Automatic handling and personnel scheduling method for operation and maintenance events of parking lot

Abstract

The invention relates to the technical field of intelligent building and parking lot management, and discloses a parking lot operation and maintenance event automatic disposal and personnel scheduling method, which comprises the following steps: firstly, obtaining operation and maintenance event information in real time through various sensing devices; secondly, a dynamic priority calculation model is constructed, and the model integrates the type severity of the events, the potential influence range based on the equipment dependency graph and the time urgency, and gives quantitative priorities to the events; secondly, vectorizing task requirements of events and multi-dimensional state information such as capabilities, workloads and positions of operation and maintenance personnel, determining an optimal matching personnel by calculating a matching score, and performing a final scheduling decision in combination with a task preemption mechanism; and finally, the system performs closed-loop parameter self-optimization on the priority calculation model and the personnel ability model according to feedback data of task completion. According to the invention, quick response and intelligent accurate scheduling of the operation and maintenance event can be realized, and the resource allocation efficiency is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent building and parking management technology, and in particular to an automatic handling and personnel dispatching method for parking lot operation and maintenance events. Background Technology

[0002] The normal operation of modern large parking lots relies heavily on the numerous electromechanical devices and environmental monitoring facilities deployed within them. The operation and maintenance of these facilities are crucial for ensuring the safe and stable operation of the parking lot.

[0003] In existing operation and maintenance management practices, the handling methods for system warnings or alarms have significant limitations. On the one hand, the system often lacks the ability to quantitatively differentiate the true urgency of different events. Whether it is a serious equipment failure alarm or a general routine reminder, it may be assigned the same response level in the initial stage, or rely entirely on the subjective judgment of the dispatcher's experience. This "one-size-fits-all" or manual judgment model cannot guarantee that truly critical events that may trigger chain reactions can be identified in the first instance and given the highest priority.

[0004] On the other hand, traditional scheduling decision-making processes are also rather crude when assigning personnel. Dispatchers typically assign tasks based solely on single-dimensional information such as the geographical distance of personnel or their simple "idle / busy" status. This approach fails to comprehensively consider the professional skills of maintenance personnel, the availability of necessary specialized tools, and the actual workload, often leading to mismatches between people and tasks. For example, assigning tasks requiring specific skills to personnel who lack those skills results in low processing efficiency and may even necessitate secondary assignments.

[0005] Even worse, the rigidity of existing operation and maintenance systems in resource allocation exposes their inadequacy in handling complex, high-concurrency scenarios. When all experienced core operation and maintenance personnel are performing tasks, if a new, high-priority emergency suddenly occurs, traditional systems lack a flexible dynamic adjustment mechanism. They cannot intelligently identify and interrupt an ongoing, relatively minor task to free up the most critical human resources for a more urgent event, which can directly lead to delays in the best response time in emergency situations. Summary of the Invention

[0006] The purpose of this invention is to provide an automatic handling and personnel scheduling method for parking lot operation and maintenance events. This method solves the problem that existing parking lot operation and maintenance relies on manual inspections or passive responses, resulting in untimely event handling and low efficiency. It also solves the problem that in the personnel scheduling process, it is impossible to combine the dynamic priority of events, the real-time status of personnel, and comprehensive capabilities for optimal matching, which leads to resource mismatch and delays in high-priority events.

[0007] To achieve the above objectives, the present invention provides an automatic handling and personnel dispatching method for parking lot operation and maintenance events, comprising the following steps: S1. Receive parking lot early warning information collected by sensing devices; Various types of sensing devices are deployed within the parking lot, such as AI cameras, smoke sensors, liquid level gauges, and gas concentration sensors. These devices are used to monitor the real-time status of the parking lot's equipment and environmental parameters, including the operating status of mobile fans, sump levels, ambient carbon dioxide concentration, and sprinkler system status, as well as to identify abnormal scenarios such as smoke and fire, water accumulation on the ground, and illegally parked vehicles. When an anomaly is detected, the sensing devices upload structured early warning information to the platform via an IoT interface.

[0008] S2. Based on the maintenance events corresponding to the parking lot warning information, calculate the dynamic priority of the maintenance events; after receiving the warning information, the platform parses it into a specific maintenance event. The system calculates a dynamic priority value representing the urgency and importance of handling this maintenance event based on its inherent attributes.

[0009] S3. Establish capability vectors for operation and maintenance personnel and establish task requirement vectors containing the dynamic priority for the operation and maintenance events. The system establishes a multi-dimensional capability vector for each on-duty operations and maintenance (O&M) personnel to quantify their overall work capabilities. Simultaneously, the system establishes a task requirement vector for O&M events generated in S2 to quantify the conditions required to handle the event, and uses the calculated dynamic priority as one dimension of this vector.

[0010] S4. Calculate the matching score based on the capability vector and the task requirement vector, and determine the optimal matching personnel; The system calculates the capability vectors of all on-duty operations and maintenance personnel against the task requirement vectors of the current pending operations and maintenance events, assigning a matching score to each personnel. The personnel with the highest matching score are selected as the optimal matching personnel to handle the event.

[0011] S5. When the optimal matching personnel is processing the current task, compare the dynamic priority of the new task corresponding to the operation and maintenance event with the dynamic priority of the current task. If the preset preemption conditions are met, send a task preemption instruction to the optimal matching personnel.

[0012] After determining the optimal matching personnel, the system checks the personnel's current work status. If they are currently processing an assigned task, the system compares the dynamic priority of newly generated maintenance events (new tasks) with the dynamic priority of the personnel's current task. When a preset preemption condition is met, the system generates and sends a task preemption command to the personnel's mobile terminal.

[0013] In a specific embodiment, the process of calculating the dynamic priority in step S2 includes: Based on the severity of the event type, the potential scope of impact, and the time urgency of the maintenance event, the dynamic priority is calculated using a preset weighted model. The specific calculation formula can be expressed as follows: P dynamic =w s ·S+w i ·I+w t ·T; Where: P dynamic S represents the final calculated dynamic priority score; S represents the event type severity score, a pre-defined quantitative value based on the event type to characterize its inherent hazard level; I represents the potential impact range score, a quantitative value to characterize the range of equipment or areas that may be affected if the event is not handled promptly; T represents the time urgency score, a quantitative value that increases with the waiting time after the event occurs; w s w i w t These represent the weight coefficients for the three dimensions mentioned above, which are preset values ​​and satisfy w. s +w i +w t =1.

[0014] Preferably, the potential impact range is determined based on a pre-built parking lot equipment and facility dependency topology map. In this map, each device or facility is a node. When a node fails, the potential impact range score I is determined based on the number of downstream nodes affected by the failed node in the topology map or a preset weight.

[0015] In one specific embodiment, the capability vector of the maintenance personnel includes at least the following dimensions: skill proficiency, real-time workload, tool possession status, and real-time geographical location. Skill proficiency can be represented as a skill vector, quantifying the personnel's mastery of different maintenance skills.

[0016] In one specific embodiment, the task requirement vector includes at least the following dimensions: required skills for the task, estimated processing time, a list of necessary tools, and the dynamic priority of the maintenance event. The required skills for the task are also represented as a skill vector, quantifying the skill requirements for processing the event.

[0017] Preferably, the step of calculating the matching score includes: calculating the matching score by weighting and combining four factors—skill matching degree, real-time workload, tool availability, and geographical distance—between the capability vector of the operations and maintenance personnel and the task requirement vector. The specific calculation formula can be expressed as follows: M score =w skill ·F skill +w load ·F load +w tool ·F tool +w dist ·F dist ; Where: M score Represents the final match score; F skill F load F tool F dist These represent the normalized scores for four factors: skill matching, workload, tool availability, and geographical distance. skill w load w tool w dist These represent the weighting coefficients of the four factors, which are preset values.

[0018] In one specific implementation scheme, the factors in the weighted combination are determined as follows: The skill matching degree is calculated by measuring the skill vector V in the capability vector of operations and maintenance personnel. ability With the required skill vector V in the task requirement vector demand The cosine similarity is used to determine this. The calculation formula is: The tool availability level is determined based on whether the tool holding status in the operation and maintenance personnel's capability vector fully meets the list of required tools in the task requirement vector. If it does, the value is 1; otherwise, the value is 0.

[0019] The real-time workload and geographical distance are directly taken as quantified values ​​and normalized so that the score is inversely proportional to the workload and distance.

[0020] Preferably, the preset preemption condition is a mathematical inequality, specifically in the form of: Pnew >P current ·K p ; Where: P new Represents the dynamic priority of new tasks; P current Represents the dynamic priority of the current task; K p This is a preset preemption sensitivity coefficient, with a value greater than or equal to 1, used to avoid frequent task switching due to small differences in task priorities.

[0021] Furthermore, after sending the task preemption instruction, the following steps are also included: The system instructs the optimal matching personnel to pause the current task and return the paused current task to the system's task pool to await rescheduling in the next round.

[0022] The method of the present invention may further include: The system receives feedback information on the processing results submitted via a mobile terminal by the optimally matched personnel after completing the task. Based on this feedback information, the system adjusts the model parameters (e.g., weight coefficients w) used to calculate dynamic priorities. s w i w t The system automatically updates and adjusts the capability vectors (such as skill proficiency) of operations and maintenance personnel.

[0023] In summary, the present invention has at least one of the following beneficial technical effects: 1. This invention achieves a quantitative distinction between the importance and urgency of different events by calculating dynamic priorities for each operational event in real time and comprehensively considering the severity of the event type, the potential scope of impact, and the time urgency. This changes the traditional operational model of using a uniform response level for all alarms or relying on manual judgment, ensuring that the most critical events can be identified and handled first.

[0024] 2. This invention establishes a multi-dimensional capability vector for operations and maintenance personnel, including skills, workload, tools, and location, and calculates a matching score with the task requirement vector of an event to determine the optimal candidate. This approach breaks through the traditional single-dimensional decision-making that relies solely on geographical location or personnel availability, achieving optimal matching between people and tasks. It ensures that the most suitable personnel are assigned to the most needed tasks, avoiding secondary scheduling due to mismatched skills or tools.

[0025] 3. This invention introduces a task preemption mechanism. When a higher-priority maintenance event occurs, the system can compare its priority with the current task being handled by the optimally matched personnel, and interrupt the current task when preset conditions are met. This mechanism ensures that even when all maintenance resources are occupied, newly emerging and more urgent events can receive the most timely response, significantly improving the robustness and emergency response capabilities of the entire maintenance system. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the system architecture of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention; Figure 3 This is a schematic diagram illustrating the dependency relationship of parking equipment and facilities according to the present invention. Detailed Implementation

[0027] See attached document Figure 1 , Figure 1 This is a functional module architecture diagram of a parking lot operation and maintenance event automatic handling and personnel dispatching system according to an embodiment of the present invention. A method provided by the present invention can run on this system. The system is deployed on one or more servers and communicates with external devices via a data network.

[0028] The system's hardware environment includes: multiple sensing devices deployed at the parking lot site, such as AI cameras, smoke sensors, liquid level gauges, and gas concentration sensors; an IoT gateway for aggregating data from the sensing devices and communicating with the server via wireless or wired networks; mobile communication terminals with specific applications installed for maintenance personnel; and a central server for hosting the system software.

[0029] See attached document Figure 1 The system of the present invention can be functionally divided into multiple modules, specifically including: The data acquisition module is configured to receive raw early warning data uploaded by one or more sensing devices via an IoT gateway. The data acquisition module parses, cleans, and standardizes the received heterogeneous data, converting it into structured early warning information containing timestamps, device IDs, location information, alarm types, and alarm values, and then sends it to the event processing engine.

[0030] The event processing engine connects to the output of the data acquisition module. It is configured to receive structured alert information and map it to a pending maintenance event. The core function of this module is to calculate a dynamic priority value for each generated maintenance event based on a preset dynamic priority calculation model. After calculation, the event processing engine outputs the maintenance event object, containing the original alert information and the dynamic priority value, to the intelligent scheduling center.

[0031] The intelligent dispatch center connects its input end to the output end of the event processing engine. As the core module for executing dispatch decisions, the intelligent dispatch center is configured to: receive maintenance event objects; obtain real-time capability vectors of all on-duty maintenance personnel from the personnel and data management module; construct task requirement vectors for the received maintenance events; perform matching score calculations based on the capability vectors and task requirement vectors to determine an optimal matching personnel; execute task preemption logic judgments; and finally generate a dispatch instruction containing the task command and the target personnel ID, and send it to the user interaction module.

[0032] The personnel and data management module is bidirectionally connected to the intelligent dispatch center. This module is configured to: store and manage basic information for each operations and maintenance (O&M) personnel, including their skill matrix; and dynamically construct and update each O&M personnel's capability vector by obtaining their geographical location, tool list, and task load status updated by the intelligent dispatch center in real time through the user interaction module. This module is also configured to receive feedback data after task completion, used to update system model parameters.

[0033] The user interaction module connects to both the intelligent dispatch center and the personnel and data management module, and communicates with the mobile communication terminals of maintenance personnel. The user interaction module is configured to: receive dispatch instructions from the intelligent dispatch center and push them as task notifications to the designated maintenance personnel's mobile communication terminals; receive real-time geographical location, task acceptance / rejection status, and task completion feedback uploaded by maintenance personnel via their mobile communication terminals, and forward this information to the personnel and data management module for processing.

[0034] See attached document Figure 2 , Figure 2 This is a flowchart of a method according to an embodiment of the present invention. Detailed embodiments of the present invention will be described in conjunction with this flowchart.

[0035] The first step in the method of this invention is the perception and generation of events, which corresponds to the attached... Figure 2Step S1. In this step, the system monitors the operating status and environment in real time through various sensing devices deployed in the parking lot. These sensing devices include, but are not limited to: AI cameras for identifying visual anomalies such as smoke, fire, water accumulation on the ground, illegal parking, or congestion; smoke sensors for monitoring smoke concentration in specific areas; level gauges for measuring water levels in sump pits or sewage pits; and air sensors for monitoring ambient temperature, carbon dioxide, or other specific gas concentrations.

[0036] Each sensing device generates a raw warning data point when it detects an abnormal state exceeding a preset threshold. To ensure data standardization and parsability, this raw warning data is constructed into a structured data packet. In a specific embodiment, this data packet can be represented as an information body M containing multiple fields, with the following structure: M = (ID...) dev ,T stamp ,L code E type E val M data ); Among them: ID dev It is a unique identifier for the sensing device; T stamp It is the precise timestamp of when the event occurred; L code It is an encoding that represents the physical location of the sensing device; E type This is the type code for the warning event; for example, code 0x01 represents a smoke alarm, and 0x02 represents a water level exceeding the limit alarm. val It is a quantitative value related to the event; for example, for a water level exceeding the limit alarm, this value is the measured water level height; for a smoke alarm, this value is the measured smoke concentration. data It is a metadata field used to carry additional information, such as the Uniform Resource Locator (URL) of a snapshot taken by an AI camera.

[0037] These structured early warning messages are aggregated through on-site IoT gateways. The IoT gateways use standardized IoT communication protocols, such as Message Queuing Telemetry Transport (MQTT), to publish information to designated topics on a central server. The data acquisition module within the central server acts as a subscriber, receiving these published early warning messages in real time.

[0038] Upon receiving information, the data acquisition module first performs integrity and format verification. If verification passes, the module parses the information, extracting the values ​​of the defined fields. Subsequently, the module encapsulates these field values ​​into a unified, standardized initial object for operational events within the system and sends this object to the event processing engine for subsequent priority analysis and calculation. In this way, the system uniformly converts various heterogeneous alarms from the physical world into standardized data objects that can be processed by subsequent processes.

[0039] See attached document Figure 2 After the event processing engine receives a standardized initial object of an operational event, it executes step S2 to calculate the dynamic priority of the event. This dynamic priority is a quantified value used to characterize the urgency and importance of the event requiring processing. The calculation process is based on a preset weighted model.

[0040] In one specific embodiment, the dynamic priority P dynamic The calculation formula is as follows: P dynamic =w s ·S+w i ·I+w t ·T; Where: P dynamic S represents the final calculated dynamic priority score; S represents the event type severity score, a pre-defined quantitative value based on the event type to characterize its inherent hazard level; I represents the potential impact range score, a quantitative value to characterize the range of equipment or areas that may be affected if the event is not handled promptly; T represents the time urgency score, a quantitative value that increases with the waiting time after the event occurs; w s w i w t These represent the weight coefficients for the three dimensions mentioned above, which are preset values ​​and satisfy w. s +w i +w t =1.

[0041] The event type severity score S is a static score determined based on the event's inherent attributes. The system internally maintains a mapping table between event types and severity scores. When the event handling engine receives an event object, it encodes it according to the event type E. type Query this table to obtain the corresponding severity score S. For example, fire alarm (E type The severity score S for the vehicle parking violation warning (=0x01) is set to 95 (out of 100), while the severity score S for the vehicle parking violation warning (E) is set to 95 (out of 100). type The S value of (=0x05) is set to 20.

[0042] See attached document Figure 3, Figure 3 This is a schematic diagram of the dependency relationships of parking equipment and facilities according to an embodiment of the present invention. The potential impact range score I is calculated based on this pre-constructed equipment and facility dependency diagram. In this diagram, each node represents a piece of equipment or a functional area, and the directed edges between nodes represent their dependencies. Each node n is pre-assigned a weight W. n This weight represents the importance of the node in the entire system. When the device corresponding to a node fails, its potential impact range I is determined by the sum of the weights of the failed node and all its affected downstream nodes. The calculation formula is as follows: Where: n fault It is the node where the failure occurred; N downstream In the dependency graph, by n f The set of all downstream nodes reachable from the ault; W n This is the preset weight of node n. For example, refer to the appendix. Figure 3 If the main drainage pump 301 malfunctions, and its downstream nodes are the main pipeline 302 and all branch pipeline sump pits 303 and 304, then the affected area I is W. 301 +W 302 +W 303 +W 304 However, a malfunction occurred in the sump pit 304 of one of the branch roads, affecting an area I of only W. 304 .

[0043] The time urgency score T is a dynamically changing function over time, with the independent variable being the time elapsed since the event occurred. This score ensures that the priority of an event that has not been processed for a long time will increase over time. In one embodiment, T is calculated using an upper limit growth function, expressed as: Where: T is the calculated time urgency score; Δt is the current system time and the event timestamp T. stamp The difference, i.e., the length of time the event has occurred; T max This is the maximum value of the score, for example, 100; k is a positive constant used to control the growth rate (steepness) of the function curve; t0 is a time constant representing the point in time where the urgency increases the fastest. This function causes the urgency of the event to increase slowly in the early stages, then rapidly increase after time t0, and finally tend towards the upper limit T. max This meets the actual needs of operation and maintenance incident handling.

[0044] After calculating the scores for S, I, and T respectively, the event processing engine substitutes them into P. dynamicThe calculation formula yields the final dynamic priority score. This score is then appended to the operations and maintenance event object, and the entire object is sent to the intelligent dispatch center for subsequent personnel dispatch decisions.

[0045] See attached document Figure 2 After the intelligent dispatch center receives the operation and maintenance event object with dynamic priority, it executes step S3. The goal of this step is to prepare two standardized data structures for subsequent matching calculations: one is a capability vector describing the real-time status of each operation and maintenance personnel, and the other is a task requirement vector describing the processing requirements of the operation and maintenance event.

[0046] First, the system constructs a capability vector V for each operations and maintenance personnel who is currently on duty. ability This vector is a multi-dimensional composite vector used to describe the comprehensive working capabilities and status of operations and maintenance personnel in real time and quantitatively. In a specific embodiment, the structure of this capability vector is defined as follows: V ability =(S vec ,L val ,T vec G coords ); Wherein: S vec This is a skill proficiency vector. The system predefines a list containing n maintenance skills (e.g., high-voltage electrician, water pump repair, network debugging, etc.). This vector is an n-dimensional vector S. vec =(s1,s2,...,s n ), where each component s i This represents the individual's proficiency rating for the i-th skill, with a value that can be an integer between 0 and 10. The data in this vector is stored in the personnel and data management module and can be updated based on the individual's qualification certifications and historical task completion records; L val This is a scalar value representing real-time workload, ranging from 0 to 1. 0 indicates the employee is completely idle, while 1 indicates their workload has reached its limit. This value is calculated by the intelligent dispatch center based on the ratio of the estimated total working hours of the employee's currently assigned tasks to the standard working hours, and is updated in real-time to the personnel and data management module; T vec This is a tool-holding state vector. The system predefines a list containing m key tools (e.g., insulation megohmmeter, drain cleaning machine, etc.). This vector is an m-dimensional binary vector T. vec =(t1,t2,...,t m ), where component t j A value of 1 indicates that the person currently possesses tool j, while 0 indicates that they do not. This status is reported by the operations and maintenance personnel via an application on their mobile terminal; G coordsThis is the real-time geographic location coordinates of the personnel, represented as a latitude-longitude pair (lon, lat). This data is periodically reported by the operations and maintenance personnel through the user interaction module on their mobile terminals.

[0047] At the same time, the intelligent scheduling center constructs a task requirement vector V for the pending operation and maintenance events. demand This vector is used to quantify the resources and conditions required to successfully handle this event. Its structure corresponds to the capability vector, and its specific definition is as follows: V demand =(S ′ vec H est ,T ′ vec ,P dynamic ); Wherein: S ′ vec It is a skill vector required for a task. Its dimension is related to the skill proficiency vector S of the operations and maintenance personnel. vec Both are n-dimensional. It is a binary vector S. ′ vec =(s ′ 1,s ′ 2,...,s ′ n ), where component s i ′ A value of 1 indicates that handling the event requires the i-th skill, while a value of 0 indicates that it does not. This vector is determined by the system based on the event type E. type Retrieved from a pre-defined "event-skill requirement" mapping table; H est This is a scalar value representing the estimated processing time in hours. This value is also determined by the system based on event type E. type and event quantification value E val Query from the preset time estimation database; T ′ vec It is a list vector of required tools. Its dimension corresponds to the tool ownership state vector T of the operations and maintenance personnel. vec They are the same, both being m-dimensional. It is a binary vector T. ′ vec =(t ′ 1,t ′ 2,...,t ′ m ), where component t j ′ A value of 1 indicates that the j-th tool is necessary to process the event, while 0 indicates that it is not necessary. This vector is determined by the system based on the event type E. type The query result is obtained from the event tool requirements mapping table; P dynamicIt is the dynamic priority score of the operation and maintenance event, that is, the result calculated by step S2.

[0048] The latest capability vector V has been generated for all on-duty maintenance personnel. ability It also generated a task requirement vector V for the current pending events. demand The intelligent scheduling center then uses these vectors as input to pass them to the subsequent matching score calculation stage.

[0049] See attached document Figure 2 After completing step S3, which involves constructing the capability vector of operations and maintenance personnel and the demand vector of tasks, the intelligent scheduling center then executes step S4 to calculate the matching score between each on-duty operations and maintenance personnel and the current pending operations and maintenance event, and determines the optimal candidate accordingly.

[0050] The intelligent dispatch center iterates through the list of all on-duty maintenance personnel, and for each person, assigns their capability vector V. ability With the task requirement vector V of the event demand As input, a comprehensive matching score M is calculated using a pre-defined weighted combination model. score The specific mathematical expression of the model is as follows: M score =∑ k∈K w k ·F k =w skill ·F skill +w load ·F load +w tool ·F tool +w dist ·F dist ; Where: M score This is the final calculated total matching score; K is the set of all considered factors, in this embodiment K = {skill, load, tool, dist}; w k ∑w is the weight coefficient of the k-th factor, a preset non-negative constant stored in the system configuration, and ∑w k =1; F k It is the score of the k-th factor after calculation and normalization, and its value range is within the interval.

[0051] Each factor score F k The specific calculation process is as follows: Skill matching score F skill The calculation is determined by assessing the match between the skills of operations and maintenance personnel and the skills required for the task. This calculation uses a cosine similarity algorithm to convert the personnel's skill proficiency vector S... vecSkill vector S′ required for the task vec Consider them as two vectors in a high-dimensional skill space, and calculate the cosine of the angle between them.

[0052] Wherein: F skill It is the calculated skill matching score; S vec It is the n-dimensional skill proficiency vector (S) of the operations and maintenance personnel. vec =(s1,s2,...,s n ));S′ vec It is the skill vector (S′) required for the n-dimensional task of this operation and maintenance event. vec =(s′1,s′2,...,s′) n )); n is the total number of predefined skill items in the system, i.e., the dimension of the vector; s i It is a vector S vec The i-th component represents the proficiency level of the operations and maintenance personnel in the i-th skill; s′ i It is a vector S′ vec The i-th component is a binary value, where 1 represents that the i-th skill is required to handle this event, and 0 represents that it is not required. The formula calculates F. skill The value ranges between 0 and 1. The closer the value is to 1, the better the match between the personnel's skill set and the task requirements.

[0053] Tool availability score F tool The calculation is a deterministic logical judgment. The system compares the list vector T of the tools required for the task one by one. ′ vec For each non-zero element in the vector, is it in the tool-holding state vector T of the operations and maintenance personnel? vec The corresponding item is also non-zero. The score is 1 only when all the necessary tools are available.

[0054] Wherein: F tool It is the calculated tool possession score; t j The tool holding state vector T of the operations and maintenance personnel vec The j-th component; t j ′ The list vector T of the tools required for the task ′ vec The j-th component; m is the total number of predefined tool items in the system, i.e., the dimension of the vector.

[0055] Workload Score F load The calculation aims to make the system favor selecting personnel who are currently less busy. This score is related to the personnel's real-time workload L. val Inversely proportional.

[0056] Fload =1-L val ; Wherein: F load It is the calculated workload score; L val It is a scalar value of the real-time workload of operations and maintenance personnel, and its value is between 0 and 1.

[0057] Geographical distance score F dist The calculation is used to quantify the proximity of personnel to the location of an incident. First, the system calculates the proximity based on the real-time geographic coordinates G of the operations and maintenance personnel. coords =(lon) p ,lat p ) and the coordinates L of the location where the event occurred coords =(lon) e ,lat e The actual physical distance d between the two objects is calculated using a spherical distance calculation formula (such as the Haversine formula). Then, this distance d is normalized to convert it into a score that is inversely proportional to the distance.

[0058] Among them, F dist This is the calculated geographical distance score. d represents the actual physical distance between the operations personnel and the location of the incident, calculated using their respective geographical coordinates G. coords =(lon) p ,lat p ) and L coords =(lon) e ,lat e )Calculated. D max It is a preset maximum scheduling distance used for normalization calculation.

[0059] The intelligent dispatch center calculates M for each on-duty personnel. score Then, the scores of all personnel are sorted in descending order. The person with the highest score is determined as the optimal matching person to handle the maintenance event. If there are ties in scores, a preset secondary sorting rule can be used (e.g., prioritizing the work load). val The final candidate will be determined by the lower-ranking candidate.

[0060] See attached document Figure 2 After the optimal matching personnel are determined in step S4, the intelligent dispatch center 30 will execute step S5. This step is a condition judgment and execution step, used to handle the specific situation where the optimal matching personnel are currently in a task execution state, to ensure that higher priority events can be handled in a timely manner.

[0061] Specifically, the intelligent dispatch center 30 first queries the real-time status of the optimal matching personnel determined in step S4. If the personnel's status is "idle," the system directly generates a dispatch instruction and assigns the task. If the personnel's status is "task in progress," the task preemption mechanism is triggered.

[0062] In this mechanism, the intelligent dispatch center 30 first obtains the task currently being executed by the optimally matched person from the system, and reads the dynamic priority of the current task, denoted as P. current Subsequently, the system assigns a dynamic priority P to the new event. new (Calculated from step S2) and P current A comparison is performed. This comparison is not a direct size comparison, but rather executed through a conditional judgment logic to avoid excessively frequent task interruptions due to minor priority differences. The mathematical expression for this preemption condition judgment is as follows: P new >P current ·(1+C preempt ); Where: P new It is the dynamic priority of newly generated operation and maintenance events; P current It is the dynamic priority of the task currently being performed by the optimally matched personnel; C preempt This is a preset preemption sensitivity coefficient, a non-negative constant stored in the system configuration. This coefficient is used to adjust the ease with which preemption is triggered. For example, when C... preempt When set to 0.2, the preemption condition is only met when the priority of a new event is more than 20% higher than the priority of the current task.

[0063] If the above preemption condition is true, that is, the judgment result is yes, the intelligent dispatch center 30 performs the following sequence of operations: sends a preemptive dispatch instruction to the user interaction module for the optimal matching personnel, which contains new, higher priority operation and maintenance event information.

[0064] In the system background, update the status of the low-priority task that the person was originally performing to "interrupted".

[0065] The interrupted tasks are put back into the task pool of the intelligent scheduling center so that they can be reassigned to other maintenance personnel in subsequent scheduling cycles.

[0066] If the above preemption condition is false, i.e., the result is negative, it means that the priority of the new event is insufficient to interrupt the current task. In this case, the intelligent scheduling center performs the following operations: Abandon assigning new events to the current best match.

[0067] From step S4, based on matching score M score From the list of people sorted in descending order, select the person ranked second as the new candidate.

[0068] For this new candidate, the system repeats the logical judgment in step S5, which involves checking their work status and, if necessary, initiating the preemption mechanism. This process proceeds sequentially along the ranking list until an available person is found, or a person who meets the preemption criteria and is currently performing a lower-priority task.

[0069] See attached document Figure 2 After an operational incident is resolved, the system will execute a closed-loop feedback and self-optimization process. This process aims to use actual task execution data to iteratively update the model parameters within the system, continuously calibrating and optimizing the accuracy of subsequent decisions.

[0070] The process begins with maintenance personnel marking a completed task as "closed" using the user interaction module on their mobile communication terminal. At this point, the system records the actual completion time of the task. Combined with the task dispatch time, the system calculates the total processing time T for the task. actual Simultaneously, a pre-set quality assessment process is triggered. For example, back-end administrators or the system score the completion quality of the task based on certain objective indicators, resulting in a quantified quality score Q. score These feedback data, including T actual and Q score It is encapsulated into a feedback data packet and sent to the personnel and data management module.

[0071] After receiving and accumulating a certain number (e.g., 100) of feedback data packets, the personnel and data management module periodically performs parameter self-optimization calculations. In a specific embodiment, this optimization process includes the following two aspects: First, adjusting the weight coefficients (w) in the dynamic priority calculation model. s w i w t Adjustments are made to the dynamic priority P calculated by the system. dynamic This allows for a more accurate reflection of the true urgency of an event. The system incorporates a "true urgency" reference score, U, assessed post-event by experienced managers. true The goal of system optimization is to minimize P. dynamic with U true The mean square error L between them.

[0072] Where: L is the mean squared error loss value; N is the number of feedback data samples used for this optimization; U true,jP is the true urgency reference score for the j-th sample task; dynamic,j It is the dynamic priority initially calculated by the system for the j-th sample task.

[0073] The system uses gradient descent to update the weights. For weight w... s Its update rules are as follows: Where: w′ s This is the updated weight value; w s η is the current weight value; η is a preset learning rate, a small positive number used to control the step size of each update; S j It is the event type severity score of the j-th sample task; weight w i and w t The system updates the weights according to a similar rule. After all weights have been updated, the system will also perform a normalization process to ensure that their sum is still 1.

[0074] Secondly, it relates to the capability vector V of operations and maintenance personnel. ability The skill proficiency vector S in vec An update will be performed. This update aims to dynamically adjust an employee's skill rating based on their actual work performance. When an operations and maintenance personnel completes a task requiring specific skill k, the system will assign a quality score Q to that task. score To update its corresponding skill components s k The update formula can use the exponential moving average method: s k,new =(1-α)·s k,old +α·(Q score ·C scale ); Where: s k,new This is the updated proficiency score for the k-th skill; s k,old Q is the score before the update; α is an update coefficient (also called the learning rate), which takes a value between 0 and 1 and is used to control the proportion of new feedback data; score This is the quality score for the completion of the task; C scale It is a scaling factor used to scale Q. score The range of values ​​is mapped to the range of skill proficiency scores.

[0075] Through the above feedback and optimization mechanisms, the system of the present invention can form a closed loop of continuous learning, making its priority assessment and personnel scheduling decisions more accurate and efficient over time.

[0076] To further clarify the technical solution of the present invention, the complete process of the present invention will be described below through a specific scenario-based embodiment. This embodiment involves appendices. Figure 1 The system architecture and appendix shown Figure 2 The method flow is shown.

[0077] Scene setting and event generation: At a certain moment, a level gauge in the sump (B2-SPO1) on level B2 of the parking lot detected a water level of 1.2 meters, exceeding the preset threshold of 1.0 meter. The level gauge immediately generated a structured raw warning data and uploaded it to the central server's data acquisition module via an IoT gateway. The data acquisition module parsed the data and generated a standardized maintenance event initialization object, which included an event type code E. type (For example, 0x02 represents water level exceeding the limit), event quantification value E val The object contains the device ID (B2-SPO1) and the timestamp of the occurrence (1.2 meters). The object is then sent to the event handling engine.

[0078] Dynamic priority calculation: After receiving the event object, the event handling engine executes step S2 to calculate its dynamic priority P. dynamic Let the system's preset weight be w. s =0.5,w i =0.3,w t =0.2.

[0079] Event type severity score S: The system queries the preset mapping table and obtains the severity score of the "water level exceeds limit" event as S = 70.

[0080] Potential Impact Scope Score I: System Reference Appendix Figure 3 The equipment dependency graph shown identifies equipment B2-SPO1 as a leaf node, meaning its failure only affects itself. The preset weight of this node is determined by querying. B2-SP01 =30, therefore I = 30.

[0081] Time urgency score T: Since the event has just occurred, assuming that the current time urgency score is T = 15 after function calculation.

[0082] Substitute the above scores into the dynamic priority calculation formula: P dynamic =w s ·S+w i ·I+w t ·T; Where: P dynamicThis represents the final dynamic priority score for the operational event; S represents the event type severity score, which is 70 in this example; I represents the potential impact range score, which is 30 in this example; T represents the event's time urgency score, which is 15 in this example; w s w i w t These are the weighting coefficients for the three dimensions mentioned above, which are 0.5, 0.3, and 0.2 in this example.

[0083] P dynamic =0.5·70+0.3·30+0.2·15=35+9+3=47. The dynamic priority of this maintenance event is determined to be 47.

[0084] Personnel matching and scheduling decisions: After receiving the maintenance event with priority 47, the intelligent scheduling center executes steps S3 and S4. First, the system constructs a task requirement vector V for this event. demand This includes information such as required skills (water pump repair) and required tools (submersible pump). Subsequently, the system obtains the capability vector V of all currently on-duty maintenance personnel. ability And calculate a matching score for each person.

[0085] Assume there are currently two maintenance personnel on duty: Personnel A and Personnel B. Let the weight of the matching score model be w. skill =0.4, w load =0.2, w tool =0.1, w dist =0.3.

[0086] Personnel A: "Water Pump Repair" score is 9 in the skill vector; Workload L val =0.8; Possessing a submersible pump; Physical distance d from the incident location = 2000 meters.

[0087] Personnel B: "Water Pump Repair" score is 6 in the skill vector; Workload L val =0.1; Possessing a submersible pump; Physical distance d from the incident location = 500 meters.

[0088] The system calculates a matching score M for person B. s core: Skill matching score F skill By calculating the cosine similarity between its skill vector and task requirement vector, F is obtained. skill =0.70.

[0089] Workload Score F load :F load =1-L val =1-0.1=0.9.

[0090] Tool availability score F tool Because it possesses the necessary submersible pump, therefore F tool =1.

[0091] Geographical distance score F dist Let the maximum scheduling distance D be... max = 5000 meters, then F dist =1 - (500 / 5000) = 0.9 Substitute the above scores into the matching score calculation formula: M score =w skill ·F skill +w load ·F load +w tool ·F tool +w dist ·F dist ; Where: M score This is the final matching total score for person B; F skill F load F tool F dist These represent the scores for person B on each dimension, which in this example are 0.70, 0.9, 1, and 0.9. skill ,w load ,w tool ,w dist These are the weighting coefficients for each dimension, which are 0.4, 0.2, 0.1, and 0.3 in this example.

[0092] The score for person B is calculated to be M. score,B =0.4·0.70+0.2·0.9+0.1·1+0.3·0.9=0.28+0.18+0.1+0.27=0.83. Calculate person A's score M in the same way. score,A =0.70. Since 0.83 > 0.70, the system determines that person B is the optimal match.

[0093] Task assignment and closed-loop: The system executes step S5 to query the status of personnel B. This is because of their workload L. val =0.1, the status is "idle", and the task preemption condition is not met. Therefore, the intelligent dispatch center directly generates a dispatch instruction and dispatches the task containing event details and location to personnel B's mobile communication terminal through the user interaction module.

[0094] Personnel B received the instruction, arrived at the site, completed the drainage task, updated the task status to "closed" via mobile terminal, and submitted a processing result report. The system recorded the actual total time T for this task. actual The complete task record is stored in the database as one of the data samples for subsequent execution of the system's closed-loop feedback and self-optimization process.

Claims

1. A method for automatically handling parking lot operation and maintenance events and dispatching personnel, characterized in that, Includes the following steps: S1. Receive parking lot early warning information collected by sensing devices; S2. Calculate the dynamic priority of the operation and maintenance event based on the operation and maintenance event corresponding to the parking lot early warning information; S3. Establish capability vectors for operation and maintenance personnel and establish task requirement vectors containing the dynamic priority for the operation and maintenance events. S4. Calculate the matching score based on the capability vector and the task requirement vector, and determine the optimal matching personnel; S5. When the optimal matching personnel is processing the current task, compare the dynamic priority of the new task corresponding to the operation and maintenance event with the dynamic priority of the current task. If the preset preemption conditions are met, send a task preemption instruction to the optimal matching personnel.

2. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, Step S2 includes: Based on the severity, potential impact, and time urgency of the operation and maintenance events, the dynamic priority is calculated using a preset weighted model.

3. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 2, characterized in that, The potential impact range is determined based on the number or weight of downstream nodes affected by the faulty node corresponding to the maintenance event in a pre-built equipment and facility dependency graph.

4. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, The capability vector of the operations and maintenance personnel includes at least: skill proficiency, real-time workload, tool possession status, and real-time geographical location.

5. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, The task requirement vector includes at least: the skills required for the task, the estimated processing time, the list of necessary tools, and the dynamic priority of the operation and maintenance events.

6. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, The steps for calculating the matching score include: The matching score is calculated by weighting the skill matching degree, real-time workload, tool availability and geographical distance in the capability vector of the operation and maintenance personnel and the task requirement vector.

7. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 6, characterized in that, The factors in the weighted combination are determined as follows: The skill matching degree is determined by calculating the cosine similarity between the skill vector in the capability vector and the required skill vector in the task requirement vector; The tool availability is determined based on whether the tool holding status in the capability vector meets the list of required tools in the task requirement vector. The real-time workload and geographical distance are directly taken as quantified values.

8. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, The preset preemption conditions are: The dynamic priority of the new task is greater than the product of the dynamic priority of the current task and the preset preemption sensitivity coefficient.

9. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 8, characterized in that, After sending the task preemption instruction, the method further includes: The instruction specifies that the optimal matching person should suspend the current task and return the suspended current task to the scheduling pool for rescheduling.

10. The automatic handling and personnel dispatching method for parking lot operation and maintenance events according to claim 1, characterized in that, Also includes: Receive the processing result feedback submitted by the optimally matched personnel after completing the task; Based on the feedback from the processing results, update the model parameters used to calculate the dynamic priority or update the capability vector of the operations and maintenance personnel.

Images