Elevator updating control method and system based on AI

Abstract

The invention relates to the technical field of elevator module updating, and discloses an AI-based elevator updating control method and system.The method comprises the steps that prediction is conducted on the basis of operation state data to obtain an adjustment factor, elevator waiting time is calculated in combination with a pre-built waiting time model, the elevator service state can dynamically reflect operation differences, and the elevator updating control accuracy is improved. And then the elevator coverage area of the target elevator group is divided according to the elevator waiting time, the migration load value is evaluated based on a queuing theory model, a dynamic upgrading path diagram is constructed, the recognition capacity of task transfer capacity differences among the elevators is enhanced, and finally the elevator upgrading sequence is determined with the migration load value minimization as the target based on a graph search algorithm. Upgrading tasks are distributed according to the sequence with the minimum system overall service influence when a plurality of elevator units need to be updated, and therefore the problems of overload, task overstock or service blind areas caused by improper scheduling are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of elevator module updating, and more particularly to an AI-based elevator updating control method and system. BACKGROUND

[0002] In intelligent buildings and high-rise buildings, as the core vertical transportation tool, the elevator system, with the increasing complexity of software functions and the continuous iteration of AI scheduling strategies, the regular upgrade of the running control program has become a necessary means to ensure its performance and safety. Unlike the traditional whole-machine shutdown upgrade method, in recent years, a modular updating elevator upgrade method has gradually developed, that is, only one or more functional modules in the elevator control system are upgraded or replaced by remote patches in a hot-swappable manner. This modular updating method does not need to shut down the whole elevator, and only needs to complete the local logic replacement in a short time, so as to realize function enhancement and system optimization, has the advantages of high availability, small upgrade window, and no influence on the main process, and has become an important trend of the new generation of elevator control systems.

[0003] However, in actual application, when multiple elevator units need to perform modular updating operations in the same time window, the existing elevator scheduling system mostly uses a fixed sequence or a simple polling strategy to determine the upgrade sequence of each elevator. Such strategies fail to consider the current running state of the system, the load difference between elevators, and the task coverage relationship, and lack an upgrade decision mechanism targeting "minimizing the overall service impact of the system". Especially during the peak period of user requests, if the scheduling is improper, it is easy to cause overloading of some elevators, task backlog, and even the emergence of service blind areas.

[0004] In view of this, the present application provides an AI-based elevator updating control method and system to solve the above problems. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the present application provides an AI-based elevator updating control method and system.

[0006] To achieve the above object, the present application provides the following technical scheme:

[0007] In a first aspect, an AI-based elevator updating control method is provided, comprising:

[0008] obtaining running state data of each elevator in a target elevator group, predicting an adjustment factor based on the running state data, and calculating an elevator waiting time based on a pre-constructed waiting time model and the adjustment factor, the adjustment factor including a speed adjustment factor and a time adjustment factor, the target elevator group referring to an elevator group that needs to be upgraded;

[0009] The target elevator group's coverage area is divided based on elevator waiting time. The migration load between each two elevators in the coverage area is evaluated using a queuing theory model. A dynamic upgrade path diagram is constructed, with elevators as nodes and migration loads as edges.

[0010] Based on the graph search algorithm, the dynamic upgrade path graph is converted into a sequential decision process. The elevator upgrade sequence is determined with the goal of minimizing the migration load value, and the preset upgrade tasks are assigned to the corresponding elevators according to the elevator upgrade sequence.

[0011] In some embodiments, a method for predicting an adjustment factor based on operating status data includes:

[0012] The operating status data is input into the pre-trained deep neural network model to obtain the adjustment factor output by the deep neural network model.

[0013] Training methods for deep neural network models include:

[0014] The input layer in the deep neural network model receives historical operating status data, and the output layer in the deep neural network model outputs historical adjustment factors;

[0015] When training the deep neural network model, the cross entropy loss function is selected as the loss function, and the loss function is minimized by the gradient descent method;

[0016] Update the weight parameters of the deep neural network model and obtain the trained deep neural network model through iterative training.

[0017] In some embodiments, a method for calculating the elevator waiting time based on a pre-built waiting time model and an adjustment factor includes:

[0018]

[0019] Where, EWT′ i is the elevator waiting time calculated after the adjustment factor is introduced for elevator i, f is the floor number of the current passenger request, and f i is the floor number where elevator i is currently located, v i is the standard design running speed of elevator i, Δv is the speed adjustment factor, n is the number of stopping floors in the current task queue of the elevator, f ij is the jth target floor to be served in the elevator task queue, f i(j-1) is the floor where the elevator is located at the j-1th task, t s is the standard average dwell time of the elevator at each normal stop, and Δt is the time adjustment factor.

[0020] In some embodiments, the method for dividing the elevator coverage area of the target elevator group according to the elevator waiting time comprises:

[0021] The elevator waiting time is subjected to sliding differential processing to obtain a waiting time change rate;

[0022] The average waiting time deviation between the elevator waiting time and the historical waiting time is calculated;

[0023] The preset coverage area is divided into Q floor sub-areas, the waiting time change rate and the average waiting time deviation are mapped to a preset two-dimensional state diagram, and an adjustment index is output through the two-dimensional state diagram;

[0024] When the adjustment index is greater than or equal to a preset index threshold, Q-K floor sub-areas are taken as the elevator coverage area, and when the adjustment index is less than the preset index threshold, Q floor sub-areas are taken as the elevator coverage area, wherein Q is greater than K, and K is an integer greater than 0.

[0025] In some embodiments, the method for dividing the elevator coverage area of the target elevator group according to the elevator waiting time comprises:

[0026] The elevator waiting time is divided into a waiting time sequence through a preset time window, the difference between every two adjacent elements in the waiting time sequence is processed, the unit time change rate between every two adjacent elements is calculated, and the average value between a plurality of unit time change rates is taken as the waiting time change rate.

[0027] In some embodiments, the method for constructing the two-dimensional state diagram comprises:

[0028] The waiting time change rate is taken as the horizontal axis of the two-dimensional state diagram, the average waiting time deviation is taken as the vertical axis of the two-dimensional state diagram, the section division boundary in the two-dimensional state diagram is determined based on a clustering algorithm, the two-dimensional state diagram is divided into P state sections according to the section division boundary, and P is an integer greater than 1.

[0029] In some embodiments, the method for evaluating the transfer load value between every two elevators in the elevator coverage area based on the queuing theory model comprises:

[0030] The queue length difference and the processing quantity difference between every two elevators are extracted based on the queuing theory model, and the transfer load value is determined according to the queue length difference and the processing quantity difference.

[0031] In some embodiments, the method for extracting the queue length difference and the processing quantity difference between every two elevators based on the queuing theory model comprises:

[0032] The running state data and the elevator waiting time of each elevator are input into the queuing theory model, and the future queue length and the future processing quantity of each elevator are predicted, the queue length difference between each two elevators is calculated by difference, and the processing quantity difference between each two elevators is calculated by difference.

[0033] In some embodiments, the method for determining the migration load value according to the queue length difference and the processing quantity difference comprises:

[0034] The first weight factor of the queue length difference is determined according to the entropy weight method, the second weight factor of the processing quantity difference is determined according to the entropy weight method, the queue length difference and the processing quantity difference are weighted and summed based on the first weight factor and the second weight factor, and the migration load value is obtained.

[0035] In the second aspect, an AI-based elevator updating control system is provided, which is used to implement the AI-based elevator updating control method described above, and comprises:

[0036] The prediction module is used to obtain the running state data of each elevator in the target elevator group, to obtain an adjustment factor based on the running state data, and to calculate the elevator waiting time based on the pre-constructed waiting time model and the adjustment factor, wherein the adjustment factor comprises a speed adjustment factor and a time adjustment factor, and the target elevator group refers to the elevator group that needs to be upgraded;

[0037] The first processing module is used to divide the elevator coverage area of the target elevator group according to the elevator waiting time, to evaluate the migration load value between each two elevators in the elevator coverage area based on the queuing theory model, to construct a dynamic upgrading path graph with the elevators as nodes and the migration load value as edges, and to determine the elevator upgrading sequence based on the graph search algorithm.

[0038] The second processing module is used to convert the dynamic upgrading path graph into a sequence decision process based on the graph search algorithm, to determine the elevator upgrading sequence with the minimum migration load value as the target, and to assign the preset upgrading task to the corresponding elevator according to the elevator upgrading sequence.

[0039] Compared with the prior art, the present application has the following advantages:

[0040] The application obtains an adjustment factor based on operation state data, calculates elevator waiting time in combination with a pre-constructed waiting time model, enables the elevator service state to dynamically reflect operation differences, divides the elevator coverage area of the target elevator group through elevator waiting time, evaluates the migration load value based on a queuing theory model, constructs a dynamic upgrade path diagram, enhances the identification ability of the task transfer ability difference between elevators, and finally determines the elevator upgrade sequence based on a graph search algorithm to minimize the migration load value, realizes the allocation of upgrade tasks in the order of minimizing the overall system service impact when multiple elevator units need to be updated, and avoids the problems of overload, task backlog or service blind area caused by improper scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 A flowchart of an AI-based elevator update control method in the application is shown in the figure.

[0042] Figure 2 A flowchart of a two-dimensional state diagram construction method in the application is shown in the figure.

[0043] Figure 3 A structural diagram of an AI-based elevator update control system in the application is shown in the figure. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions and advantages of the application clearer, further detailed description will be made below in combination with specific embodiments and with reference to the drawings, in the following detailed description, many specific details are set forth to provide a thorough understanding of the described exemplary embodiments, however, it is obvious to those skilled in the art that the described embodiments can be practiced without some or all of these specific details, in other exemplary embodiments, well-known structures are not described in detail to avoid unnecessary obscuring of the concepts of the present disclosure, it should be understood that the specific embodiments described herein are only used to explain the application and not to limit the application, and the aspects described in the embodiments can be combined arbitrarily without conflict.

[0045] Embodiment 1

[0046] Please refer to Figure 1 The embodiment disclosed in the figure provides an AI-based elevator update control method, which comprises:

[0047] S10: Obtain the operation state data of each elevator in the target elevator group, obtain an adjustment factor based on the operation state data, and calculate the elevator waiting time according to the pre-constructed waiting time model and the adjustment factor, the adjustment factor includes a speed adjustment factor and a time adjustment factor, and the target elevator group refers to the elevator group that needs to be upgraded;

[0048] In the embodiment, the target elevator group refers to an elevator group that needs to be upgraded, which can be a plurality of elevator units deployed in the same building or building group and having a mutual cooperative scheduling relationship, or an elevator set marked as the same functional partition or logical scheduling unit in the scheduling system, such as high-zone elevators, middle-zone elevators, and low-zone elevators in a building according to service floors.

[0049] It should be noted that the running state data includes but is not limited to current floor information, real-time running direction, current load rate, and current idle time length. The real-time running direction refers to the movement trend of the elevator relative to the floor in the current running period, including "up", "down", or "still" state, which is used to determine whether the elevator is responding to a passenger request in a certain direction. The current load rate refers to the proportional relationship between the current number of passengers or the load of the elevator and its maximum rated carrying capacity, which is usually expressed in percentage form, and is used to reflect the current crowdedness or carrying capacity of the elevator. The current idle time length refers to the duration that the elevator has not been assigned an effective task or has not responded to a user request in the current period, which is used to evaluate its schedulability and service idle degree.

[0050] The method for predicting the adjustment factor based on the running state data includes:

[0051] The running state data is input into the pre-trained deep neural network model to obtain the adjustment factor output by the deep neural network model.

[0052] The training method of the deep neural network model includes:

[0053] The input layer of the deep neural network model receives historical running state data, and the output layer of the deep neural network model outputs historical adjustment factors;

[0054] When training the deep neural network model, a cross-entropy loss function is selected as the loss function, and the loss function is minimized by the gradient descent method;

[0055] The weight parameters of the deep neural network model are updated, and the trained deep neural network model is obtained through iterative training.

[0056] It can be understood that the adjustment factor includes a speed adjustment factor and a time adjustment factor. The speed adjustment factor represents the dynamic correction amplitude required relative to the standard design speed under the current elevator running state, and the time adjustment factor represents the correction degree of the average stay time of the elevator at each stop floor relative to the standard stop time under the current scheduling environment.

[0057] So taking the current load rate and the current idle time as examples, the current load rate is one of the key factors affecting the speed adjustment factor. When the elevator load rate is high (such as close to full load), the system often reduces the running speed of the elevator to ensure the safety and stability of operation, so the model will predict a lower speed adjustment factor. At the same time, under high load conditions, the time spent by passengers getting on and off is longer, and the frequency of opening and closing the elevator door increases, resulting in an increase in the time spent at each stop, thereby directly increasing the output value of the time adjustment factor. The current idle time is mainly used to reflect whether the elevator is in an idle state recently. If the elevator is idle for a long time, it means that the task pressure is low, and it can usually run at a speed close to the design speed. The model tends to output a smaller speed correction value. At the same time, the idle elevator usually responds to new requests quickly and stops smoothly, so the corresponding time adjustment factor is low. On the contrary, if the idle time is very short, it means that the task is frequent and the state is busy, and the time adjustment factor is increased.

[0058] The method for calculating the elevator waiting time according to the pre-constructed waiting time model and the adjustment factor includes:

[0059]

[0060] EWT' = f - f + v - v + n * t - n * t + Δt i EWT' is the elevator waiting time calculated after introducing the adjustment factor, f is the floor number where the current passenger request is located, f is the floor number where the elevator i is currently located, v is the standard design running speed of the elevator i, Δv is the speed adjustment factor, n is the number of stop floors in the current task queue of the elevator, f is the jth target floor to be served in the elevator task queue, f is the floor where the elevator is located when the j-1th task is performed, t is the standard average stay time of the elevator at each normal stop floor, and Δt is the time adjustment factor. i i ij i(j-1) s

[0061] ​​​​​In this embodiment, by introducing a deep neural network model, the adjustment factor predicted based on the real-time collected running state data (including the current load rate, the current idle time, the current floor and the running direction, etc.) is used to dynamically correct the elevator design speed and the standard stopping time, so as to convert the elevator waiting time from the traditional static estimation mechanism to a dynamic prediction mechanism. For example, when the elevator load rate is high, the model predicts a downward trend of the corresponding speed and an extension trend of the stopping time, which is reflected in the waiting time model as a decrease in the overall service capacity. When the elevator is in an idle state for a long time, the model can output a small correction value to restore the waiting time to a level close to the design benchmark, thereby significantly improving the accuracy and robustness of the waiting time estimation, making the subsequent task migration, coverage area division, upgrade timing arrangement and other links more consistent with the actual running state of the system, effectively avoiding the problem of system service capacity decline or local congestion caused by scheduling estimation errors, and realizing intelligent protection of service continuity and user response experience during the program module upgrade process of the elevator system.

[0062] S20: Dividing the elevator coverage area of the target elevator group according to the elevator waiting time, evaluating the migration load value between each two elevators in the elevator coverage area based on a queuing theory model, and constructing a dynamic upgrade path graph with the elevators as nodes and the migration load value as edges;

[0063] It should be noted that the elevator coverage area refers to the set of floors that the elevator can serve or preferentially respond to in the dispatching system. The elevator coverage area can be the set of all floors that the elevator is structurally allowed to stop at, or the set of target floors that the elevator is preferentially assigned tasks in the dispatching strategy.

[0064] The method for dividing the elevator coverage area of the target elevator group according to the elevator waiting time comprises the following steps:

[0065] The elevator waiting time is subjected to a sliding difference process to obtain a waiting time change rate;

[0066] The average waiting time deviation between the elevator waiting time and the historical waiting time is calculated;

[0067] The preset coverage area is divided into Q floor sub-areas, the waiting time change rate and the average waiting time deviation are mapped to a preset two-dimensional state diagram, and an adjustment index is output through the two-dimensional state diagram;

[0068] When the adjustment index is greater than or equal to a preset index threshold, Q-K floor sub-areas are taken as the elevator coverage area, and when the adjustment index is less than the preset index threshold, Q floor sub-areas are taken as the elevator coverage area, wherein Q is greater than K, and K is an integer greater than 0.

[0069] It can be understood that the preset coverage area refers to a set of serviceable floors of each elevator in the elevator group control system, which is preset based on the physical structure, stopping capacity, scheduling strategy and service partition rule of the elevator, such as high-zone elevator serving 20-40 floors and low-zone elevator serving 1-20 floors. The floor sub-area refers to a plurality of continuous floor paragraphs obtained by dividing the preset coverage area of the elevator according to a predetermined granularity, each sub-area is composed of a group of adjacent floors, and each sub-area is the smallest movable service unit, which can be divided by a fixed number of floors, such as every 5 floors.

[0070] The method for performing sliding difference processing on the elevator waiting time to obtain the waiting time change rate comprises the following steps:

[0071] The elevator waiting time is divided into a waiting time sequence by a preset time window, the difference between each two adjacent elements in the waiting time sequence is processed, the unit time change rate between each two adjacent elements is calculated, and the average value between a plurality of unit time change rates is taken as the waiting time change rate.

[0072] In the embodiment, the preset time window refers to a continuous time period set by the system for statistics and analysis when performing dynamic trend analysis on the elevator waiting time. The elements in the waiting time sequence represent the average waiting time value of the elevator at a specific sampling time, which can be obtained by statistics or averaging of the queuing time of all tasks to be responded at that time.

[0073] It should be noted that the historical average waiting time is calculated by the historical waiting time, and the difference between the elevator waiting time and the historical average waiting time is taken as the average waiting time deviation.

[0074] Specifically, the two-dimensional state diagram refers to a two-dimensional feature space region division model constructed by taking the "waiting time change rate" and the "average waiting time deviation" as two orthogonal dimensions, which is used to reflect the joint evaluation result of the current service state of the elevator in the trend and stability directions. The horizontal axis of the two-dimensional state diagram is the "waiting time change rate", which is used to describe whether the elevator service trend is rising, falling or stable, and the vertical axis is the "average waiting time deviation", which is used to judge whether the current load of the elevator deviates from its historical running baseline. The two-dimensional space is divided into a plurality of state sections (such as low-risk area, critical area, high-risk area, etc.), and each section corresponds to a preset adjustment index interval value.

[0075] The application guides the dynamic release of the floor sub-regions by adjusting the index output, specifically, the system first maps the current service state of the elevator to a two-dimensional state diagram based on the waiting time change rate and the average waiting time deviation of the elevator, and obtains a corresponding adjustment index. The adjustment index is not only used to determine whether to trigger the regional release operation, but also is linked with the real-time service state of each sub-region in the covered region. When the release condition is met, the sub-region with a larger historical waiting time deviation is preferentially released, or the sub-region corresponding to the high-risk state section is released, so as to realize the dynamic mapping between the adjustment index and the specific floor release strategy.

[0076] As shown in Figure 2 The construction method of the two-dimensional state diagram includes:

[0077] The waiting time change rate is taken as the horizontal axis of the two-dimensional state diagram, and the average waiting time deviation is taken as the vertical axis of the two-dimensional state diagram. The section division boundary in the two-dimensional state diagram is determined based on a clustering algorithm. The two-dimensional state diagram is divided into P state sections according to the section division boundary, and P is an integer greater than 1.

[0078] The application introduces two service state indicators, the waiting time change rate and the average waiting time deviation, which have dynamic and trend perception capabilities, to accurately depict the running trend and load stability of the elevator in the current period. On this basis, a two-dimensional state diagram model based on a clustering algorithm is constructed. The above two indicators are mapped to the horizontal axis and the vertical axis of the two-dimensional space, respectively. Through adaptive learning and clustering analysis, the two-dimensional space is divided into multiple state sections with scheduling significance (for example, low load area, fluctuation area, high risk area, etc.), and each state section is assigned a corresponding adjustment index output interval.

[0079] The adjustment index, as a fused quantitative result of the elevator service state, is used to drive the flexible dynamic division strategy of the elevator covered region. When the adjustment index is higher than the set threshold, the system automatically triggers the release of part of the floor sub-regions. When the adjustment index is lower, the service coverage range of the elevator is maintained, thereby realizing the adaptive adjustment of the region based on the service pressure state.

[0080] Through this mechanism, the system not only realizes the visual risk classification of the elevator running state, but also realizes multi-dimensional service risk identification and trend prediction control without relying on fixed rules, greatly improving the scheduling flexibility, service continuity and overall response efficiency of the group control system in the case of multi-elevator program upgrade, temporary shutdown or high load burst.

[0081] The method for evaluating the migration load value between each two elevators in the elevator covered region based on the queuing theory model includes:

[0082] The queue length difference and the processing quantity difference between each two elevators are extracted based on the queuing theory model, and a migration load value is determined according to the queue length difference and the processing quantity difference.

[0083] The method for extracting the queue length difference and the processing quantity difference between each two elevators based on the queuing theory model comprises:

[0084] The running state data and the elevator waiting time of each elevator are input into the queuing theory model, the future queue length and the future processing quantity of each elevator are predicted, the future queue length difference between each two elevators is calculated by difference, and the future processing quantity difference between each two elevators is calculated by difference.

[0085] It should be noted that the queuing theory model is a mathematical modeling method based on the theory of operations research, which is used to analyze the running process of a system with the characteristics of task arrival, queuing waiting, resource service and completion departure. In the elevator dispatching system, the passenger call request can be regarded as the task arrival, and the response and transportation process of the elevator to the request can be regarded as the service process. Through the queuing theory model, the service capacity, queue length, waiting time and other key dispatching performance indicators of the elevator system in a future time period can be predicted, which is an important evaluation tool for task allocation and dispatching optimization in the field.

[0086] In the present application, the future queue length represents the number of tasks accumulated by the elevator in a future preset time window, and the number of tasks can correspond to the floor call, passenger request or dispatching instruction required to be responded by the elevator in the current or future period of time. The future processing quantity refers to the number of dispatching tasks that the elevator is expected to complete per unit time, which is a key indicator of the service capacity of the elevator. The future processing quantity can dynamically reflect the task emptying speed and service response capacity of the elevator in the future period of time. The running state data of the elevator (such as the current task queue length, the current load rate and the current idle time) and the elevator waiting time jointly constitute the core input variables of the queuing theory model. The task queue length can be used to initialize the current queuing state, the current load rate and the current idle time reflect the service capacity and task saturation degree of the elevator, and the elevator waiting time reflects the system response efficiency. Based on these input data, the queuing theory model can dynamically estimate the task arrival rate and the service rate of the elevator, and further predict the queue length accumulated by the elevator in the future preset time window and the task processing quantity per unit time, thereby providing accurate support for cost evaluation in the task migration path.

[0087] The method for determining the migration load value according to the queue length difference and the processing quantity difference comprises:

[0088] The first weight factor of the queue length difference is determined according to the entropy weight method, the second weight factor of the processing quantity difference is determined according to the entropy weight method, the queue length difference and the processing quantity difference are weighted and summed based on the first weight factor and the second weight factor, and a migration load value is obtained.

[0089] It should be noted that the queue length difference is taken as an example, assuming that the queue length difference data of the target elevator group is standardized and the proportion distribution is [0.4, 0.3, 0.2, 0.1], the sample number r = 4, and the entropy value is calculated:

[0090]

[0091] w L = 1-0.85 = 0.15;

[0092] wherein E is the entropy value, w L is the first weight factor.

[0093] Similarly, the second weight factor of the processing quantity difference can be determined according to the entropy weight method, and the embodiment does not make too much repetition.

[0094] The migration load value is used to measure the service pressure required to migrate the task of the elevator to be upgraded to another elevator, and is an important quantitative index of the task migration cost between elevators. The present application predicts the future queue length and future processing quantity of each elevator through the queuing theory model, calculates the queue length difference and the processing quantity difference based on the difference between the two elevators, represents the task supply and demand imbalance state, the queue length difference reflects the acceptable space of the target elevator, and the processing quantity difference represents the processing capacity per unit time in the future. The two together constitute the "resistance strength" of the elevator to task migration. The system then weights and fuses the two indexes according to the entropy weight method, and outputs the migration load value as the edge weight value in the dynamic upgrade path graph, providing a path reference with the smallest service burden for subsequent upgrade sequence optimization.

[0095] In the embodiment, by introducing the queuing theory model, combining the elevator running state data and waiting time information, the dynamic prediction of the future queue length and future processing capacity of the elevator is realized, replacing the task allocation and scheduling evaluation mode based on the current static parameters in the traditional scheme.

[0096] In addition, in order to avoid the deviation caused by the subjective setting of the task migration cost weight in the traditional scheduling method, the entropy weight method is further introduced in the present application, the weight coefficients of the queue length difference and the processing quantity difference are adaptively calculated according to the discreteness and information entropy under different scheduling states, and a more objective and reasonable "migration load value" index is constructed.

[0097] Unlike the prior art which only calculates the migration cost based on fixed formula or empirical rules, the method proposed in the present application realizes dynamic modeling, accurate quantification and state adaptive adjustment of task migration cost evaluation, can significantly improve the perception ability of system load changes in the task migration process, effectively reduce the risk of service congestion caused by unreasonable scheduling, and guarantee the service continuity and response balance of the elevator group control system during program upgrade.

[0098] S30: converting the dynamic upgrade path graph into a sequence decision process based on a graph search algorithm, determining an elevator upgrade sequence with the goal of minimizing the migration load value, and distributing the preset upgrade task to the corresponding elevator according to the elevator upgrade sequence.

[0099] In the embodiment, the graph search algorithm can be A* algorithm or Dijkstra algorithm. The graph search algorithm (such as Dijkstra algorithm or A* heuristic search algorithm) is a classic algorithm specially used for finding the shortest path or minimum total cost path in a graph structure. In the present application, the graph search algorithm is used to search the dynamic upgrade path graph with the goal of minimizing the migration load value, which is equivalent to finding an upgrade access sequence with the minimum cumulative task migration cost among multiple elevators.

[0100] Since the edge weight in the graph objectively reflects the scheduling influence strength between elevators, the graph search algorithm always preferentially selects the path extension direction with the minimum current cost during node traversal, thereby outputting a globally optimal or approximately optimal elevator upgrade sequence, ensuring that the overall service burden of the system is minimized during the upgrade process, and avoiding the problems of scheduling imbalance or enlarged blind area caused by the priority upgrade of high-load elevators. In this way, the system triggers the upgrade process of each elevator in sequence according to the arrangement of each elevator in the sequence, and schedules the corresponding upgrade task (such as AI control module update, function patch installation, etc.) to the elevator node in the current sequence for execution.

[0101] The embodiment obtains the adjustment factor by prediction based on the running state data, calculates the elevator waiting time in combination with the pre-constructed waiting time model, enables the elevator service state to dynamically reflect the running difference, divides the elevator coverage area of the target elevator group through the elevator waiting time, evaluates the migration load value based on the queuing theory model, constructs the dynamic upgrade path graph, enhances the recognition ability of the task transfer ability difference between elevators, and finally determines the elevator upgrade sequence based on the graph search algorithm with the goal of minimizing the migration load value, realizes the distribution of upgrade tasks in the order of minimizing the overall system service impact when multiple elevator units need to be updated, and thereby avoids the problems of overload, task backlog or service blind area caused by improper scheduling.

[0102] Embodiment 2

[0103] Please refer to Figure 3As shown, based on the same inventive concept, this embodiment discloses an AI-based elevator update control system. For details not provided in this embodiment, please refer to the description of the relevant parts in Example 1. The system includes:

[0104] Prediction module: used to obtain the operating status data of each elevator in the target elevator group, predict the adjustment factor based on the operating status data, and calculate the elevator waiting time according to the pre-built waiting time model and the adjustment factor, including the speed adjustment factor and the time adjustment factor. The target elevator group is the elevator group that needs to be upgraded;

[0105] In this embodiment, the target elevator group refers to the elevator group that needs to be upgraded. The elevator group can be multiple elevator units deployed in the same building or building group with a mutually coordinated scheduling relationship, or it can be a collection of elevators marked as the same functional partition or logical scheduling unit in the scheduling system, such as the high-zone elevator, middle-zone elevator and low-zone elevator divided according to the service floors in a building.

[0106] Methods for obtaining adjustment factors based on predictions of operating status data include:

[0107] The operating status data is input into the pre-trained deep neural network model to obtain the adjustment factor output by the deep neural network model.

[0108] Training methods for deep neural network models include:

[0109] The input layer in the deep neural network model receives historical operating status data, and the output layer in the deep neural network model outputs historical adjustment factors;

[0110] When training the deep neural network model, the cross entropy loss function is selected as the loss function, and the loss function is minimized by the gradient descent method;

[0111] Update the weight parameters of the deep neural network model and obtain the trained deep neural network model through iterative training.

[0112] Methods for calculating elevator waiting time based on a pre-built waiting time model and adjustment factors include:

[0113]

[0114] Where, EWT′ i is the elevator waiting time calculated after the adjustment factor is introduced for elevator i, f is the floor number of the current passenger request, and f i is the floor number that elevator i is currently on, v i is the standard design running speed of elevator i, Δv is the speed adjustment factor, n is the number of stopping floors in the current task queue of the elevator, fij f is the jth target floor to be served in the elevator task queue, i(j-1) t is the floor where the elevator is located at the j-1th task, s Δt is the standard average stay time of each normal stop floor of the elevator, and is a time adjustment factor.

[0115] The first processing module is used to divide the elevator coverage area of the target elevator group according to the elevator waiting time, evaluate the migration load value between each two elevators in the elevator coverage area based on a queuing theory model, and construct a dynamic upgrade path graph with the elevators as nodes and the migration load values as edges.

[0116] It should be noted that the elevator coverage area refers to a range of floor sets that can be served or preferentially responded by the elevator in the dispatching system. The elevator coverage area can be the entire set of floors that the elevator is structurally allowed to stop at, or the set of target floors that the elevator is preferentially assigned tasks in the dispatching strategy.

[0117] The method for dividing the elevator coverage area of the target elevator group according to the elevator waiting time comprises the following steps:

[0118] The elevator waiting time is subjected to a sliding difference processing to obtain a waiting time change rate;

[0119] The average waiting time deviation between the elevator waiting time and the historical waiting time is calculated;

[0120] The preset coverage area is divided into Q floor sub-areas, the waiting time change rate and the average waiting time deviation are mapped to a preset two-dimensional state diagram, and an adjustment index is output through the two-dimensional state diagram;

[0121] When the adjustment index is greater than or equal to a preset index threshold, Q-K floor sub-areas are taken as the elevator coverage area, and when the adjustment index is less than the preset index threshold, Q floor sub-areas are taken as the elevator coverage area, wherein Q is greater than K, and K is an integer greater than 0.

[0122] The method for obtaining the waiting time change rate by subjecting the elevator waiting time to a sliding difference processing comprises the following steps:

[0123] The elevator waiting time is divided into a waiting time sequence through a preset time window, each two adjacent elements in the waiting time sequence are subjected to a difference processing, the unit time change rate between each two adjacent elements is calculated, and the average value between a plurality of unit time change rates is taken as the waiting time change rate.

[0124] The method for constructing the two-dimensional state diagram comprises the following steps:

[0125] The waiting time change rate is taken as the horizontal axis of the two-dimensional state diagram, and the average waiting time deviation is taken as the vertical axis of the two-dimensional state diagram, a section division boundary in the two-dimensional state diagram is determined based on a clustering algorithm, and the two-dimensional state diagram is divided into P state sections according to the section division boundary, where P is an integer greater than 1.

[0126] The method for evaluating the migration load value between each two elevators in the elevator coverage area based on the queuing theory model comprises:

[0127] The queue length difference and the processing quantity difference between each two elevators are extracted based on the queuing theory model, and the migration load value is determined according to the queue length difference and the processing quantity difference.

[0128] The method for extracting the queue length difference and the processing quantity difference between each two elevators based on the queuing theory model comprises:

[0129] The running state data of each elevator and the elevator waiting time are input into the queuing theory model, the future queue length and the future processing quantity of each elevator are predicted, the queue length difference between each two elevators is calculated by difference, and the processing quantity difference between each two elevators is calculated by difference.

[0130] The method for determining the migration load value according to the queue length difference and the processing quantity difference comprises:

[0131] The first weight factor of the queue length difference is determined according to the entropy weight method, the second weight factor of the processing quantity difference is determined according to the entropy weight method, the queue length difference and the processing quantity difference are weighted and summed based on the first weight factor and the second weight factor, and the migration load value is obtained.

[0132] The second processing module is configured to: convert the dynamic upgrade path graph into a sequence decision process based on a graph search algorithm, determine an elevator upgrade sequence for minimizing the migration load value, and distribute a preset upgrade task to a corresponding elevator according to the elevator upgrade sequence.

[0133] The detailed description set forth above in connection with the appended drawings describes examples and does not represent all of the examples that can be implemented or that are within the scope of the claims. The terms “example” and “exemplary” are used herein to mean “serving as an example, instance, or illustration,” and not to imply “preferred” or “superior” over other examples. The detailed description includes specific details for the purpose of providing an understanding of the described techniques.

[0134] “one embodiment” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application, the appearances of the phrase “in one embodiment” or “in an embodiment” in various places in the specification are not necessarily all referring to the same embodiment, and some embodiments can be widely different from other embodiments.

[0135] It should also be noted that these embodiments can be described as a process which is depicted as a flowchart, a structure chart, or a block diagram, although the flowchart can describe an operational flow, the process can be a method or a function, although a function can be a method with a return value or an identification code, the flowchart can depict a computer program product or instructions stored in a computer readable medium, although a flowchart can depict a machine or a computer program product, although a data processing system can syntactically, textually, or programmatically implement the flowchart, and the various embodiments can be implemented by various means, such as hardware, software, firmware, or any combination of hardware, software and firmware

Claims

1. An AI-based elevator update control method, characterized in that: include: Obtaining operating status data of each elevator in a target elevator group, performing a prediction based on the operating status data to obtain an adjustment factor, and calculating the elevator waiting time according to a pre-built waiting time model and the adjustment factor, wherein the adjustment factor includes a speed adjustment factor and a time adjustment factor. The target elevator group is the elevator group that needs to be upgraded; The target elevator group's coverage area is divided based on elevator waiting time. The migration load between each two elevators in the coverage area is evaluated using a queuing theory model. A dynamic upgrade path diagram is constructed, with elevators as nodes and migration loads as edges. Based on the graph search algorithm, the dynamic upgrade path graph is converted into a sequential decision process. The elevator upgrade sequence is determined with the goal of minimizing the migration load value, and the preset upgrade tasks are assigned to the corresponding elevators according to the elevator upgrade sequence.

2. The AI-based elevator update control method according to claim 1, characterized in that: The method for obtaining the adjustment factor by prediction based on the operating status data includes: The operating status data is input into the pre-trained deep neural network model to obtain the adjustment factor output by the deep neural network model. Training methods for deep neural network models include: The input layer in the deep neural network model receives historical operating status data, and the output layer in the deep neural network model outputs historical adjustment factors; When training the deep neural network model, the cross entropy loss function is selected as the loss function, and the loss function is minimized by the gradient descent method; Update the weight parameters of the deep neural network model and obtain the trained deep neural network model through iterative training.

3. The AI-based elevator update control method according to claim 2, characterized in that: The method for calculating the elevator waiting time based on the pre-built waiting time model and the adjustment factor includes: Where, EWT′ i is the elevator waiting time calculated after the adjustment factor is introduced for elevator i, f is the floor number of the current passenger request, and f i is the floor number that elevator i is currently on, v i is the standard design running speed of elevator i, Δv is the speed adjustment factor, n is the number of stopping floors in the current task queue of the elevator, f ij is the jth target floor to be served in the elevator task queue, f i(j-1) is the floor where the elevator is located at the j-1th task, t s is the standard average dwell time of the elevator at each normal stop, and Δt is the time adjustment factor.

4. The AI-based elevator update control method according to claim 1, characterized in that: The method for dividing the elevator coverage area of ​​the target elevator group according to the elevator waiting time includes: Perform sliding difference processing on the elevator waiting time to obtain the waiting time change rate; Calculate the average waiting time deviation between elevator waiting time and historical waiting time; Divide the preset coverage area into Q floor sub-areas, map the waiting time change rate and average waiting time deviation to a preset two-dimensional state diagram, and output the adjustment index through the two-dimensional state diagram; When the adjustment index is greater than or equal to the preset index threshold, Q-K floor sub-areas are used as elevator coverage areas. When the adjustment index is less than the preset index threshold, Q floor sub-areas are used as elevator coverage areas, where Q is greater than K and K is an integer greater than 0.

5. The AI-based elevator update control method according to claim 4, characterized in that: Methods for performing sliding difference processing on the elevator waiting time to obtain the waiting time change rate include: Divide the elevator waiting time into waiting time sequences through preset time windows; Perform difference processing between every two adjacent elements in the waiting time series and calculate the unit time change rate between every two adjacent elements; The average value of the multiple unit time change rates is taken as the waiting time change rate.

6. The AI-based elevator update control method according to claim 4, characterized in that: The construction methods of the two-dimensional state diagram include: The waiting time change rate is used as the horizontal axis of the two-dimensional state diagram, and the average waiting time deviation is used as the vertical axis of the two-dimensional state diagram; Based on the clustering algorithm, the segment division boundary in the two-dimensional state graph is determined, and the two-dimensional state graph is divided into P state segments according to the segment division boundary, where P is an integer greater than 1.

7. The AI-based elevator update control method according to claim 1, characterized in that: The method for evaluating the migration load value between every two elevators in an elevator coverage area based on a queuing theory model includes: Based on the queuing theory model, the queue length difference and the processing quantity difference between every two elevators are extracted, and the migration load value is determined according to the queue length difference and the processing quantity difference.

8. The AI-based elevator update control method according to claim 7, characterized in that: The method for extracting the queue length difference and processing quantity difference between each two elevators based on the queuing theory model includes: Input the operating status data and waiting time of each elevator into the queuing theory model to predict the future queue length and future processing quantity of each elevator; The difference between the future queue lengths of each two elevators is calculated to obtain the queue length difference, and the difference between the future processing quantities of each two elevators is calculated to obtain the processing quantity difference.

9. The AI-based elevator update control method according to claim 8, characterized in that: The method for determining the migration load value based on the queue length difference and the processing quantity difference includes: Determine a first weighting factor for the queue length difference based on the entropy weight method, and determine a second weighting factor for the processing quantity difference based on the entropy weight method; The queue length difference and the processing quantity difference are weighted and summed based on the first weight factor and the second weight factor to obtain a migration load value.

10. An AI-based elevator update control system, used to implement the AI-based elevator update control method according to any one of claims 1 to 9, characterized in that: include: Prediction module: used to obtain the operating status data of each elevator in the target elevator group, predict the adjustment factor based on the operating status data, and calculate the elevator waiting time according to the pre-built waiting time model and the adjustment factor, including the speed adjustment factor and the time adjustment factor. The target elevator group is the elevator group that needs to be upgraded; The first processing module is used to divide the elevator coverage area of ​​the target elevator group according to the elevator waiting time. Based on the queuing theory model, the migration load value between each two elevators in the elevator coverage area is evaluated. The dynamic upgrade path diagram is constructed with the elevators as nodes and the migration load values ​​as edges. The second processing module is used to convert the dynamic upgrade path graph into a sequence decision process based on a graph search algorithm, determine the elevator upgrade sequence with the goal of minimizing the migration load value, and assign the preset upgrade tasks to the corresponding elevators according to the elevator upgrade sequence.

Images