Sharing monitoring method and device for ecological flow data

Abstract

The invention relates to the technical field of big data processing, in particular to a sharing monitoring method and device for ecological flow data, and the method comprises the steps: collecting the ecological flow data when the resource loss of an ecological environment is recognized; judging a monitoring recovery type associated with the ecological flow data; under the condition that the monitoring recovery type is the first type, determining an available ecological sensor queue in a preset strategy mapping table according to the ecological flow data; under the condition that the monitoring recovery type is a second type, purifying and restoring the ecological flow data by using a deep adversarial sample generation network deployed in advance, and determining an available ecological sensor queue; each ecological sensor in the queue of available ecological sensors is enabled to perform shared monitoring for ecological flow data. Mass ecological flow operation data are fully utilized, monitoring is accurate and timely, and more effective information is provided for production, operation and management of a multi-source ecological system enterprise.

Description

Technical Field

[0001] The present invention relates to the field of big data processing technology, and in particular to a shared monitoring method and device for ecological flow data. Background Art

[0002] The rapid development of my country's multi-source ecosystems, the rapid growth of multi-source ecosystem regions, the special location of multi-source ecosystems and the volatility of multi-source ecosystems have brought huge challenges to the monitoring of multi-source ecosystem failures. Traditional monitoring systems have the problems of excessive monitoring, inaccurate monitoring and untimely monitoring, while existing big data systems have not been fully utilized and have the following defects and deficiencies: Massive amounts of ecological flow operation data, maintenance data and environmental data are wasted, but the huge amount of data cannot be filtered, processed, analyzed and formed into more effective information for the production and operation management of multi-source ecosystem enterprises through software systems within a reasonable time.

[0003] The existing monitoring does not associate related components and cannot achieve the advance and accurate monitoring.

[0004] There is no multi-source ecosystem unit health management platform based on a network algorithm platform, and the unpredictability of the health status of multi-source ecosystem units in multi-source ecosystem fields poses a management risk. Summary of the invention

[0005] To achieve the above object, the present invention provides the following technical solutions: According to a first aspect of the present invention, the present invention claims a method for sharing and monitoring ecological flow data, the method comprising: When resource loss is identified in the ecological environment, collect ecological flow data; Determining the monitoring restoration type associated with the ecological flow data; Under the condition that the monitoring recovery type is the first type, determining the available ecological sensor queue in a preset strategy mapping table according to the ecological flow data, wherein the first type refers to a type associated with resource loss with a pre-configured monitoring strategy and a resource loss cause; Under the condition that the monitoring and restoration type is the second type, the ecological flow data is purified and repaired using a pre-deployed deep adversarial sample generation network to determine an available ecological sensor queue, wherein the second type refers to a type of resource loss associated with no inducement for resource loss or no pre-configured monitoring strategy; Each ecological sensor in the available ecological sensor queue is enabled to perform shared monitoring of ecological flow data.

[0006] Furthermore, the strategy mapping table stores the association between the ecological traffic resource id and the sensor queue id; The determining of the available ecological sensor queue in a preset strategy mapping table according to the ecological flow data includes: Obtaining an ecological flow resource id from the ecological flow data; According to the ecological flow resource ID, the association relationship between the ecological flow resource ID and the sensor queue ID is searched to determine the associated ecological sensor ID; An associated available ecological sensor queue is determined according to the associated ecological sensor ID.

[0007] Furthermore, the use of the pre-deployed deep adversarial sample generation network to purify and repair the ecological traffic data and determine the available ecological sensor queue includes: Identify ecological zones where resource loss occurs from the ecological flow data; Collecting variable values ​​of various parameters of the ecological zone where resource loss occurs; Based on each parameter, determining the health of the parameter according to the variable value of the parameter; According to the health of the parameter, the GAN matrix is ​​queried, and the GAN value associated with each ecological sensor ID is output, the GAN matrix is ​​a behavioral health sequence, the columns of the GAN matrix are each ecological sensor ID, and the GAN value associated with the health and ecological sensor ID binary in the GAN matrix is ​​output through an adversarial network algorithm; Selecting an available ecological sensor ID according to the GAN value; An available ecological sensor queue is constructed according to the available ecological sensor IDs.

[0008] Furthermore, the process of outputting the GAN value in the GAN matrix through the adversarial network algorithm includes: Pre-cleaning and repairing the GAN values ​​in the GAN matrix; Based on each health metric in a row of the GAN matrix, an eco-sensor id is arbitrarily adopted; Enable the eco-sensor associated with the eco-sensor id; Collecting the ecological environment health and discrimination value of the ecological environment after the ecological sensor is activated; The updated GAN value of the health degree and the ecological sensor ID tuple is determined according to the ecological environment health degree, the discrimination value and the pre-configured GAN value calculation strategy, and the process of arbitrarily adopting an ecological sensor ID is returned to complete until the loop stop condition is met.

[0009] Furthermore, the GAN value calculation strategy is as follows: ; Among them, s is the original health, is the adversarial sample generation rate, is the discriminant value, γ is the overfitting suppression coefficient, New health The maximum GAN value under ; After the ecological sensor is enabled, the resource loss is stopped and no new resource loss is obtained, the judgment value is a positive value. If the resource loss does not stop after the ecological sensor is enabled, or new resource loss is obtained after the ecological sensor is enabled, or the ecological environment performance cannot meet the standard activation conditions, the discrimination value Is a negative value.

[0010] Furthermore, the ecological flow data includes ecological flow resource variables; The enabling of each ecological sensor in the available ecological sensor queue includes: Inputting the ecological flow resource variables into the respective ecological sensors; The ecological sensors are enabled one by one according to the positioning position and data causal relationship of each ecological sensor in the available ecological sensor queue.

[0011] Furthermore, the method further comprises: Collect resource loss purification and repair results; The deep adversarial sample generation network is optimized using the eco-sensor IDs included in the available eco-sensor queue and the resource loss purification and repair results.

[0012] According to a second aspect of the present invention, the present invention claims protection for a shared monitoring device for ecological flow data, the device comprising: The ecological flow data collection unit collects ecological flow data when it identifies resource loss in the ecological environment; A monitoring and restoration type determination unit, for determining a monitoring and restoration type associated with the ecological flow data; A policy library purification and restoration unit, under the condition that the monitoring and restoration type is the first type, determines the available ecological sensor queue in a preset policy mapping table according to the ecological flow data, wherein the first type refers to a type associated with resource loss with a pre-configured monitoring policy and a resource loss cause; Strengthening the adversarial sample generation purification and repair unit, under the condition that the monitoring and recovery type is the second type, using the pre-deployed deep adversarial sample generation network to purify and repair the ecological flow data, and determine the available ecological sensor queue, the second type refers to the type of resource loss associated with no resource loss inducement or no pre-configured monitoring strategy; A sensor enabling unit, which enables each ecological sensor in the available ecological sensor queue to perform shared monitoring of ecological flow data; The shared monitoring device for ecological flow data completes the shared monitoring method for ecological flow data.

[0013] The present invention relates to the field of big data processing technology, and in particular to a shared monitoring method and device for ecological flow data. When it is identified that there is resource loss in the ecological environment, the ecological flow data is collected; the monitoring recovery type associated with the ecological flow data is determined; under the condition that the monitoring recovery type is the first type, the available ecological sensor queue is determined in a preset strategy mapping table according to the ecological flow data; under the condition that the monitoring recovery type is the second type, the ecological flow data is purified and repaired using a pre-deployed deep adversarial sample generation network to determine the available ecological sensor queue; each ecological sensor in the available ecological sensor queue is enabled to perform shared monitoring of the ecological flow data. The present invention makes full use of massive ecological flow operation data, monitors accurately and timely, and provides more effective information for the production, operation and management of multi-source ecosystem enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A workflow diagram of a shared monitoring method for ecological flow data claimed in an embodiment of the present invention; Figure 2 A structural module diagram of a shared monitoring device for ecological flow data claimed in an embodiment of the present invention; Reference numerals: Ecological flow data collection unit-21, monitoring and recovery type judgment unit-22, strategy library purification and repair unit-23, enhanced adversarial sample generation purification and repair unit-24, sensor activation unit-25. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field without creative work according to the embodiments of the present invention are within the scope of protection of the present invention.

[0016] The terms "first", "second" and "third" in the present invention only describe the purpose, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention only explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, ecological environment, product or equipment that includes a series of processes or units is not limited to the listed processes or units, but optionally also includes processes or units that are not listed, or optionally also includes other processes or units inherent to these processes, methods, products or equipment.

[0017] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0018] An embodiment of the present invention provides a shared monitoring method for ecological flow data, the method comprising: collecting ecological flow data when resource loss is identified in the ecological environment; determining the monitoring recovery type associated with the ecological flow data; under the condition that the monitoring recovery type is the first type, determining the available ecological sensor queue in a preset strategy mapping table based on the ecological flow data, wherein the first type refers to the type of resource loss associated with the presence of resource loss inducements and a pre-configured monitoring strategy; under the condition that the monitoring recovery type is the second type, using a pre-deployed deep adversarial sample generation network to purify and repair the ecological flow data to determine the available ecological sensor queue, wherein the second type refers to the type of resource loss associated with the absence of resource loss inducements or a pre-configured monitoring strategy; enabling each ecological sensor in the available ecological sensor queue to perform shared monitoring of the ecological flow data.

[0019] After collecting ecological flow data, the type of resource loss is determined based on the ecological flow data. If the cause of resource loss exists and there is a pre-configured monitoring strategy, the pre-configured monitoring strategy is used to determine the ecological sensors that need to be completed, and the above ecological sensors are completed to conduct shared monitoring of ecological flow data. If the cause of resource loss does not exist or there is no related monitoring strategy, the ecological flow data is purified and repaired using the enhanced adversarial sample generation algorithm, the ecological sensors that need to be completed are determined, and the above ecological sensors are completed to conduct shared monitoring of ecological flow data. Not only does it rely on pre-configured expert strategies to deal with existing resource loss, but it can also dynamically purify and repair non-existent or complex resource loss situations, significantly improving the ability of the healthy cloud platform virtual ecosystem ecological environment to deal with non-existent resource loss and the flexibility of the cloud platform virtual ecosystem ecological environment to purify and repair resource loss.

[0020] The following is a detailed description of the shared monitoring method for ecological flow data provided by the present invention in conjunction with the accompanying drawings and specific implementation methods.

[0021] Embodiment 1: Figure 1 This is a flow chart of a shared monitoring method for ecological flow data in an embodiment of the present invention. This embodiment can be applied to the situation where resource loss in a cloud platform virtual ecosystem is purified and repaired. The method can be completed by a shared monitoring device for ecological flow data. The shared monitoring device for ecological flow data can be implemented in software and / or hardware. The shared monitoring device for ecological flow data can be configured in a cloud platform virtual ecosystem or an intelligent service distributed system. In the embodiment of the present invention, the shared monitoring method for ecological flow data is applied in a cloud platform virtual ecosystem as an example for explanation.

[0022] The shared monitoring method for ecological flow data provided by an embodiment of the present invention is applied in a cloud platform virtual ecosystem. After receiving an ecological environment startup instruction and the cloud platform virtual ecosystem is started, the shared monitoring method for ecological flow data begins to be completed and automatically repairs the resource loss in the cloud platform virtual ecosystem.

[0023] like Figure 1 As shown, the shared monitoring method for ecological flow data provided by the embodiment of the present invention mainly includes processes S101-S105.

[0024] S101. When resource loss in the ecological environment is identified, collect ecological flow data.

[0025] The above ecological environment may refer to a service distributed system including one or more cloud terminals. In the embodiment of the present invention, the ecological environment is taken as an example of a cloud platform virtual ecosystem.

[0026] Resource loss in the ecological environment means that any one or more ecological resource elements in the service distributed system have resource loss and cannot be used in a standard manner. The above ecological resource elements include: any cloud terminal in the computing array, any cloud terminal in the storage array, the cloud terminal of the login node, or the cloud terminal of the management node.

[0027] Among them, the computing array is mainly responsible for supporting the activation of healthy performance computing applications in the cloud platform virtual ecosystem ecological environment. The login node is generally used by users to log in to the cloud platform virtual ecosystem ecological environment in order to use computing resources. The management node mainly deploys the ecological environment management software of the cloud platform virtual ecosystem, and the storage array stores the data obtained by the healthy performance computing application.

[0028] Among them, the resource loss in the virtual ecosystem of the cloud platform may include but is not limited to the following incentives: resource loss in the original ecological environment, resource loss caused by application sensor anomalies, resource loss caused by concurrent consumption anomalies, resource loss caused by unbalanced multi-source coordination, resource loss caused by poor resource circulation, resource loss caused by resource management problems, resource loss caused by invasion of alien species, resource loss caused by competition among local species, and resource loss caused by environmental pollution.

[0029] Resource loss in the cloud platform virtual ecosystem also includes: resource loss without inducements or resource loss in complex situations. Resource loss without inducements means that only resource loss in the ecological environment can be identified, but the inducement of resource loss cannot be determined. For example: there is wind energy resource loss in the cloud platform virtual ecosystem, but the inducement of wind energy loss cannot be determined. Another example: the land managed by the cloud platform virtual ecosystem is in an overloaded state for a long time, but the specific inducement of the land overload for a long time cannot be determined. Resource loss in complex situations means that the resource loss in the cloud platform virtual ecosystem is caused by multiple inducements.

[0030] After the cloud platform virtual ecosystem is started, the activation health of the cloud platform virtual ecosystem is identified in real time to determine whether there is resource loss in the supercomputing. After determining that there is resource loss, the ecological flow data is collected. Among them, the ecological flow data represents information related to resource loss. For example, the ecological flow data may include ecological resource elements with resource loss, resource loss phenomena, and the causes of resource loss.

[0031] In a possible implementation, an ecological traffic resource listener and a listening strategy associated with the ecological traffic resource listener are set in advance, and setting information based on the ecological traffic resource listener can be stored in a configuration memory.

[0032] After the cloud platform virtual ecosystem is started, the ecological traffic resource listener uses the ecological traffic resource listener strategy to identify and evaluate the health of the cloud platform virtual ecosystem in real time, and collects ecological traffic data when it is identified that there is resource loss in the cloud platform virtual ecosystem.

[0033] According to an ecological flow resource intensity type definition and example provided in an embodiment of the present invention, the ecological flow resource type mainly includes four types of ecological flow resource intensities, namely notification, warning, resource loss, and disaster.

[0034] Since the monitoring status of the ecological flow resource listener is set in advance, during the activation of the cloud platform virtual ecosystem, the activation health of the cloud platform virtual ecosystem is identified in real time, and the level of ecological flow data is determined based on the activation health and the monitoring status set in advance. For example: during the activation of the cloud platform virtual ecosystem, if it is identified that the ozone layer integrity rate exceeds 70%, then the warning monitoring status is reached, and ecological flow data related to the reminder is generated. For another example: during the activation of the cloud platform virtual ecosystem, if a service ecological resource element is identified to be deactivated, then the resource loss monitoring status is reached, and ecological flow data related to resource loss is generated.

[0035] S102. Determine the monitoring and restoration type associated with the ecological flow data.

[0036] In a possible implementation, when the ecological flow resource listener and the monitoring strategy associated with the ecological flow resource listener are set in advance, the resource loss type associated with the ecological flow resource listener can be set at the same time. The setting information based on the ecological flow resource listener can be stored in the configuration memory.

[0037] Resource loss types can be divided according to whether the causes of resource loss exist and the complexity of resource loss. Monitoring and recovery types are divided into three categories, including the third type, the second type, and the first type.

[0038] The first type represents the matching monitoring and restoration type, that is, there is a clear expert monitoring strategy based on the existing resource loss in the strategy mapping table, and the ecological environment can be restored to the standard health level through automated ecological sensors.

[0039] The second type indicates the possible monitoring and restoration type. That is, it is possible to self-purify and restore the ecological environment, but there is no clear purification and restoration strategy. For example, resource losses without various inducements basically belong to the possible monitoring and restoration type. The deep adversarial sample generation network is used to self-purify and restore this type of resource loss.

[0040] The third type represents the unmonitorable recovery type, that is, the resource loss that cannot be repaired by automated purification and repair methods, and requires manual intervention for purification and repair. Exemplary, the unmonitorable resource loss may include non-renewable resource loss, such as ozone layer damage, oil resource loss, stratum disconnection, etc., which cannot be repaired by automated purification and repair methods and can only be repaired by manual intervention.

[0041] Before the virtual ecosystem of the cloud platform is enabled, the monitoring and recovery types associated with each resource loss are set in advance. For example, the monitoring and recovery types associated with the loss of oil resources and the damage to the ozone layer are of the third type, that is, the non-monitorable recovery type. Another example: the ecological environment of the virtual ecosystem of the cloud platform is attacked by a disaster, and the disaster cannot be intercepted and cleaned up. The monitoring and recovery type associated with this resource loss is of the second type, that is, the possible monitoring and recovery type. Another example: due to improper resource allocation, a certain land load is too large. This type of resource loss has an expert monitoring strategy. Therefore, the monitoring and recovery type associated with this resource loss is of the first type, that is, the matching monitoring and recovery type.

[0042] The monitoring and recovery types associated with each resource loss are set in advance and stored in the configuration storage device. After identifying that there is resource loss in the cloud platform virtual ecosystem, the above-mentioned association relationship is queried according to the resource loss in the cloud platform virtual ecosystem to determine the monitoring and recovery type associated with the resource loss in the cloud platform virtual ecosystem, that is, the resource loss type associated with the ecological flow data.

[0043] Furthermore, the monitoring and recovery types associated with each resource loss can be set before the cloud platform virtual ecosystem is started, or can be set or updated after the cloud platform virtual ecosystem is started. The above update includes adding new resource losses and their associated monitoring and recovery types, modifying the monitoring and recovery types associated with existing resource losses, or updating resource losses.

[0044] S103. Under the condition that the monitoring recovery type is the first type, determine the available ecological sensor queue in the preset strategy mapping table according to the ecological flow data, wherein the first type refers to the type associated with resource loss that has a resource loss inducement and a pre-configured monitoring strategy.

[0045] The monitoring and restoration type of ecological flow data is the first type, which means that the resource loss in the virtual ecosystem of the cloud platform can be determined by the policies in the existing policy mapping table. The available ecological sensor queue can include multiple ecological sensors, and the positioning positions and data causal relationships between each ecological sensor are configured.

[0046] The policy mapping table sets the association relationship between the ecological flow data and the ecological sensor queue ID associated with the ecological flow data. After determining that the monitoring recovery type associated with the ecological flow data is the first type, the above policy data is queried according to the ecological flow data, the queried ecological sensor queue ID is used as the available ecological sensor queue ID, and the available ecological sensor queues associated with the available ecological sensor queue ID are collected.

[0047] In a possible implementation, a shared monitoring strategy pair for ecological flow data is summarized, that is, the ecological flow data that can be monitored and the ecological sensor queue ID that can stop resource loss are bound together as a monitoring strategy and input into a strategy mapping table. The ecological sensor queue may include multiple ecological sensor IDs and the positioning positions and data causal relationships between each ecological sensor.

[0048] The association between the ecological traffic data and the ecological sensor queue ID associated with the ecological traffic data is stored in the expert strategy table in the configuration memory. The expert strategy table specifically includes the following information: strategy creation time, operation script, whether to enable, and strategy update time.

[0049] After each ecological sensor included in the ecological sensor queue is enabled, the cloud platform virtual ecosystem ecological environment can be restored to the health before resource loss, that is, the shared monitoring of ecological flow data of the cloud platform virtual ecosystem is realized. In other words, the monitoring strategy stored in the strategy mapping table is a series of ecological sensor queues. The above-mentioned available ecological sensor queue represents the monitoring strategy for purifying and repairing the current resource loss of the cloud platform virtual ecosystem.

[0050] In a possible implementation, the association between the ecological traffic resource id and the sensor queue id is stored in the policy mapping table; the available ecological sensor queue is determined in the preset policy mapping table based on the ecological traffic data, including: obtaining the ecological traffic resource id from the ecological traffic data; querying the association between the ecological traffic resource id and the sensor queue id based on the ecological traffic resource id to determine the associated ecological sensor id, and determining the associated available ecological sensor queue based on the ecological sensor id.

[0051] The ecological flow resource template may be directly used as the ecological flow resource id, or a sequence number may be assigned to each ecological flow resource template as the ecological flow resource id. In this embodiment, the ecological flow resource template is used as the ecological flow resource id as an example for description.

[0052] In a possible implementation, after collecting the ecological flow data, the ecological flow data is split into two parts: the ecological flow resource template and the ecological flow resource variable. According to the ecological flow resource template, the ecological sensor ID associated with the ecological flow resource template is queried in the policy mapping table to determine the ecological sensor ID associated with the ecological flow resource template, and the ecological sensors associated with each ecological sensor ID are used as the available ecological sensor queue. For example, the ecological sensor IDs associated with the ecological flow resource template include (a1, a2, a3, a4), then the 4 ecological sensors associated with (a1, a2, a3, a4) are collected as the available ecological sensor queue.

[0053] In another possible implementation, the eco-sensor queue id associated with the eco-flow resource template is queried in the policy mapping table according to the eco-flow resource template, the eco-sensor queue id is associated with multiple eco-sensor ids, and the eco-sensors associated with each eco-sensor id are used as available eco-sensor queues. For example: the eco-sensor queue id associated with the eco-flow resource template is A1, and A1 is configured in advance to include 4 eco-sensor ids (a1, a2, a3, a4), then the 4 eco-sensors associated with (a1, a2, a3, a4) included in A1 are collected as available eco-sensor queues.

[0054] S104. Under the condition that the monitoring and recovery type is the second type, the ecological flow data is purified and repaired using the deep adversarial sample generation network deployed in advance to determine the available ecological sensor queue. The second type refers to the type of resource loss associated with no inducement for resource loss or no pre-configured monitoring strategy.

[0055] The monitoring recovery type associated with ecological flow data is the second type, which indicates the resource loss in the virtual ecosystem of the cloud platform. There is currently no clear expert monitoring strategy, but it is possible to identify an available ecological sensor fleet that may achieve shared monitoring of ecological flow data through the enabled health of supercomputing and deep adversarial sample generation networks.

[0056] In a possible implementation, a pre-deployed deep adversarial sample generation network is used to purify and repair ecological flow data to determine an available ecological sensor queue, including: determining ecological zones with resource loss from ecological flow data; collecting variable values ​​of various parameters of ecological zones with resource loss; determining the health of the parameters based on the variable values ​​of the parameters; querying the GAN matrix based on the health of the parameters, and outputting the GAN values ​​associated with each ecological sensor id, the rows of the GAN value matrix are each health, the columns of the GAN matrix are each ecological sensor id, and the GAN values ​​associated with the health and ecological sensor id binary in the GAN matrix are output through an adversarial network algorithm; selecting available ecological sensor ids based on the GAN values; and constructing an available ecological sensor queue based on multiple available ecological sensor ids.

[0057] An ecological zone with resource loss refers to a cloud terminal or land where resource loss occurs. In the shared monitoring scenario for ecological flow data, the above parameters include but are not limited to the following information: land planting rate, memory integrity rate, ozone layer I / O, network traffic, all specific software ecological environment monitoring parameters and log information in the resource loss ecological zone. For example: a specific software ecological environment monitoring parameter indicates that the deep adversarial sample generation network is expected to purify and repair the file ecological environment software resource loss, so it is necessary to add the relevant parameters of the file ecological environment software when deploying the model.

[0058] The parameter health is determined using the variable values ​​of each parameter provided above. The parameter variable value refers to the specific value of each parameter. The parameter health is represented as the id of the parameter range. The relationship between the range of each parameter variable value and the parameter health is set in advance. After collecting the parameter variable values ​​of each ecological zone, based on each parameter and according to the parameter variable value, the relationship between the range of the parameter variable value and the parameter health is queried to determine the associated parameter health.

[0059] For example, let's take the parameter land planting rate as an example. When the land planting rate is less than 70%, the health of the associated land planting rate is sub-healthy; when the land planting rate is less than 90% and greater than 70%, the health of the associated land planting rate is medium; when the land planting rate is greater than 90%, the health of the associated land planting rate is healthy. When the land planting rate of the resource loss cloud terminal is collected to be 78%, it can be determined that the health of the land planting rate is medium.

[0060] Based on the variable values ​​of each parameter in the resource loss cloud terminal that has been collected, the above process is completed to determine the health associated with each parameter, and the health s is constructed according to the health associated with each parameter, that is, the queue of the health of all parameters, that is, s={parameter 1 health, parameter 2 health, parameter 3 health, ..., parameter n health}. Exhaust all possible sequences S={s1, s2, ..., s n} as the health set in the GAN table.

[0061] The main part of the deep adversarial sample generation network is the GAN matrix, and the adversarial network method is mainly used to determine the deep adversarial sample generation network GAN matrix. The core idea of ​​the adversarial network algorithm is to generate the optimal strategy for adversarial samples through interaction with the environment.

[0062] Each row in the GAN matrix is ​​a sequence of health, for example: the first row is the health sequence s1, the second row is the health sequence s2, ..., the nth row is the health sequence sn. Among them, health sequence s1 = {sub-healthy, sub-healthy, sub-healthy, ..., sub-healthy}; health sequence s2 = {sub-healthy, healthy, sub-healthy, ..., sub-healthy}; health sequence sn = {healthy, healthy, healthy, ..., healthy}. Enumerate all possible health sequences in the rows of the GAN matrix. Each column in the GAN matrix is ​​the ID of an ecological sensor. Enumerate all the complete ecological sensor queue IDs in the columns of the GAN matrix. For example: A1, A2, A3, ..., A n .

[0063] The GAN values ​​associated with all health sequences and ecological sensor queue id binary pairs are pre-cleaned and repaired to 0, and adversarial samples are generated for the GAN values ​​in the GAN matrix according to the enhanced adversarial sample generation algorithm, and the updated GAN values ​​are output.

[0064] After determining the health sequence S composed of the health of the parameters associated with the ecological zones where resource loss exists in the virtual ecosystem of the cloud platform, a query is performed in the GAN matrix based on the health sequence to find the row where the health sequence S is located, and then the GAN values ​​associated with each ecological sensor queue id in the row are collected; and the available ecological sensor queue id is used for each ecological sensor queue id.

[0065] Specifically, when the eco-sensor queue id needs to be completed, an arbitrary number n is generated. If the arbitrary number n is less than ε, an eco-sensor queue id is arbitrarily selected from each eco-sensor queue id as the available eco-sensor queue id. If the arbitrary number n is greater than or equal to ε, the eco-sensor queue id with the largest GAN value is used as the available eco-sensor queue id. If the GAN values ​​associated with multiple eco-sensor queue ids are the same, an eco-sensor queue id is arbitrarily selected from the eco-sensor queue ids with the same GAN value as the available eco-sensor queue id. Among them, ε is generated according to the discriminant value adversarial sample output in the enhanced adversarial sample generation algorithm.

[0066] A plurality of ecological sensor IDs included in the available ecological sensor queue ID are collected, and ecological sensors associated with the plurality of ecological sensor IDs are used as the available ecological sensor queue.

[0067] The following is an introduction to the enhanced adversarial sample generation algorithm.

[0068] In one possible implementation, the process of outputting the GAN value in the GAN matrix through the enhanced adversarial sample generation algorithm includes: pre-purifying and repairing the GAN value in the GAN matrix; arbitrarily adopting an ecological sensor id based on each health in the row of the GAN matrix; enabling the ecological sensor associated with the ecological sensor id; collecting the ecological environment health and discrimination value of the ecological environment after the ecological sensor is enabled; determining the updated GAN value of the health and ecological sensor id binary based on the ecological environment health, discrimination value and pre-configured GAN value calculation strategy, and returning to complete the process of arbitrarily adopting an ecological sensor id until the loop stop condition is met.

[0069] The embodiment of the present invention is described by taking an example that there are four cases of health degree sequences and four ecological sensor queue IDs.

[0070] In the embodiment of the present invention, the following operations are performed based on each of the above health sequences s1, s2, s3 and s4.

[0071] The eco-sensor queue id that may be completed by the health sequence is adopted. The initial value of ε is configured to be 0.1. Each time the eco-sensor queue id needs to be adopted, an arbitrary number n is generated. If the arbitrary number is less than ε, an eco-sensor queue id is adopted at random. If the arbitrary number n is greater than or equal to ε, the eco-sensor queue id with the largest GAN value is adopted. If the GAN values ​​of multiple eco-sensor queue ids are the same, an eco-sensor queue id is adopted at random from the eco-sensor queue ids with the same GAN value.

[0072] After the ecological sensor queue id is selected according to the above process, the ecological sensors associated with the ecological sensor queue id are collected and completed. After the ecological sensors are completed, the activation health of the ecological environment is collected, and the GAN value associated with the ecological sensor queue id is updated according to the activation health of the ecological environment.

[0073] Specifically, the GAN value calculation strategy is as follows: ; Among them, s is the original health, is the adversarial sample generation rate, is the discriminant value, γ is the overfitting suppression coefficient, New health The maximum GAN value under the condition that after the ecological sensor is enabled, the resource loss is terminated and no new resource loss is obtained, the discriminant value If the resource loss is not stopped after the ecological sensor is enabled, or if new resource loss is obtained after the ecological sensor is enabled, or if the ecological environment performance cannot meet the standard activation conditions, the judgment value Is a negative value.

[0074] S105: Activate each ecological sensor in the available ecological sensor queue to perform shared monitoring of ecological flow data.

[0075] Before the shared monitoring method for ecological flow data provided by the embodiment of the present invention is completed, the ecological sensors that can be completed are compiled in advance, and the queue of the sensors that can be completed for ecological environment management is used as an operation and maintenance action set.

[0076] Each eco-sensor can receive action parameters, which include, for example, the name of the cloud terminal where the action is completed, and the parameter format is json. The eco-sensor is stored in the specified path of the cloud terminal where the configuration memory is located in the form of text. The metadata of the eco-sensor is stored in the configuration memory, mainly including the storage action id, action name, action storage path, and addition time.

[0077] In a possible implementation, the ecological flow data includes ecological flow resource variables; enabling each ecological sensor in the available ecological sensor queue includes: inputting the ecological flow resource variables into each ecological sensor; enabling the ecological sensors one by one according to the positioning position of each ecological sensor in the available ecological sensor queue and the data causal relationship.

[0078] The eco-sensor queue is composed of multiple or one eco-sensors, and the completion order and data causal relationship between the eco-sensors involved in the eco-sensor queue can be indicated through a configuration file in the yml format.

[0079] The ecological flow resource variable is input into the ecological sensor to indicate where each ecological sensor needs to repair the resource loss. For example, if the ecological flow resource variable is "cloud terminal A", cloud terminal A is input into the ecological sensor to indicate that the ecological sensor is completed for cloud terminal A and the resource loss occurring in cloud terminal A is monitored.

[0080] After the ecological flow resource variables are input into each ecological sensor, each ecological sensor can be completed in the main cloud terminal, which is a cloud terminal that completes the shared monitoring method for ecological flow data. Each ecological sensor can also be sent to a cloud terminal where resource loss exists.

[0081] On the basis of the above embodiment, it also includes: collecting resource loss purification and repair results; optimizing the deep adversarial sample generation network using the available ecological sensor queue ID and resource loss purification and repair results.

[0082] Among them, the resource loss purification and repair results may include whether resource loss is monitored, whether new resource loss is obtained, and other automatically obtained purification and repair results, and may also include evaluation information related to the resource loss purification and repair results input by the operation and maintenance personnel, such as: scoring value.

[0083] A preliminary assessment of the ecological environment will be conducted based on whether the ecological flow resources have recovered and the loss of ecological environmental health. The operation and maintenance personnel will add appropriate human assessments based on the actual response of the ecological environment in addition to the preliminary assessment, and feed back the comprehensive assessment results to the deep adversarial sample generation network deployment unit as a reference for the next model cycle.

[0084] At the same time, for each shared monitoring operation on ecological flow data, an audit function for shared monitoring operations on ecological flow data is set up to record the script ID, operation time, operation results, etc. of each operation.

[0085] An embodiment of the present invention provides a shared monitoring method for ecological flow data, the method comprising: when it is identified that there is resource loss in a virtual ecosystem of a cloud platform, collecting ecological flow data; determining the monitoring recovery type associated with the ecological flow data; under the condition that the monitoring recovery type is a first type, determining an available ecological sensor queue in a preset strategy mapping table based on the ecological flow data, wherein the first type refers to a type of resource loss associated with the presence of a resource loss inducement and a pre-configured monitoring strategy; under the condition that the monitoring recovery type is a second type, using a pre-deployed deep adversarial sample generation network to purify and repair the ecological flow data, and determining an available ecological sensor queue, wherein the second type refers to a type of resource loss associated with no resource loss inducement or no pre-configured monitoring strategy; enabling each ecological sensor in the available ecological sensor queue to perform shared monitoring of the ecological flow data.

[0086] The shared monitoring method for ecological flow data provided by the embodiment of the present invention has the following advantages: (1) After collecting ecological flow data, the type of resource loss is determined based on the ecological flow data. If the cause of resource loss exists and there is a pre-configured monitoring strategy, the pre-configured monitoring strategy is used to determine the ecological sensors that need to be completed, and the above ecological sensors are completed to conduct shared monitoring of ecological flow data. If the cause of resource loss does not exist or there is no related monitoring strategy, the ecological flow data is purified and repaired using the enhanced adversarial sample generation algorithm, the ecological sensors that need to be completed are determined, and the above ecological sensors are completed to conduct shared monitoring of ecological flow data. Not only does it rely on pre-configured expert strategies to deal with existing resource loss, but it can also dynamically purify and repair non-existent or complex resource loss situations, significantly improving the ability of the healthy cloud platform virtual ecosystem ecological environment to deal with non-existent resource loss and the flexibility of the cloud platform virtual ecosystem ecological environment to purify and repair resource loss.

[0087] (2) Integrating enhanced adversarial sample generation for decision optimization not only increases the flexibility of purification and restoration of resource loss, but also enables the ecological environment to continuously generate adversarial samples and adapt to new resource loss patterns and solution strategies, continuously improving the efficiency and effectiveness of resource loss purification and restoration over time.

[0088] (3) A complete feedback loop is designed to evaluate the effect of each resource loss response and feed the results back to the deep adversarial sample generation network for the next round of cyclic deployment. This self-evaluation mechanism not only ensures the immediate effect evaluation of resource loss purification and repair actions, but also optimizes the long-term performance of the entire ecosystem through continuous adversarial sample generation and adjustment.

[0089] Embodiment 2: Figure 2FIG. 1 is a schematic diagram of a structure of a shared monitoring device for ecological flow data in an embodiment of the present invention. Figure 2 As shown, the shared monitoring device 20 for ecological flow data provided by the embodiment of the present invention mainly includes: Among them, the ecological flow data collection unit 21 collects ecological flow data when it is identified that there is resource loss in the ecological environment; the monitoring and restoration type judgment unit 22 judges the monitoring and restoration type associated with the ecological flow data; the policy library purification and restoration unit 23, under the condition that the monitoring and restoration type is the first type, determines the available ecological sensor queue in the preset policy mapping table based on the ecological flow data, wherein the first type refers to the type of resource loss associated with the existence of resource loss inducements and a pre-configured monitoring strategy; the enhanced adversarial sample generation purification and restoration unit 24, under the condition that the monitoring and restoration type is the second type, uses the pre-deployed deep adversarial sample generation network to purify and repair the ecological flow data, and determines the available ecological sensor queue, wherein the second type refers to the type of resource loss associated with no resource loss inducements or no pre-configured monitoring strategy; the sensor activation unit 25 activates each ecological sensor in the available ecological sensor queue to perform shared monitoring of ecological flow data.

[0090] The embodiment of the present invention provides a shared monitoring device for ecological flow data, which completes the following process: when it is identified that there is resource loss in the ecological environment, collect ecological flow data; determine the monitoring recovery type associated with the ecological flow data; under the condition that the monitoring recovery type is the first type, determine the available ecological sensor queue in the preset strategy mapping table according to the ecological flow data, wherein the first type refers to the type of resource loss associated with the presence of resource loss inducement and the monitoring strategy configured in advance; under the condition that the monitoring recovery type is the second type, use the deep adversarial sample generation network deployed in advance to purify and repair the ecological flow data, and determine the available ecological sensor queue, and the second type refers to the type of resource loss associated with no resource loss inducement or no monitoring strategy configured in advance; enable each ecological sensor in the available ecological sensor queue to perform shared monitoring for ecological flow data. After the ecological flow data is collected, the resource loss type is determined according to the ecological flow data, and under the condition that the resource loss inducement exists and there is a monitoring strategy configured in advance, the ecological sensor that needs to be completed is determined by using the monitoring strategy configured in advance, and the above ecological sensor is completed to perform shared monitoring for ecological flow data. Under the condition that the cause of resource loss does not exist or is not associated with the monitoring strategy, the ecological flow data is purified and repaired using the enhanced adversarial sample generation algorithm, the ecological sensors that need to be completed are determined, and the above ecological sensors are completed to conduct shared monitoring of ecological flow data. Not only does it rely on pre-configured expert strategies to deal with existing resource loss, but it can also dynamically purify and repair non-existent or complex resource loss situations, significantly improving the ability of the healthy cloud platform virtual ecosystem ecological environment to deal with non-existent resource loss and the flexibility of the cloud platform virtual ecosystem ecological environment to purify and repair resource loss.

[0091] In a possible implementation, the association between the ecological traffic resource id and the sensor queue id is stored in the policy mapping table; the policy library purification and repair unit 23 specifically obtains the ecological traffic resource id from the ecological traffic data; based on the ecological traffic resource id, the association between the ecological traffic resource id and the sensor queue id is queried to determine the associated ecological sensor id; based on the associated ecological sensor id, the associated available ecological sensor queue is determined.

[0092] In a possible implementation, the adversarial sample generation purification and repair unit 24 is strengthened, specifically determining ecological zones with resource loss from ecological flow data; collecting variable values ​​of various parameters of ecological zones with resource loss; based on various parameters, determining the health of the parameters according to the variable values ​​of the parameters; querying in the GAN matrix according to the health of the parameters, outputting the GAN values ​​associated with each ecological sensor id, the behavioral health sequence of the GAN value matrix, the columns of the GAN matrix are each ecological sensor id, and the GAN values ​​associated with the health and ecological sensor id binary in the GAN matrix are output through the adversarial network algorithm; selecting available ecological sensor ids according to the GAN values; and constructing an available ecological sensor queue according to the available ecological sensor ids.

[0093] In a possible implementation, it also includes a deep adversarial sample generation network deployment unit, which outputs the GAN value in the GAN matrix through the enhanced adversarial sample generation algorithm, specifically, pre-purifies and repairs the GAN value in the GAN matrix; based on each health degree in the row of the GAN matrix, arbitrarily adopts an ecological sensor id; enables the ecological sensor associated with the ecological sensor id; collects the ecological environment health degree and discrimination value of the ecological environment after the ecological sensor is enabled; determines the updated GAN value of the health degree and ecological sensor id binary based on the ecological environment health degree, discrimination value and the pre-configured GAN value calculation strategy, and returns to complete the process of arbitrarily adopting an ecological sensor id until the loop stops.

[0094] In one possible implementation, the GAN value calculation strategy is as follows: ; Among them, s is the original health, is the adversarial sample generation rate, is the discriminant value, γ is the overfitting suppression coefficient, New health The maximum GAN value under the condition that after the ecological sensor is enabled, the resource loss is terminated and no new resource loss is obtained, the discriminant value If the resource loss is not stopped after the ecological sensor is enabled, or if new resource loss is obtained after the ecological sensor is enabled, or if the ecological environment performance cannot meet the standard activation conditions, the judgment value Is a negative value.

[0095] In a possible implementation, the ecological flow data includes ecological flow resource variables; the sensor activation unit 25 specifically inputs the ecological flow resource variables into each ecological sensor; and the ecological sensors are activated one by one according to the positioning position of each ecological sensor in the available ecological sensor queue and the data causal relationship.

[0096] In a possible implementation, it also includes: a model optimization unit that collects resource loss purification and repair results; and optimizes the deep adversarial sample generation network using the various ecological sensor IDs and resource loss purification and repair results included in the available ecological sensor queue.

[0097] The shared monitoring device for ecological flow data provided in the embodiment of the present invention can complete the shared monitoring method for ecological flow data provided in any embodiment of the present invention, and has the corresponding functional units and beneficial effects of the completion method.

[0098] In the several embodiments provided by the present invention, it should be understood that the disclosed ecological environment, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or ecological resource elements can be combined or integrated into another ecological environment, or some features can be ignored or not completed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0099] In addition, each functional unit in each embodiment of the present invention may be integrated into a purification and repair unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of a software functional unit. The above is only an implementation method of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of the present invention, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present invention.

[0100] The specific implementation methods of the invention are described in detail above, but they are only examples, and the invention is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the invention, therefore, the equalization, modification, improvement, etc. made without departing from the spirit and principle of the invention should be included in the scope of the invention.

Claims

1. A shared monitoring method for ecological flow data, characterized in that: The method comprises: When resource loss is identified in the ecological environment, collect ecological flow data; Determining the monitoring restoration type associated with the ecological flow data; Under the condition that the monitoring recovery type is the first type, determining the available ecological sensor queue in a preset strategy mapping table according to the ecological flow data, wherein the first type refers to a type associated with resource loss with a pre-configured monitoring strategy and a resource loss cause; Under the condition that the monitoring and restoration type is the second type, the ecological flow data is purified and repaired using a pre-deployed deep adversarial sample generation network to determine an available ecological sensor queue, wherein the second type refers to a type of resource loss associated with no inducement for resource loss or no pre-configured monitoring strategy; Each ecological sensor in the available ecological sensor queue is enabled to perform shared monitoring of ecological flow data.

2. A method for sharing and monitoring ecological flow data according to claim 1, characterized in that: The strategy mapping table stores the association between the ecological traffic resource ID and the sensor queue ID; The determining of the available ecological sensor queue in a preset strategy mapping table according to the ecological flow data includes: Obtaining an ecological flow resource id from the ecological flow data; According to the ecological flow resource ID, the association relationship between the ecological flow resource ID and the sensor queue ID is searched to determine the associated ecological sensor ID; An associated available ecological sensor queue is determined according to the associated ecological sensor ID.

3. A method for sharing and monitoring ecological flow data according to claim 1, characterized in that: The method of using the pre-deployed deep adversarial sample generation network to purify and repair the ecological traffic data and determine the available ecological sensor queue includes: Identify ecological zones where resource loss occurs from the ecological flow data; Collecting variable values ​​of various parameters of the ecological zone where resource loss occurs; Based on each parameter, determining the health of the parameter according to the variable value of the parameter; According to the health of the parameter, the GAN matrix is ​​queried, and the GAN value associated with each ecological sensor ID is output, the GAN matrix is ​​a behavioral health sequence, the columns of the GAN matrix are each ecological sensor ID, and the GAN value associated with the health and ecological sensor ID binary in the GAN matrix is ​​output through an adversarial network algorithm; Selecting an available ecological sensor ID according to the GAN value; An available ecological sensor queue is constructed according to the available ecological sensor IDs.

4. A method for sharing and monitoring ecological flow data as claimed in claim 3, characterized in that: The process of outputting the GAN value in the GAN matrix through the adversarial network algorithm includes: Pre-cleaning and repairing the GAN values ​​in the GAN matrix; Based on each health metric in a row of the GAN matrix, an eco-sensor id is arbitrarily adopted; Enable the eco-sensor associated with the eco-sensor id; Collecting the ecological environment health and discrimination value of the ecological environment after the ecological sensor is activated; The updated GAN value of the health degree and the ecological sensor ID tuple is determined according to the ecological environment health degree, the discrimination value and the pre-configured GAN value calculation strategy, and the process of arbitrarily adopting an ecological sensor ID is returned to complete until the loop stop condition is met.

5. A method for sharing and monitoring ecological flow data as claimed in claim 4, characterized in that: The GAN value calculation strategy is as follows: ； Among them, s is the original health, is the adversarial sample generation rate, is the discriminant value, γ is the overfitting suppression coefficient, New health The maximum GAN value under ; After the ecological sensor is enabled, the resource loss is stopped and no new resource loss is obtained, the judgment value is a positive value. If the resource loss does not stop after the ecological sensor is enabled, or new resource loss is obtained after the ecological sensor is enabled, or the ecological environment performance cannot meet the standard activation conditions, the discrimination value Is a negative value.

6. A method for sharing and monitoring ecological flow data according to any one of claims 1 to 3, characterized in that: The ecological flow data include ecological flow resource variables; The enabling of each ecological sensor in the available ecological sensor queue includes: Inputting the ecological flow resource variables into the respective ecological sensors; The ecological sensors are enabled one by one according to the positioning position and data causal relationship of each ecological sensor in the available ecological sensor queue.

7. A method for sharing and monitoring ecological flow data according to claim 1, characterized in that: Also includes: Collect resource loss purification and repair results; The deep adversarial sample generation network is optimized using the eco-sensor IDs included in the available eco-sensor queue and the resource loss purification and repair results.

8. A shared monitoring device for ecological flow data, characterized in that: The device comprises: The ecological flow data collection unit collects ecological flow data when it identifies resource loss in the ecological environment; A monitoring and restoration type determination unit, for determining a monitoring and restoration type associated with the ecological flow data; A policy library purification and restoration unit, under the condition that the monitoring and restoration type is the first type, determines the available ecological sensor queue in a preset policy mapping table according to the ecological flow data, wherein the first type refers to a type associated with resource loss with a pre-configured monitoring policy and a resource loss cause; Strengthening the adversarial sample generation purification and repair unit, under the condition that the monitoring and recovery type is the second type, using the pre-deployed deep adversarial sample generation network to purify and repair the ecological flow data, and determine the available ecological sensor queue, the second type refers to the type of resource loss associated with no resource loss inducement or no pre-configured monitoring strategy; A sensor enabling unit, which enables each ecological sensor in the available ecological sensor queue to perform shared monitoring of ecological flow data; The shared monitoring device for ecological flow data completes the shared monitoring method for ecological flow data as described in any one of claims 1-7.

Images