Distributed machine room fire fighting equipment remote centralized management method, system and equipment

Abstract

The invention provides a distributed machine room fire fighting equipment remote centralized management method, system and equipment, and the method comprises the steps: periodically obtaining environment parameters from all sensors through a monitoring agent, and transmitting the environment parameters to a decision agent; performing preliminary analysis on the environmental parameters through a decision agent, and judging whether a potential fire risk exists or not by using a machine learning model; when determining that the fire risk exists, making a resource scheduling plan by using the coordination agent based on the existing resource distribution and the distributed machine room position, and sending the resource scheduling plan to the decision agent; a generated resource scheduling plan is converted into a corresponding operation command through a decision agent, the operation command is transmitted to a corresponding execution agent through a network, fire extinguishing is executed, potential risks are identified through a machine learning model, and through mutual cooperation of different agents, the overall cooperative fire fighting capability is improved, and the fire fighting efficiency is improved. And the management efficiency of the distributed machine room fire fighting equipment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and particularly to a method, system and device for remote centralized management of distributed computer room fire-fighting equipment. Background Art

[0002] With the rapid development of information technology, the number of weak current computer rooms in large organizations such as universities and enterprises has increased rapidly. These computer rooms are usually distributed in different buildings and campuses, and each computer room is equipped with various key devices. To ensure the safe operation of these devices, especially to prevent fire risks, traditional fire management systems face many challenges: Limitations of centralized management systems: Traditional systems rely on a single central controller to collect data, make decisions and issue instructions. This architecture is unable to cope with large-scale distributed environments, and is prone to problems such as single-point failures and response delays. Serious information island phenomenon: The fire management systems between different campuses or departments often operate independently, lacking effective data sharing and collaborative working mechanisms, resulting in waste of resources and low emergency response efficiency. Insufficient adaptability in complex environments: Factors such as the building layout and personnel flow on campus change over time, and static planning methods are difficult to adapt to this dynamic environment, which may lead to unreasonable resource scheduling or low emergency response efficiency. Unstable network communication: Due to problems with the network infrastructure, there may be data transmission delays or communication interruptions, affecting information synchronization and coordination between agents.

[0003] Therefore, how to improve the centralized management efficiency of distributed computer room fire-fighting equipment has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0004] The present invention provides a method, system and device for remote centralized management of distributed computer room fire-fighting equipment to solve the defect of low management efficiency of distributed computer room fire-fighting equipment in the prior art.

[0005] In a first aspect, the present invention provides a method for remote centralized management of distributed computer room fire-fighting equipment, including: Periodically obtaining environmental parameters from each sensor through a monitoring agent, and sending the environmental parameters to a decision-making agent; Preliminarily analyzing the environmental parameters through a decision-making agent, and using a machine learning model to judge whether there is a potential fire risk; When it is determined that there is a fire risk, based on the existing resource distribution and the location of the distributed computer room, using a coordination agent to formulate a resource scheduling plan and send it to the decision-making agent; Converting the generated resource scheduling plan into corresponding operation commands through a decision-making agent, and transmitting them to the corresponding execution agent through the network to perform fire extinguishing.

[0006] A remote centralized management method for distributed computer room fire-fighting equipment provided by the present invention, the machine learning model includes: a first module, a second module, a first output layer, and a second output layer; Among them, the first module is used to input the environmental parameters and output the first hidden feature to the second module; Each of the decision agents includes a second module, a first output layer, and a second output layer, and the coordination agent configures the second module.

[0007] A remote centralized management method for distributed computer room fire-fighting equipment provided by the present invention, the formula of the l-th layer of the first module is as follows: ; Among them, represents the object recognition feature of the v-th object in the l-th layer, represents the object recognition feature of the u-th object in the l-th layer, represents the set of objects that have object associations with objects v and u respectively, and represent the cardinality of the set, represents the plane recognition weight matrix of the l-th layer of the first data structure recognition layer, , E represents the total number of layers, when , , represents the feature associated with the u-th object, =E when is equal to the first hidden feature of object v, is the sigmoid function.

[0008] A remote centralized management method for distributed computer room fire-fighting equipment provided by the present invention, the formula of the second module is as follows: ; ; ; ; Among them, and represent the activation vectors of the reset gate and the update gate at the t-th step respectively; represents the candidate state generated at the t-th step; represents the second hidden feature at the t-th step; represent the first, second, third, fourth, fifth, and sixth transformation matrices (trainable parameters) respectively; Denote the 1st, 2nd, and 3rd deviations (trainable parameters); \(r\in G\), where \(G\) represents the set of all objects, \(t\in\{1, 2, 3, \ldots, n\}\), when \(t = 1\) ; Denote the data of the computer room at time \(t\); Denote the Sigmoid activation function; Denote the tanh activation function.

[0009] According to a method for remote centralized management of distributed computer room fire-fighting equipment provided by the present invention, the operation of the second module is divided into a first stage and a second stage; When running the first stage, only the data of the computer room where the decision agent is located is input. When running the second stage, the data of other computer rooms is introduced as input according to the output result of the second output layer during the operation of the first stage. When running the second stage, the output of the first output layer is used as the result of determining whether there is a potential fire risk this time; When running the first stage, it is the data of the computer room where the decision agent is located. When running the second stage, it is the sum of the data of the computer room where the decision agent is located and the data of other computer rooms.

[0010] According to a method for remote centralized management of distributed computer room fire-fighting equipment provided by the present invention, the expression of the first output layer is: ; where, Denote the fully connected layer, Denote the second hidden feature at the \(n\)th step, \(n\) is the length of the time series, Denote the first output vector, and its \(i\)th component represents the probability of the \(i\)th fire event occurring. If the probability value is greater than 0.5, it means that this type of fire event occurs. The fire event includes no event.

[0011] According to a method for remote centralized management of distributed computer room fire-fighting equipment provided by the present invention, the expression of the second output layer is: ; where, Denote the concatenation function, Denote the fully connected layer, Denote the first hidden feature of object \(v\), Denote the set of all objects, Denote the second hidden feature at the \(n\)th step, \(n\) is the length of the time series, Denote the second output vector, and its \(i\)th component represents the probability that the data of the \(i\)th computer room needs to be introduced into the second stage. If the probability value is greater than 0.5, it means that introduction is required.

[0012] A remote centralized management method for distributed computer room fire-fighting equipment provided by the present invention, the coordination agent further includes a third output layer, and the expression of the third output layer is: ; wherein, represents the fully connected layer, represents the output matrix, and the element in the i-th row and c-th column of the output matrix represents the probability value of the i-th fire-fighting resource called by the i-th computer room. If the probability value is greater than 0.5, it means that the fire-fighting resource needs to be called.

[0013] In a second aspect, the present invention further provides a remote centralized management system for distributed computer room fire-fighting equipment, including: A monitoring agent module, configured to periodically obtain environmental parameters from each sensor through the monitoring agent and send the environmental parameters to the decision-making agent; A decision-making agent module, configured to preliminarily analyze the environmental parameters through the decision-making agent and use a machine learning model to determine whether there is a potential fire risk; A coordination agent module, configured to, when it is determined that there is a fire risk, formulate a resource scheduling plan based on the existing resource distribution and the location of the distributed computer room by using the coordination agent and send it to the decision-making agent; An execution agent module, configured to convert the generated resource scheduling plan into corresponding operation commands through the decision-making agent and transmit them to the corresponding execution agent through the network to perform fire extinguishing.

[0014] In a third aspect, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the remote centralized management method for distributed computer room fire-fighting equipment as described in any one of the above.

[0015] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the remote centralized management method for distributed computer room fire-fighting equipment as described in any one of the above.

[0016] In a fifth aspect, the present invention further provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the remote centralized management method for distributed computer room fire-fighting equipment as described in any one of the above.

[0017] A remote centralized management method, system and device for distributed computer room fire-fighting equipment provided by the present invention obtain environmental parameters from various sensors periodically through a monitoring agent, and send the environmental parameters to a decision-making agent; the decision-making agent conducts a preliminary analysis on the environmental parameters and uses a machine learning model to judge whether there is a potential fire risk; when it is determined that there is a fire risk, based on the existing resource distribution and the location of the distributed computer room, a resource scheduling plan is formulated by a coordination agent and sent to the decision-making agent; the decision-making agent converts the generated resource scheduling plan into corresponding operation commands and transmits them to the corresponding execution agent through the network to execute fire extinguishing. By identifying potential risks through the machine learning model and through the mutual cooperation between different agents, the overall collaborative fire-fighting ability is improved, and the management efficiency of the distributed computer room fire-fighting equipment is enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 is a schematic flowchart of the remote centralized management method for distributed computer room fire-fighting equipment provided in this embodiment; Figure 2 is a schematic structural diagram of the remote centralized management system for distributed computer room fire-fighting equipment provided in this embodiment; Figure 3 is a schematic structural diagram of the electronic device provided in this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions in the present invention in conjunction with the drawings in the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0021] Figure 1 is a schematic flowchart of the remote centralized management method for distributed computer room fire-fighting equipment provided in this embodiment.

[0022] Such as Figure 1As shown in the figure, the remote centralized management method for distributed computer room fire protection equipment provided by the embodiments of the present invention may have a multi-agent system as the execution subject. Specifically, there are 4 executors, including Executor 1, i.e., the monitoring agent, Executor 2, i.e., the decision-making agent, Executor 3, i.e., the execution agent, and Executor 4, i.e., the coordination agent.

[0023] The role of the multi-agent system (MAS) in distributed computer room fire protection management is mainly reflected in its ability to simulate and implement complex, self-organizing collaborative behaviors, thereby improving the system's response speed, flexibility, and robustness.

[0024] Executor 1: Monitoring Agents - There is one or more monitoring agents in each computer room responsible for collecting and preprocessing the sensor data of the local computer room. These agents will send the processed information to the decision-making agent.

[0025] Executor 2: Decision-making Agents - Decision-making agents not only analyze local data but also exchange information with decision-making agents in neighboring areas to jointly determine the best course of action. They use machine learning models or thresholds to judge whether there is a potential fire risk.

[0026] Executor 3: Execution Agents - Execution agents receive instructions from decision-making agents and execute specific fire protection measures, such as activating the sprinkler system or notifying relevant personnel. They can also cooperate with other execution agents. For example, when the fire extinguishers in one computer room are insufficient, they can request support from adjacent computer rooms.

[0027] Executor 4: Coordination Agents - Coordination agents ensure the collaborative work between different agents and maintain the orderly operation of the entire system. They are responsible for cross-computer room communication and resource scheduling to ensure the consistency and timeliness of information.

[0028] The method mainly includes the following steps: 101. Periodically obtain environmental parameters from each sensor through the monitoring agent and send the environmental parameters to the decision-making agent.

[0029] In a specific implementation process, data collection is first carried out. The collected data includes: real-time environmental data from each computer room, such as temperature, smoke concentration, flame monitoring, etc. Historical maintenance records and alarm logs, which are used as the basis for risk assessment and prediction. The location and status (availability) of currently available fire protection resources. Maps and building layout diagrams to help plan the shortest path and the optimal evacuation route. And the MAS can control the monitoring agents to communicate with each other and share the data they monitor respectively, so as to understand the overall situation more comprehensively.

[0030] 102. The decision-making agent conducts a preliminary analysis of the environmental parameters and uses a machine learning model to determine whether there is a potential fire risk.

[0031] MAS is used to control the decision-making agent. By communicating with neighboring agents, it can obtain more context information and thus make more accurate judgments. If an agent believes that a certain computer room poses a high risk, it can request additional monitoring data from nearby agents or directly suggest increasing the patrol frequency in that area.

[0032] The decision-making agent can analyze the environmental parameters. By analyzing the environmental parameters through a machine learning model and then in the way of big data, it determines whether there is a fire risk. The specific method can be to input the environmental parameter data, output the fire risk index, and determine the magnitude of the fire risk according to the size of the fire index, so as to achieve data analysis and early warning.

[0033] 103. When it is determined that there is a fire risk, based on the existing resource distribution and the location of the distributed computer rooms, the coordination agent formulates a resource scheduling plan and sends it to the decision-making agent.

[0034] Specifically, MAS can control the decision-making agents to optimize resource allocation through negotiation. For example, if a computer room lacks fire extinguishers, nearby decision-making agents may actively provide support to ensure the optimal configuration of resources, thus realizing resource scheduling control. As for how many fire extinguishers should be specifically called or from which computer room to retrieve them, it can be determined according to the magnitude of the fire risk index and the fire situation to ensure that the scheduling and fire extinguishing can be completed quickly and in a timely manner.

[0035] 104. The decision-making agent converts the generated resource scheduling plan into corresponding operation commands and transmits them through the network to the corresponding execution agents to perform fire extinguishing.

[0036] After obtaining the resource scheduling plan, the commands are issued and executed. After receiving the commands, the execution agents immediately start the relevant equipment or contact the on-site personnel to perform the fire extinguishing tasks. MAS can control the execution agents to cooperate with each other to ensure the smooth completion of the tasks. For example, if an execution agent cannot complete a certain task independently, it can request the help of other execution agents to form a temporary cooperation group.

[0037] By adopting the method of the present invention, due to the existence of a multi-agent system for overall planning, the following effects can be achieved: Effect 1: Distributed decision-making - MAS allows each agent to make a preliminary judgment based on local information and share this information with other agents to reach a consensus or a joint decision. This reduces the dependence on a single central controller and improves the speed and efficiency of decision-making.

[0038] Function 2: Adaptation and Learning - Agents can adjust their behavior patterns based on past experiences and newly acquired data, enabling the system to continuously optimize its performance over time.

[0039] Function 3: Fault Tolerance - Even if some agents fail, other agents can still continue to work, maintaining the normal operation of the system and enhancing its reliability.

[0040] Function 4: Task Allocation and Collaboration - Agents can allocate tasks through negotiation to ensure that the most suitable agent executes the most appropriate task. At the same time, they can also cooperate to complete complex tasks.

[0041] MAS not only achieves comprehensive coverage of distributed computer room fire management but also significantly improves the efficiency and quality of emergency management. The key to MAS lies in its ability to support distributed decision-making, adaptive learning, fault tolerance, and task collaboration, which are advantages that traditional centralized management systems are difficult to match.

[0042] Furthermore, the machine learning model in this embodiment includes: a first module, a second module, a first output layer, and a second output layer; wherein, the first module is used to input environmental parameters and output first hidden features to the second module; each decision agent includes a second module, a first output layer, and a second output layer, and the coordination agent configures the second module.

[0043] Among them, the layer of the first module has the calculation formula as (1): (1); Among them, represents the object recognition feature of the v-th object in the layer, represents the object recognition feature of the u-th object in the layer, and respectively represent the sets of objects that have object connections with objects v and u, represents the cardinality of the set, represents the planar recognition weight matrix of the l-th layer of the first data structure recognition layer, , E represents the total number of layers, when , represents the feature associated with the u-th object, =E when is equal to the first hidden feature of object v, is the sigmoid function.

[0044] Among them, the operation of the second module is divided into a first stage and a second stage; when running the first stage, only the data of the computer room where the decision agent is located is input. When running the second stage, the data of other computer rooms is introduced as input according to the output result of the second output layer during the operation of the first stage. When running the second stage, the output of the first output layer is used as the result of determining whether there is a potential fire risk; when running the first stage, it is the data of the computer room where the decision agent is located, and when running the second stage, it is the sum of the data of the computer room where the decision agent is located and the data of other computer rooms.

[0045] The calculation formulas of the second module are as shown in (2), (3), (4), and (5): (2); (3); (4); (5); Among them, and respectively represent the activation vectors of the reset gate and the update gate at the t-th step; represents the candidate state generated at the t-th step; represents the second hidden feature at the t-th step; respectively represent the 1st, 2nd, 3rd, 4th, 5th, and 6th transformation matrices (trainable parameters); represents the 1st, 2nd, and 3rd biases (trainable parameters); r ∈ G, where G represents the set of all objects, t ∈ {1, 2, 3,..., n}, and when t = 1 ; represents the data of the computer room at time t. During the first stage of operation, it is the data of the computer room where the decision agent is located, and during the second operation, it is the data of the computer room where the decision agent is located and the data of other computer rooms; represents the Sigmoid activation function; represents the tanh activation function.

[0046] And the expression of the first output layer is (6): (6); Among them, represents the fully connected layer, represents the second hidden feature at the n-th step, where n is the length of the time series, represents the first output vector, and its i-th component represents the probability of the i-th fire event occurring. If the probability value is greater than 0.5, it means that this type of fire event occurs, and the fire events include no event.

[0047] The expression of the second output layer is (7): (7); Among them, represents a splicing function, represents a fully connected layer, represents the first hidden feature of object v, represents the set of all objects, represents the second hidden feature at the nth step, where n is the length of the time series, represents the second output vector, and the i-th component of it represents the probability that the data of the i-th computer room needs to be introduced into the second stage. If the probability value is greater than 0.5, it means that it needs to be introduced.

[0048] The coordination agent also includes a third output layer, and the expression of the third output layer is (8): (8); Among them, represents a fully connected layer, represents the output matrix, and the element in the i-th row and c-th column of the output matrix represents the probability value of the i-th fire protection resource called by the i-th computer room. If the probability value is greater than 0.5, it means that the fire protection resource needs to be called.

[0049] During the execution process, the execution agent continuously monitors the task progress and reports the latest situation to the decision-making agent. The decision-making agent dynamically adjusts the strategy according to the feedback information to ensure the effectiveness and flexibility of the response measures. The decision-making agent can re-evaluate the situation based on the real-time feedback and discuss new action plans with other agents. This dynamic adjustment ability enables the system to flexibly respond to changing situations.

[0050] After each event ends, all relevant data and operation records will be saved as the basis for subsequent analysis and improvement. The knowledge base of the decision-making agent is updated using the reinforcement learning algorithm, enabling the system to make better choices in future similar situations. Through self-learning and experience accumulation, the agent can continuously improve its decision-making ability and the ability to handle complex situations. In addition, the agents can also share the learning results to promote the evolution of the entire system.

[0051] Based on the same general inventive concept, the present invention also protects a distributed computer room fire protection equipment remote centralized management system. The distributed computer room fire protection equipment remote centralized management system described below can be mutually corresponding and referred to with the distributed computer room fire protection equipment remote centralized management method described above.

[0052] Figure 2 is the structural schematic diagram of the distributed computer room fire protection equipment remote centralized management system provided in this embodiment.

[0053] As Figure 2 shown, a distributed computer room fire protection equipment remote centralized management system provided in this embodiment includes: The monitoring agent module 201 is used to periodically obtain environmental parameters from each sensor through the monitoring agent and send the environmental parameters to the decision-making agent; The decision-making agent module 202 is used to preliminarily analyze the environmental parameters through the decision-making agent and use a machine learning model to determine whether there is a potential fire risk; The coordination agent module 203 is used to, when it is determined that there is a fire risk, formulate a resource scheduling plan based on the existing resource distribution and the location of the distributed computer room by using the coordination agent and send it to the decision-making agent; The execution agent module 204 is used to convert the generated resource scheduling plan into corresponding operation commands through the decision-making agent and transmit them to the corresponding execution agent through the network to perform fire extinguishing.

[0054] Figure 3 It is a schematic structural diagram of the electronic device provided in this embodiment.

[0055] As Figure 3 shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete mutual communication through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute the method for remote centralized management of fire-fighting equipment in a distributed computer room. The method includes: periodically obtaining environmental parameters from each sensor through the monitoring agent and sending the environmental parameters to the decision-making agent; preliminarily analyzing the environmental parameters through the decision-making agent and using a machine learning model to determine whether there is a potential fire risk; when it is determined that there is a fire risk, formulating a resource scheduling plan based on the existing resource distribution and the location of the distributed computer room by using the coordination agent and sending it to the decision-making agent; converting the generated resource scheduling plan into corresponding operation commands through the decision-making agent and transmitting them to the corresponding execution agent through the network to perform fire extinguishing.

[0056] In addition, when the logical instructions in the above-mentioned memory 330 can be implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0057] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the remote centralized management method for distributed computer room fire-fighting equipment provided by the above-mentioned various methods. The method includes: periodically obtaining environmental parameters from each sensor through a monitoring agent and sending the environmental parameters to a decision-making agent; preliminarily analyzing the environmental parameters through the decision-making agent and using a machine learning model to determine whether there is a potential fire risk; when it is determined that there is a fire risk, based on the existing resource distribution and the location of the distributed computer room, using a coordination agent to formulate a resource scheduling plan and send it to the decision-making agent; converting the generated resource scheduling plan into corresponding operation commands through the decision-making agent and transmitting them to the corresponding execution agent through the network to execute fire extinguishing.

[0058] In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is implemented to execute the remote centralized management method for distributed computer room fire-fighting equipment provided by the above-mentioned various methods. The method includes: periodically obtaining environmental parameters from each sensor through a monitoring agent and sending the environmental parameters to a decision-making agent; preliminarily analyzing the environmental parameters through the decision-making agent and using a machine learning model to determine whether there is a potential fire risk; when it is determined that there is a fire risk, based on the existing resource distribution and the location of the distributed computer room, using a coordination agent to formulate a resource scheduling plan and send it to the decision-making agent; converting the generated resource scheduling plan into corresponding operation commands through the decision-making agent and transmitting them to the corresponding execution agent through the network to execute fire extinguishing.

[0059] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0060] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A remote centralized management method for distributed computer room fire protection equipment, characterized in that: include: The monitoring agent periodically obtains environmental parameters from each sensor and sends the environmental parameters to the decision agent; Performing a preliminary analysis of the environmental parameters through a decision-making agent and using a machine learning model to determine whether there is a potential fire risk; When it is determined that there is a fire risk, a resource scheduling plan is formulated using the coordination agent based on the existing resource distribution and the location of the distributed computer rooms, and sent to the decision agent; The generated resource scheduling plan is converted into corresponding operation commands through the decision agent, and transmitted to the corresponding execution agent through the network to execute fire fighting.

2. The remote centralized management method for distributed computer room fire fighting equipment according to claim 1 is characterized in that: The machine learning model includes: a first module, a second module, a first output layer and a second output layer; Wherein, the first module is used to input the environmental parameters and output the first hidden features to the second module; Each of the decision agents includes a second module, a first output layer and a second output layer, and the coordination agent configures the second module.

3. The remote centralized management method for distributed computer room fire fighting equipment according to claim 2 is characterized in that: The first module The calculation formula of the layer is as follows: ； in, Indicates The object recognition features of the v-th object in the layer, Indicates The object recognition features of the u-th object in the layer, and Respectively represent the set of objects that have object relationships with object v and object u, represents the cardinality of a set, represents the plane recognition weight matrix of the first data structure recognition layer, , E represents the total number of layers, hour , represents the features associated with the u-th object, =E is equal to the first hidden feature of object v, is the sigmoid function.

4. The remote centralized management method for distributed computer room fire fighting equipment according to claim 2 is characterized in that: The calculation formula for the second module is as follows: ； ； ； ； in, and Represent the activation vectors of the reset gate and update gate of the tth step respectively; represents the candidate state generated in step t; represents the second hidden feature of the tth step; Respectively represent the 1st, 2nd, 3rd, 4th, 5th, and 6th transformation matrices (trainable parameters); represents the 1st, 2nd, and 3rd deviations (trainable parameters); r∈G, G represents the set of all objects, t∈{1, 2, 3, …, n}, when t=1 ; Represents the data of the computer room at time t; Represents the Sigmoid activation function; Represents the tanh activation function.

5. The method for remote centralized management of distributed computer room fire fighting equipment according to claim 4 is characterized in that: The operation of the second module is divided into the first stage and the second stage; When running the first stage, only the data of the computer room where the decision agent is located is input. When running the second stage, the data of other computer rooms are introduced as input according to the output result of the second output layer when running the first stage. When running the second stage, the output of the first output layer is used as the result of judging whether there is a potential fire risk. When the first stage is run, the data is from the computer room where the decision-making agent is located, and when the second stage is run, the data is the sum of the data from the computer room where the decision-making agent is located and the data from other computer rooms.

6. The method for remote centralized management of distributed computer room fire fighting equipment according to claim 4, characterized in that: The expression of the first output layer is: ； in, represents the fully connected layer, represents the second hidden feature of the nth step, where n is the length of the time series, Represents the first output vector, whose i-th component represents the probability of the i-th fire event occurring. If the probability value is greater than 0.5, it means that a fire event of this type has occurred, and the fire event includes no event.

7. The method for remote centralized management of distributed computer room fire fighting equipment according to claim 6, characterized in that: The expression of the second output layer is: ； in, represents the concatenation function, represents the fully connected layer, represents the first hidden feature of object v, Represents the collection of all objects, represents the second hidden feature of the nth step, where n is the length of the time series, Represents the second output vector, and its i-th component represents the probability that the data of the i-th computer room needs to be introduced into the second stage. If the probability value is greater than 0.5, it means that it needs to be introduced.

8. The method for remote centralized management of distributed computer room fire fighting equipment according to claim 7, characterized in that: The coordination agent also includes a third output layer, and the expression of the third output layer is: ； in, represents the fully connected layer, It represents an output matrix, and the element in the i-th row and c-th column of the output matrix represents the probability value of the i-th fire-fighting resource called by the i-th computer room. If the probability value is greater than 0.5, it means that the fire-fighting resource needs to be called.

9. A remote centralized management system for distributed computer room fire fighting equipment, characterized in that: include: A monitoring agent module, used for periodically acquiring environmental parameters from various sensors through the monitoring agent, and sending the environmental parameters to the decision agent; A decision agent module, used for performing a preliminary analysis of the environmental parameters through a decision agent, and using a machine learning model to determine whether there is a potential fire risk; A coordination agent module, for, when it is determined that there is a fire risk, formulating a resource scheduling plan using the coordination agent based on the existing resource distribution and the location of the distributed computer room, and sending the plan to the decision agent; The execution agent module is used to convert the generated resource scheduling plan into corresponding operation commands through the decision agent, and transmit it to the corresponding execution agent through the network to execute fire fighting.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the remote centralized management method for distributed computer room fire protection equipment as described in any one of claims 1 to 8 is implemented.

Images