a high-relevance application combination scene; the intelligent layout strategy module is in communication connection with the system data acquisition module, the application scene generation module and the scene layout restoration module to generate multiple sets of window layout snapshots; and the scene layout restoration module is in communication connection with the system data acquisition module and the intelligent layout strategy module and is responsible for scene layout restoration. According to the method, a software combination semantic network is established, and multi-screen application automatic layout based on application data flow and user interaction is realized.

Description

TECHNICAL FIELD

[0001] The application relates to multi-application-scene window layout restoration, in particular to a multi-application-scene window layout cross-screen cooperative nonlinear restoration system, and belongs to the technical field of computers. BACKGROUND

[0002] With the progress of hardware design and manufacturing processes such as computers, graphics cards and display devices, a graphics card can support multiple display output devices, and the demand of users for multi-screen display scenes of multiple display devices has gradually increased. Rich desktop applications provide diversified functions for different users, and multi-application collaborative work has become the norm.

[0003] At present, the multi-screen display setting generally only records the physical display parameters, such as resolution, arrangement position between different displays, etc., and ignores the correlation between software ecosystems, for example, the user opens two software, Photoshop and Lightroom, which are usually combined in the fields of photography and design, and can be recognized as a design workflow. And it cannot handle dynamic scene switching, such as manually rearranging the windows on multiple displays when switching from "programming IDE + database tools" to "video conference + document demonstration". Different accounts of the same hardware device cannot isolate development / entertainment scene configurations, and cross-user scene confusion occurs. SUMMARY

[0004] In view of the above problems, the application provides a multi-application-scene window layout cross-screen cooperative nonlinear restoration system to realize automatic layout of multiple screens based on application data flow and user interaction.

[0005] To achieve the above purpose, the technical scheme of the application is as follows: a multi-application-scene window layout cross-screen cooperative nonlinear restoration system, the system comprises a system data acquisition module, an application scene generation module, an intelligent layout strategy module and a scene layout restoration module; wherein, The system data acquisition module, the application scene generation module, the intelligent layout strategy module and the scene layout restoration module are all in communication connection, and are responsible for collecting application-related data; The application scene generation module is in communication connection with the system data acquisition module and the intelligent layout strategy module, analyzes the data collected by the system data acquisition module, performs multi-dimensional correlation strength analysis on the application, finally generates a high-correlation application combination scene, and delivers it to the intelligent layout strategy module; The intelligent layout strategy module is in communication connection with the system data collection module, the application scenario generation module and the scenario layout restoration module, generates multiple sets of window layout snapshots for each scenario through the window layout data collected by the system data collection module in combination with the application combination scenario, and dynamically updates the snapshot priority through the priority algorithm for the scenario layout restoration module to restore the scenario window layout. The scenario layout restoration module is in communication connection with the system data collection module and the intelligent layout strategy module, is responsible for the restoration of the scenario window layout, identifies the scenario when detecting the change of the application scenario, and then dynamically feeds back the recommended best layout to the system data collection module and restores according to the snapshot priority delivered by the intelligent layout strategy module.

[0006] Further, the system data collection module collects application-related data including application usage data and window layout data. The application usage data includes application process ID and process name AppID, application start timestamp, application close timestamp, application active duration, application foreground and background switching time, application CPU usage, application memory usage, and application network access. The window layout data includes the coordinate position (x, y) of the application window, the width (w) and height (h) of the application window, the display MonitorID where the application window is located, the layering order, the window state, the split screen information and the special area occupation. The system data collection module collects application-related data also including application file system access, system clipboard usage, and display connection.

[0007] Further, the working steps of the application scenario generation module include the following: Step 1. Calculate the correlation strength between each pair of applications to construct an application correlation strength matrix. Step 2. Convert the correlation strength matrix into a distance matrix, perform hierarchical clustering, and organize the clustering results. Step 3. Analyze the clustering results, obtain high-correlation application groups, and generate high-correlation application combination scenario information. Step 4. Store the associated application group data.

[0008] Further, the correlation strength between each pair of applications is denoted as application A and application B, and the calculation formula is as follows: CorrelationStrength(A,B) = w1 T(A,B) + w2 R(A,B) + w3 I(A,B), wherein: T(A,B) is the time dimension correlation degree of application A and application B, R(A,B) is the resource dimension correlation degree of application A and application B, I(A,B) is the interaction dimension correlation degree of application A and application B, w1, w2, w3 are corresponding weight coefficients, and w1+w2+w3=1.

[0009] Further, the calculation formula of the time dimension correlation degree of application A and application B is as follows: T(A,B) = 0.4 life cycle overlap degree + 0.4 active time coincidence degree + 0.2 start sequence correlation degree, wherein, life cycle overlap degree = overlap time / total time, the overlap time refers to the overlap time of the life cycle of application A and application B, and the total time is the sum of the life cycle of application A and application B; active time coincidence degree = common active time / total active time, the common active time refers to the overlap time of application A and application B being in the foreground at the same time or the overlap time of application A and application B being in the background at the same time, and the total active time refers to the sum of the application active time of application A and application B; start sequence correlation degree = adjacent start times / total start times, the adjacent start times refer to the number of times that application A and application B are started adjacently within a certain period of time, and the total start times refer to the sum of the start times of application A and application B within a certain period of time.

[0010] Further, the calculation formula of the resource dimension correlation degree of application A and application B is as follows: R(A,B) = 0.4 resource usage correlation degree + 0.3 IO operation correlation degree + 0.3 network access correlation degree, wherein, resource usage correlation degree = pearson(CPU_A, CPU_B) 0.5 + pearson(MEM_A, MEM_B) 0.5, pearson(CPU_A, CPU_B) is the Pearson correlation coefficient calculated for two groups of data of the CPU usage rate of application A and the CPU usage rate of application B; MEM_A and MEM_B are application memory usage rates, and pearson(MEM_A, MEM_B) is the Pearson correlation coefficient calculated for two groups of application memory usage rates; IO operation relevance = common IO operation times / total IO operation times, common IO operation times refers to the number of times that application A and application B simultaneously occur IO operation in the same time period; total IO operation times refers to the sum of IO operation times of application A and application B respectively; Network access relevance = common network access times / total network access times, common network access times refers to the number of times that application A and application B access common domain names or the same ip+port, total network access times refers to the sum of network access times of application A and application B.

[0011] Further, the interaction dimension correlation degree calculation formula of application A and application B is as follows: I(A,B) = 0.4 Window switching frequency + 0.3 Data exchange frequency + 0.3 User operation relevance, wherein, Window switching frequency = switching times between application A and application B / total switching times, data exchange frequency = data exchange times between application A and application B / total data exchange times, user operation relevance = related operation times / total operation times, related operation includes window state adjustment, window state adjustment includes maximization, minimization, window size adjustment, window edge split screen.

[0012] Further, the working steps of the intelligent layout strategy module include the following: Step 1, window state vector: according to the window layout data collected by the system data collection module, the window state is vectorized, and a state vector is generated for each window; Step 2, generate scene window layout snapshot: based on the application combination scene generated by the application scene generation module, all relevant window state vectors in each application combination scene are collected as a layout snapshot, and a parameter is added to each layout snapshot; Step 3, layout similarity deduplication: calculate the similarity between different layouts, and merge the records when the similarity is more than 90%; Step 4, store scene window layout snapshot.

[0013] Further, the parameters in step 2 include: layout ID, scene ID, creation time, update time, cumulative use times, time period use count, duration, satisfaction score, window configuration set, display configuration, layout priority; The layout priority is calculated by the following formula: Layout priority score = (a x recent use frequency + b x cumulative use duration) x time decay coefficient + c x satisfaction score + d x time matching degree, wherein, The time matching degree is the coincidence degree of the current use time of the layout snapshot and the historical use time period of the layout snapshot, the period use rate = period use count / accumulative use count, the time matching degree = historical use rate of the current period x (1+ recent use adjustment factor), a, b, c, d are weight coefficients; the time decay coefficient = e^(-λ x unused days), wherein λ is a decay rate.

[0014] Further, the scene layout restoration module detects software scene changes, identifies a current application combination scene according to the application combination, queries an application combination scene ID and a window layout thereof, matches an optimal layout according to a layout priority, and acquires a window configuration set to complete layout restoration.

[0015] The multi-application scene window layout cross-screen coordination nonlinear restoration system of the application has the following advantages: The application discloses a multi-application scene window layout cross-screen coordination nonlinear restoration system based on a domestic desktop operating system. Application scene generation: collecting use data and related window layout data of a system application, combining a domestic operating system, and generating high-correlation application combination scene information through multi-dimensional correlation strength analysis; Application window layout strategy: combining application combination scene information, capturing and storing multiple layout states, and providing layout recommendation according to layout priority and different users through memory data; Supporting scene conversion layout restoration: acquiring an optimal layout after detecting scene conversion and smoothly restoring window layout; Scene strategy continuous optimization: continuously collecting system data, optimizing application combination scenes and layout strategies, and providing more intelligent restoration strategies.

[0016] The application supports creating multiple layout versions (such as "deep coding mode" vs "code review mode") for the same group of applications and supports seamless integration of a window manager under a Kirin system. BRIEF DESCRIPTION OF DRAWINGS

[0017] The application will be further described in detail below in combination with the drawings and specific embodiments.

[0018] Figure 1 The application is a multi-application scene window layout cross-screen coordination nonlinear restoration system structure composition schematic diagram. DETAILED DESCRIPTION

[0019] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced according to other different modes without departing from the scope of the present application, and those skilled in the art can make similar extensions without departing from the concept of the present application, so the present application is not limited to the specific embodiments disclosed below.

[0020] The present application is based on the operating system level window management technology and event listening technology, and relates to application behavior analysis, scenario layout strategy and user habit modeling. In view of the problem that the traditional multi-screen display is limited to physical parameter setting and memory, a software combined semantic network (such as WPS+WeChat+DingTalk for judging as an office collaboration scenario) is established to realize the automatic layout of multi-screen applications based on application data flow and user interaction. At the same time, the nonlinear memory of the scenario state is realized, and multiple sets of layouts of the same application combination can be saved in different scenarios. The multi-application scenario window layout cross-screen collaborative nonlinear restoration system of the present application is described in detail through the following embodiments.

[0021] Embodiment 1 In combination Figure 1 As shown in the figure, the multi-application scenario window layout cross-screen collaborative nonlinear restoration system of the present application includes a system data acquisition module, an application scenario generation module, an intelligent layout strategy module and a scenario layout restoration module, wherein: The system data acquisition module is in communication connection with the application scenario generation module, the intelligent layout strategy module and the scenario layout restoration module, and is responsible for data acquisition, including application usage data and window layout data, etc. The application scenario generation module is in communication connection with the system data acquisition module and the intelligent layout strategy module, and through the analysis of the data collected by the system data acquisition module, the application is subjected to multi-dimensional correlation strength analysis, and finally a high-correlation application combination scenario is generated and delivered to the intelligent layout strategy module. The intelligent layout strategy module is in communication connection with the system data acquisition module, the application scenario generation module and the scenario layout restoration module, and through the window layout data collected by the system data acquisition module, in combination with the application combination scenario, multiple sets of window layout snapshots are generated for each scenario; at the same time, through a priority algorithm, the snapshot priority is dynamically updated for the scenario layout restoration module to restore the scenario window layout.

[0022] The scenario layout restoration module is in communication connection with the system data acquisition module and the intelligent layout strategy module, and is responsible for scenario window layout restoration. When the application scenario changes are detected, the scenario is identified, and then the snapshot priority delivered by the intelligent layout strategy module is dynamically fed back to the system data acquisition module to recommend the best layout and restore.

[0023] Through the cooperation of the four modules, the window layout restoration scheme based on the application scene is realized. For different users and different use scenes, the window layout is automatically restored, reducing the tedious operation of manually adjusting the window layout. At the same time, the system learns and iterates during use, providing more intelligent restoration schemes.

[0024] Embodiment 2 On the basis of embodiment 1, the system data acquisition module is further introduced in detail. The system data acquisition module is responsible for collecting system user application usage data and window layout data, and provides data input for the application scene generation module and the intelligent layout strategy module.

[0025] The collected application usage data includes: Application process ID and process name AppID; Application start timestamp; Application close timestamp; Application active duration (total of focus time and background running time); Application foreground and background switching time; Application CPU usage; Application memory usage; Application network access; The collected window layout data includes the following parameters: Application window coordinate position (x, y); Application window width (w) and height (h); Application window MonitorID; Stacking order: record the Z-axis order value z-order of the window; Window state: record the maximum, minimum, full screen and other special states state; Split screen information: record the system split screen or user manual split screen state; Special area occupation: record the taskbar, sidebar and other system UI element positions.

[0026] Other system information collection: Application file system access; System clipboard usage; Display connection.

[0027] Embodiment 3 On the basis of embodiment 2, the application scene generation module is further introduced in detail. The application scene generation module obtains the data of the system data acquisition module, analyzes the data, obtains the application combination with high correlation, and forms the application combination scene with high correlation. The main working steps of this module include the following: Step 1. Calculate the correlation strength between each pair of applications, and construct the correlation strength matrix between applications; Step 2. Convert the correlation strength matrix to a distance matrix (in order to perform the following clustering), perform hierarchical clustering, and sort the clustering results; Step 3. Analyze the clustering results to obtain high correlation application groups, and generate high correlation application group combination scenario information (application combination scenario ID, application list, average correlation strength); Step 4. Store the associated application group data.

[0028] The above steps complete the application correlation strength analysis according to various data of the system data acquisition module, which is the most important task of the application scenario generation module. The following examples illustrate the correlation strength calculation formula of applications A and B as follows: CorrelationStrength(A,B) = w1 T(A,B) + w2 R(A,B) + w3 I(A,B) Where: T(A,B): Time dimension correlation R(A,B): Resource dimension correlation I(A,B): Interaction dimension correlation w1, w2, w3: Corresponding weight coefficients, and w1+w2+w3=1, w1, w2, w3 take values of 0.3, 0.2, and 0.5, respectively; The above time dimension correlation T(A,B) calculation formula is as follows: T(A,B) = 0.4 Life cycle overlap degree + 0.4 Active time coincidence degree + 0.2 Start sequence correlation Wherein, the life cycle overlap degree = overlap time / total time, the life cycle is the application closing timestamp minus the application starting timestamp; the overlap time refers to the overlap time of the life cycle of application A and application B, that is, the duration of the "time intersection", that is, the duration of the overlapping part of the life cycle of the two applications; the total time is the sum of the life cycle of application A and application B.

[0029] The active time coincidence degree = common active time / total active time, the common active time and the total active time are calculated by the application foreground and background switching time obtained in embodiment 1, the common active time refers to the overlap time of application A and application B being in the foreground at the same time or in the background at the same time, and the total active time refers to the total duration of the application activity of application A and application B.

[0030] The start sequence correlation degree = adjacent start times / total start times, the adjacent start times refer to the number of times that the application A and the application B are started adjacently within a specific time period, and the total start times refer to the sum of the start times of the application A and the application B within the specific time period.

[0031] The resource dimension correlation degree R(A, B) is calculated according to the following formula: R(A, B) = 0.4 Resource usage correlation degree + 0.3 IO operation correlation degree + 0.3 Network access correlation degree Resource usage correlation degree = pearson(CPU_A, CPU_B) 0.5 + pearson(MEM_A, MEM_B) 0.5, pearson(CPU_A, CPU_B) is a Pearson correlation coefficient calculated for the CPU usage rate (CPU_A) of the application A and the CPU usage rate (CPU_B) of the application B, used to quantify the linear correlation degree of the two applications in CPU resource consumption. MEM_A and MEM_B are application memory usage rates, and the data is obtained according to the application memory usage collected in Embodiment 1; pearson(MEM_A, MEM_B) is a Pearson correlation coefficient calculated for the two sets of application memory usage rates, quantifying the linear correlation strength and direction of the two applications in memory resource consumption.

[0032] IO operation correlation degree = common IO operation times / total IO operation times, the IO operation times are obtained according to the application file system access collected in Embodiment 1, the common IO operation times refer to the number of times that the application A and the application B simultaneously perform IO operations within the same time period (such as simultaneously reading and writing the same file or directory, or having IO behavior within the same time window); and the total IO operation times refer to the sum of the IO operation times of the application A and the application B.

[0033] Network access correlation degree = common network access times / total network access times, the network access times are obtained from the application network access data collected in Embodiment 1, the common network access times refer to the number of times that the application A and the application B access the same domain name or the same ip+port, and the total network access times refer to the sum of the network access times of the application A and the application B.

[0034] The interaction dimension correlation degree I(A, B) is calculated according to the following formula: I(A, B) = 0.4 Window switching frequency + 0.3 Data exchange frequency + 0.3 User operation relevance Wherein, the window switching frequency = the number of switching between application A and application B / the total number of switching, the data exchange frequency = the number of data exchange between application A and application B / the total number of data exchange, the user operation relevance = the number of relevant operations / the total number of operations, the relevant operation mainly refers to the window state adjustment, including: maximizing, minimizing, window size adjustment, window edge split screen, etc.

[0035] The above user operation relevance reflects the proportion of the operation between application A and application B in the whole, the number of relevant operations is the number of relevant operations between application A and application B (i.e. the operation of maximizing, minimizing, window edge split screen, etc. involving application A and application B), the total number of operations refers to the total number of relevant operations between all applications, which is used for normalization, specifically: The number of relevant operations between application A and application B, for example: application A and application B window edge split screen, application A minimized and switched to application B, application A and application B window size adjustment, etc. The total number of relevant operations between all applications in the statistical period, for example: the total number of relevant operations of application A and application B, application A and application C, application B and application C, application D and application E, etc.

[0036] That is, the number of relevant operations is the number of relevant operations between application A and application B, and the total number of operations is the total number of relevant operations between all applications.

[0037] Embodiment 4 On the basis of embodiment 3, the intelligent layout strategy module is further introduced in detail. The intelligent layout strategy module obtains the window layout data of the system data collection module, combines the application combination scene generated by the application scene generation module, analyzes and processes to form a layout strategy, and realizes a dynamic adjustment layout priority algorithm, and provides an intelligent layout strategy for the application combination scene.

[0038] The working steps of the intelligent layout strategy module include the following: Step 1, window state vector: according to the window layout data collected by the system data acquisition module, the window state is vectorized, and a state vector [AppID, MonitorID, x, y, w, h, z-order, state] is generated for each window; wherein AppID is the application process name, MonitorID refers to the display ID where the application window is located, x, y refers to the coordinate position of the application window, w is the width of the application window, h is the height of the application window, z-order is the Z-axis order value of the recorded window, and state refers to the window state (including recording maximization, minimization, full screen, etc. Special state); Step 2, generate scene window layout snapshot: based on the application combination scene generated by the application scene generation module, all relevant window state vectors in each application combination scene are collected as a layout snapshot, (all relevant windows refer to the window arrangement state of the associated application combination scene, such as opening wps and browser, adjusting to a window distribution state that you like, wps on the right, browser on the left, etc.), and each layout snapshot is attached with parameters, including: a. Layout ID: unique identifier of the layout scheme; b. Scene ID: associated application combination scene ID (one application combination scene ID can be associated with multiple layout snapshots); c. Creation time: first recording time; d. Update time: the latest update time; e. Cumulative use times: cumulative value of the number of times the current layout snapshot is applied and restored by the user, that is, the number of times the user selects to use after the user is recommended by the associated application combination scene The window layout; f. Time period use count: in the morning (9am-12pm), noon (1pm-3pm), afternoon (3pm-6pm) and evening (6pm-9pm) time period, the number of times the current window layout snapshot is applied and restored by the user; g. Duration: cumulative use time, the scene window layout snapshot can store multiple for an associated application combination scene, the duration refers to the cumulative time of using the current window layout snapshot; the scene layout restoration module detects that the scene has changed, then selects a different snapshot, this snapshot is strongly related to the associated application combination scene, when using this snapshot, the applications of the corresponding application combination scene are running; h. Satisfaction score: satisfaction score (0-100) derived from user modification behavior; i. Window configuration set: all window state vector sets; j. Display configuration: current display number, resolution and relative position; k. Layout priority: priority score of the current layout snapshot; Step 3, layout similarity deduplication: calculate the similarity between different layouts, and merge the records when the similarity exceeds 90%; Step 4, store the scene window layout snapshot.

[0039] The priority dynamic adjustment algorithm adopted in the intelligent layout strategy module for the priority of the layout snapshot dynamically adjusts the priority score of the layout snapshot, and its comprehensive score formula is as follows: Layout priority score = (a x recent use frequency + b x cumulative use time) x time decay coefficient + c x satisfaction score + d x time matching degree; Among them, the time matching degree is the relevance of collecting the use frequency and time characteristics of the historical layout snapshot, that is, the degree of coincidence between the current use time of the layout snapshot and the historical use time period of the layout snapshot, such as the user likes to use this application group of this layout snapshot in the morning; According to the period use count and cumulative use count, the period use rate can be calculated, and then the time matching degree can be calculated: Period use rate = period use count / cumulative use count Time matching degree = historical use rate of current period x (1 + recent use adjustment factor), recent use adjustment factor: if the layout has use records in the same period in the last 3 days, it is 0.5, otherwise it is 0; Where a, b, c, d are weight coefficients, initially set to 0.368, 0.211, 0.263, 0.158, and keep a+b+c+d=1.

[0040] Time decay function: use an exponential decay function to reduce the priority of long-term unused layouts, Time decay coefficient = e^(-λ x days not used) Where λ is the decay rate, usually set to 0.1-0.3.

[0041] Embodiment 5 On the basis of embodiment 4, the scene layout restoration module is further introduced in detail. The scene layout restoration module detects software scene changes, identifies the current application combination scene according to the application combination, queries the application combination scene ID and its window layout, matches the best layout according to the layout priority (if the priority is the same, then according to the recently used layout to decide), and obtains the window configuration set to complete the layout restoration.

[0042] Specifically, when the scene layout restoration module detects a software scene change, the best layout given by the intelligent layout strategy module is used for restoration. After the user determines, there may be fine tuning, which may be window layout, or increase the opened application. The result is dynamically fed back to the system data collection module to collect data, update the application combination scene or layout strategy for subsequent selection. Through this cycle, the system slowly learns and approaches the user's usage habits, and then gives a more intelligent restoration.

[0043] Embodiment 6 This embodiment is a specific application example of the system described in the above embodiments. The following describes the working process of the system of the present application with a typical user usage scenario as an example, including the following steps: Step 1. The user opens the machine to enter the desktop system, opens the VSCode application + terminal application + browser application. Step 2. According to the scene window layout stored in the intelligent layout strategy module, the best layout of the current application combination is matched and automatically executed: (1) Fix VSCode on the left half of the screen; (2) Fix the browser on the upper half of the right half of the screen; (3) Fix the terminal on the lower half of the right half of the screen.

[0044] Step 3. The scene layout restoration module detects that a new display is connected, the screen is set to an extended mode, and automatically executes: (1) Fix VSCode on the first screen; (2) Fix the browser on the upper half of the second screen; (3) Fix the terminal on the lower half of the second screen.

[0045] Step 4. The user closes VSCode and the terminal and opens WPS. Step 5. The current running is WPS and the browser, the scene layout restoration module detects a software scene change, and according to the scene window layout stored in the intelligent layout strategy module, the application combination scene with the highest correlation degree is obtained by matching the current application combination, and automatically executed: (1) Fix WPS on the first screen; (2) Fix the browser on the second screen; (3) Open the file manager and suspend it in the middle of the second screen.

[0046] Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor belong to the scope of protection of the present application.

Claims

1. A multi-application scenario oriented window layout cross-screen cooperative nonlinear restoration system, characterized in that, The system comprises a system data collection module, an application scenario generation module, an intelligent layout strategy module and a scenario layout restoration module, wherein The system data collection module is in communication connection with the application scenario generation module, the intelligent layout strategy module and the scenario layout restoration module, and is responsible for collecting application-related data; The application scenario generation module is in communication connection with the system data collection module and the intelligent layout strategy module, analyzes the data collected by the system data collection module, performs multi-dimensional correlation strength analysis on the application, finally generates a high-correlation application combination scenario, and delivers it to the intelligent layout strategy module; The intelligent layout strategy module is in communication connection with the system data collection module, the application scenario generation module and the scenario layout restoration module, generates multiple sets of window layout snapshots for each scenario by combining the application combination scenario with the window layout data collected by the system data collection module; simultaneously, the snapshot priority is dynamically updated by a priority algorithm, so as to provide the scenario layout restoration module with the restoration of the scenario window layout; The scenario layout restoration module is in communication connection with the system data collection module and the intelligent layout strategy module, is responsible for the restoration of the scenario window layout, identifies the scenario when detecting the change of the application scenario, and then dynamically feeds back the recommended best layout to the system data collection module and restores it according to the snapshot priority delivered by the intelligent layout strategy module. 2.The multi-application scenario oriented window layout cross-screen cooperative nonlinear reduction system according to claim 1, characterized in that, The system data collection module collects application-related data including application usage data and window layout data; The application usage data includes application process ID and process name AppID, application start timestamp, application close timestamp, application active duration, application foreground and background switching time, application CPU usage, application memory usage, and application network access; The window layout data includes the coordinate position (x, y) of the application window, the width (w) and height (h) of the application window, the display MonitorID where the application window is located, the layering order, the window state, the split-screen information and the special area occupation; The system data collection module collects application-related data, which further includes application file system access, system clipboard usage and display connection. 3.The multi-application scenario oriented window layout cross-screen cooperative nonlinear reduction system according to claim 2, characterized in that, The working steps of the application scenario generation module include the following: Step 1. Calculate the correlation strength between each pair of applications, and construct the correlation strength matrix between the applications; Step 2. Convert the correlation strength matrix into a distance matrix, perform hierarchical clustering, and sort the clustering results; Step 3. Analyze the clustering results, obtain a high-correlation application group, and generate high-correlation application combination scenario information; Step 4. Store the correlation application group data.

4. The multi-application-scene-oriented window layout cross-screen coordination non-linear restoration system according to claim 3, characterized in that, Step 1 The correlation strength between each pair of applications is denoted as application A and application B, and the calculation formula is as follows: CorrelationStrength(A, B) = w1 T(A, B) + w2 R(A, B) + w3 I(A, B), wherein: T(A, B) is the time dimension correlation degree of application A and application B, R(A, B) is the resource dimension correlation degree of application A and application B, I(A, B) is the interaction dimension correlation degree of application A and application B, w1, w2, w3 are corresponding weight coefficients, and w1+w2+w3=1.

5. The multi-application scenario oriented window layout cross-screen coordination non-linear restoration system according to claim 4, characterized in that, The calculation formula of the time dimension correlation degree of application A and application B is as follows: T(A,B) = 0.4 Life cycle overlap + 0.4 Active time overlap + 0.2 Launch sequence correlation, wherein, Life cycle overlap degree = overlap time / total time, wherein the overlap time refers to the overlap time of the life cycles of application A and application B, and the total time refers to the sum of the life cycles of application A and application B; Active time coincidence degree = common active time / total active time, wherein the common active time refers to the overlap time of the foreground or the overlap time of the background when application A and application B are in the foreground at the same time, and the total active time refers to the sum of the application active durations of application A and application B; The start sequence correlation degree = adjacent start times / total start times, the adjacent start times refer to the number of times that the application A and the application B are started adjacently within a specific time period, and the total start times refer to the sum of the start times of the application A and the application B within the specific time period. 6.The multi-application scenario oriented window layout cross-screen cooperative nonlinear reduction system according to claim 4, characterized in that, The resource dimension correlation degree of the application A and the application B is calculated according to the following formula: R(A,B) = 0.4 Resource usage relevance + 0.3 IO operation relevance + 0.3 Network access relevance, wherein, Resource usage correlation = pearson(CPU_A, CPU_B) 0.5 + pearson(MEM_A, MEM_B) 0.5, pearson(CPU_A, CPU_B) is the Pearson correlation coefficient calculated for the CPU usage of application A and the CPU usage of application B; MEM_A, MEM_B are the application memory usage, and pearson(MEM_A, MEM_B) is the Pearson correlation coefficient calculated for the two sets of application memory usage. The IO operation correlation degree = common IO operation times / total IO operation times, the common IO operation times refer to the number of times that the application A and the application B simultaneously perform IO operations within the same time period, and the total IO operation times refer to the sum of the IO operation times of the application A and the application B; The network access correlation degree = common network access times / total network access times, the common network access times refer to the number of times that the application A and the application B access the same domain name or the same ip+port, and the total network access times refer to the sum of the network access times of the application A and the application B. 7.The multi-application scenario oriented window layout cross-screen cooperative non-linear restoration system according to claim 4, characterized in that, The interaction dimension correlation degree of the application A and the application B is calculated according to the following formula: I(A,B) = 0.4 Window switching frequency + 0.3 Data exchange frequency + 0.3 User operation relevance, wherein, The window switching frequency = the number of times of switching between the application A and the application B / the total switching times, the data exchange frequency = the number of times of data exchange between the application A and the application B / the total data exchange times, and the user operation correlation degree = the number of related operations / the total operation times, the related operations including window state adjustment, window state adjustment including maximization, minimization, window size adjustment, and window edge split screen. 8.The multi-application scenario oriented window layout cross-screen cooperative nonlinear reduction system according to claim 3, characterized in that, The working steps of the intelligent layout strategy module include the following: Step 1, window state vector: according to the window layout data collected by the system data collection module, the window state is vectorized, and a state vector is generated for each window; Step 2, generate scene window layout snapshot: based on the application combination scene generated by the application scene generation module, all relevant window state vectors in each application combination scene are collected as a layout snapshot, and a parameter is added to each layout snapshot; Step 3, layout similarity deduplication: calculate the similarity between different layouts, and merge the records when the similarity is more than 90%; Step 4, store the scene window layout snapshot. 9.The multi-application scenario oriented window layout cross-screen collaboration nonlinear reduction system according to claim 8, characterized in that, The parameters in step 2 include: layout ID, scene ID, creation time, update time, cumulative use times, time period use count, duration, satisfaction score, window configuration set, display configuration, and layout priority; The layout priority is calculated by the following formula: Layout priority score = (a x recent use frequency + b x cumulative use duration) x time decay coefficient + c x satisfaction score + d x time matching degree, wherein, The time matching degree is the degree of overlap between the current use time of the layout snapshot and the historical use time period of the layout snapshot, the time period use rate = time period use count / cumulative use count, and the time matching degree = current time period historical use rate x (1 + recent use adjustment factor), a, b, c, and d are weight coefficients; the time decay coefficient = e^(-λ x unused days), wherein λ is the decay rate.

10. The multi-application-scene window layout cross-screen coordination non-linear restoration system according to claim 9, characterized in that, The scene layout restoring module detects software scene change, identifies current application combination scene according to application combination, queries application combination scene ID and window layout, matches optimal layout according to layout priority, and acquires window configuration set to complete layout restoring.

Images