Elevator dispatching system, elevator dispatching engine and electronic equipment

Abstract

The invention provides an elevator dispatching system, an elevator dispatching engine and electronic equipment, and the elevator dispatching system comprises an elevator data collection module which is used for collecting real-time operation data of an elevator; the elevator state recognition module is used for recognizing the current elevator state according to the elevator real-time operation data; the resource obtaining module is used for obtaining the request of the elevator on each floor, the elevator operation system load and the elevator resources; the strategy adjusting module is used for dynamically adjusting the elevator dispatching strategy according to the current elevator state, the request of the elevator on each floor and the elevator operation system load and the elevator resources, the dispatching strategy can be automatically switched according to the actual requirement, a self-adaptive and real-time optimized elevator dispatching decision is achieved, and the dispatching efficiency is improved. And the scheduling efficiency and flexibility of the system are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent building management and artificial intelligence technology, and in particular to an elevator dispatching system, an elevator dispatching engine and electronic equipment. Background Art

[0002] As the height and scale of urban buildings continue to increase, the elevator system, as an important component of modern buildings, has become an increasingly critical issue in terms of scheduling. Traditional elevator scheduling methods, such as shortest response time scheduling and priority scheduling, have some effects, but they can only be applied to the same scenarios, such as peak demand scenarios, sudden emergency scenarios, and dynamically changing passenger demand scenarios. In actual use, elevators will have the problem of switching between different scenarios. Existing elevator methods often fail to achieve optimized resource allocation and intelligent decision-making when facing the switching of operation scenarios, resulting in insufficient adaptive capabilities in multiple scenarios, resulting in inefficient elevator scheduling systems, long waiting times for passengers, and waste of resources. Summary of the invention

[0003] The present invention provides an elevator dispatching system, an elevator dispatching engine and electronic equipment, which are used to solve the defect that a traditional elevator dispatching engine system has insufficient self-adaptation capability in various scenarios.

[0004] The present invention provides an elevator dispatching system, comprising: Elevator data acquisition module, used to collect real-time elevator operation data; An elevator status recognition module, used to recognize the current elevator status according to the real-time operation data of the elevator; Resource acquisition module, used to obtain the elevator request on each floor as well as the elevator operation system load and elevator resources; The strategy adjustment module is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the elevator request at each floor, the elevator operation system load and the elevator resources.

[0005] According to the elevator dispatching system provided by the present invention, the strategy adjustment module includes: A strategy initial selection unit, including at least two scheduling algorithms, for outputting a current scheduling strategy according to different scheduling algorithms; A reward acquisition unit, used to obtain the reward obtained by each scheduling algorithm after executing an action; The scheduling strategy optimization unit is used to optimize the current scheduling strategy according to the rewards obtained by each scheduling algorithm after executing the action.

[0006] According to the elevator dispatching system provided by the present invention, the strategy preliminary selection unit includes: The reinforcement learning framework is used for the elevator system to select an action based on the current state at each moment. After executing the action, the system enters a new state. The execution result of each action will give a reward signal, and the scheduling strategy is optimized by continuously iterating and updating the Q value. The action space corresponding to the reinforcement learning framework includes the running direction of the elevator and the scheduling order of the elevator.

[0007] According to the elevator dispatching system provided by the present invention, the dispatching strategy selection unit further includes: A Q network model, wherein the Q network model includes a neural network, an experience replay mechanism, and a target network; The input of the neural network is the current state of the elevator system, and the output is the Q value of each possible action; Experience replay mechanism: used to store multivariate data sets in an experience pool, where the multivariate data sets consist of historical state-action-reward-next state data. The data in the experience pool is used to provide random training samples during the training of the Q network model. The weight update module is used to periodically update the weights of the target network.

[0008] According to the elevator dispatching system provided by the present invention, the reward obtaining unit includes: A positive reward obtaining unit, used to give positive rewards when the dispatching strategy selected by the elevator reduces the waiting time of passengers, or when resources can be used more efficiently within a certain time period; The negative reward acquisition unit is used to give negative rewards if the scheduling strategy causes the passenger to wait too long, the resources are unevenly used, or the elevator runs idle for a time greater than a preset threshold.

[0009] According to the elevator dispatching system provided by the present invention, the resource acquisition module further comprises: acquiring the request priority of the elevator on each floor; The strategy adjustment module is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the request priority of the elevator on each floor, the request of the elevator on each floor, and the elevator operation system load and elevator resources.

[0010] The present invention also provides an elevator dispatching engine, comprising an elevator dispatching system as described in any one of the above items, and a simulation unit for simulating the elevator operation conditions under different scenarios of the current dispatching strategy; The strategy adjustment module is also used to optimize the current scheduling strategy according to the elevator operation conditions in different scenarios of the current scheduling strategy simulated by the simulation engine.

[0011] According to the elevator dispatching engine provided by the present invention, the elevator dispatching engine further includes: Step-by-step simulation mode for analyzing the process of each scheduling decision step by step; Double-speed simulation mode: used to quickly verify the effect of the scheduling scheme under high load conditions; Real-time simulation mode: used to display scheduling status and waiting time in real time.

[0012] The elevator dispatching engine further comprises: An instance creation unit, used to independently configure simulation parameters, scheduling strategies and scenario settings according to the needs of each user, and create a multi-user independent simulation instance according to the simulation parameters, scheduling strategies and scenarios; The user management and resource allocation unit is used to independently manage the simulation instance of each user.

[0013] The present invention also provides an electronic device, comprising the elevator dispatching engine as described in any one of the above items.

[0014] The elevator dispatching system, elevator dispatching engine and electronic equipment provided by the present invention include an elevator data acquisition module for collecting real-time elevator operation data; an elevator state identification module for identifying the current elevator state according to the real-time elevator operation data; a resource acquisition module for acquiring the elevator request on each floor and the elevator operation system load and elevator resources; a strategy adjustment module for dynamically adjusting the elevator dispatching strategy according to the current elevator state, the elevator request on each floor, the elevator operation system load and the elevator resources, and being able to automatically switch the dispatching strategy according to actual needs, realize adaptive and real-time optimized elevator dispatching decisions, and improve the dispatching efficiency and flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 is a functional structure diagram of an elevator dispatching system provided by an embodiment of the present invention; Figure 2 is a functional structure diagram of an elevator dispatching engine provided by an embodiment of the present invention; Figure 3 It is a functional structure diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Figure 1 The functional structure diagram of the elevator dispatching system provided by the embodiment of the present invention is as follows: Figure 1 As shown, the elevator dispatching system provided by the embodiment of the present invention includes: The elevator data collection module 101 is used to collect the real-time operation data of the elevator; In an embodiment of the present invention, each scheduling decision is made based on the current state of the elevator, which includes real-time elevator operation data in multiple dimensions, such as the current floor of the elevator, passenger demand on each floor, load in the elevator, elevator operation direction, etc.

[0019] An elevator state identification module 102, used to identify the current elevator state according to the real-time operation data of the elevator; Resource acquisition module 103, used to obtain the request of the elevator on each floor and the load of the elevator operation system and the elevator resources; In an embodiment of the present invention, passenger request: The request for each floor will be taken as input, and the system will consider the passenger's request and the priority of the request.

[0020] System load and elevator resources: The resource utilization of elevators (such as load) and the coordinated scheduling of multiple elevators are also important factors in decision-making.

[0021] The strategy adjustment module 104 is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the elevator request at each floor, the elevator operation system load and the elevator resources.

[0022] Traditional dispatching systems cannot adapt to different scenarios such as peak hours and emergency evacuation. The embodiment of the present invention can identify and switch different dispatching strategies through the strategy adjustment module. For example, a short response time strategy is adopted in normal operation, while a priority evacuation strategy is adopted in emergency situations. The system can automatically select the optimal strategy according to the current demand and elevator status to cope with various complex operation scenarios.

[0023] Traditional elevator dispatching methods are applicable to only a few scenarios. For example, elevator dispatching models trained based on peak demand scenario data often fail to achieve optimized resource allocation and intelligent decision-making when the scenario switches to an emergency scenario, which can easily lead to inefficient elevator dispatching systems, long waiting times for passengers, and waste of resources.

[0024] The elevator dispatching system provided by the embodiment of the present invention includes an elevator data acquisition module, which is used to collect real-time elevator operation data; an elevator state identification module, which is used to identify the current elevator state according to the real-time elevator operation data; a resource acquisition module, which is used to obtain the elevator's request on each floor and the elevator operation system load and elevator resources; a strategy adjustment module, which is used to dynamically adjust the elevator dispatching strategy according to the current elevator state, the elevator's request on each floor, the elevator operation system load and the elevator resources, and can automatically switch the dispatching strategy according to actual needs, realize adaptive and real-time optimized elevator dispatching decisions, and improve the dispatching efficiency and flexibility of the system.

[0025] Based on any of the above embodiments, the policy adjustment module includes: A strategy initial selection unit, including at least two scheduling algorithms, for outputting a current scheduling strategy according to different scheduling algorithms; A reward acquisition unit, used to obtain the reward obtained by each scheduling algorithm after executing an action; The scheduling strategy optimization unit is used to optimize the current scheduling strategy according to the rewards obtained by each scheduling algorithm after executing the action.

[0026] Traditional elevator dispatching algorithms are often static and cannot respond to changing passenger demands in real time. The present invention uses a reinforcement learning algorithm to enable the elevator dispatching system to dynamically learn and optimize according to real-time data streams (such as elevator status, floor requests, etc.) to achieve intelligent dispatching. Through the reinforcement learning model, the system can adaptively adjust the dispatching strategy during training, reduce waiting time, and improve resource utilization.

[0027] In an embodiment of the present invention, the strategy initial selection unit includes: The reinforcement learning framework is used for the elevator system to select an action based on the current state at each moment. After executing the action, the system enters a new state. The execution result of each action will give a reward signal, and the scheduling strategy is optimized by continuously iterating and updating the Q value. The action space corresponding to the reinforcement learning framework includes the running direction of the elevator and the scheduling order of the elevator.

[0028] In the embodiment of the present invention, the execution result of each action, such as reduced passenger waiting time and improved resource utilization, will give a reward signal, and the reinforcement learning algorithm optimizes the scheduling strategy by continuously iterating and updating the Q value. The running direction of the elevator includes up, down, stop, transfer, etc., as well as the scheduling order of the elevator. Each floor can set a priority separately to meet the different needs of users. Each action is selected based on the current state, and the elevator dispatching system learns the optimal action by exploring and utilizing strategies.

[0029] In an embodiment of the present invention, the scheduling strategy selection unit further includes: A Q network model, wherein the Q network model includes a neural network, an experience replay mechanism, and a target network; The input of the neural network is the current state of the elevator system, and the output is the Q value of each possible action; Experience replay mechanism: used to store multivariate data sets in an experience pool, where the multivariate data sets consist of historical state-action-reward-next state data. The data in the experience pool is used to provide random training samples during the training of the Q network model. The weight update module is used to periodically update the weights of the target network.

[0030] The embodiment of the present invention adopts the DQN (Deep Q-learning) algorithm to realize the learning of scheduling strategy. DQN uses a deep neural network to approximate the Q function. The input of the neural network is the current state of the elevator system, and the output is the Q value of each possible action. The larger the Q value, the higher the expected return of the action, which means the action is better. In order to reduce the data correlation in reinforcement learning, DQN introduces an experience replay mechanism, and all historical state-action-reward-next state data are stored in an experience pool. During the training process, DQN randomly extracts these historical experiences for training, thereby avoiding the correlation problem between samples. DQN introduces the concept of a target network, and the weight of the target network is updated regularly to avoid the oscillation of the Q value and improve the learning stability.

[0031] Traditionally, some research and applications use machine learning algorithms for elevator scheduling optimization, such as neural network-based elevator scheduling methods. However, these methods usually require a large amount of historical data to train the model, and most of them rely on manually designed reward functions and feature selection.

[0032] The present invention adopts the DQN algorithm and makes full use of the exploration and utilization characteristics of reinforcement learning. DQN continuously interacts with the elevator environment, gains experience from real-time feedback, automatically optimizes the scheduling strategy, and self-adjusts according to different scenarios in each simulation, avoiding the limitations of manually designed reward functions. It can adaptively learn and optimize scheduling strategies, dynamically adjust scheduling decisions in various complex environments, and has stronger adaptability and flexibility.

[0033] In an embodiment of the present invention, the reward acquisition unit includes: A positive reward obtaining unit, used to give positive rewards when the dispatching strategy selected by the elevator reduces the waiting time of passengers, or when resources can be used more efficiently within a certain time period; The negative reward acquisition unit is used to give negative rewards if the scheduling strategy causes the passenger to wait too long, the resources are unevenly used, or the elevator runs idle for a time greater than a preset threshold.

[0034] In an embodiment of the present invention, a reward function is used to measure the quality of a scheduling strategy. Specifically, the reward function takes into account passenger waiting time, resource utilization, and scheduling efficiency. In passenger waiting time, if the scheduling strategy selected by the elevator reduces the waiting time of passengers, the system will give a positive reward. In resource utilization, if the elevator system can use resources more efficiently within a certain period of time (for example, reduce idle operation, reduce unnecessary dwell time, etc.), the system will give a higher reward. In scheduling efficiency, the system will reward based on dimensions such as the frequency of elevator operation and the accuracy of scheduling. The goal is to allow the elevator system to maintain a high scheduling efficiency under different loads such as peak and trough periods.

[0035] Negative rewards: If the scheduling strategy causes passengers to wait too long, resources to be used unevenly, or elevators to run empty frequently, the system will give negative rewards. In this way, the system can avoid inefficient scheduling strategies and gradually optimize the elevator scheduling plan through reinforcement learning.

[0036] In the embodiment of the present invention, the resource acquisition module further comprises: acquiring the request priority of the elevator on each floor; The strategy adjustment module is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the request priority of the elevator on each floor, the request of the elevator on each floor, and the elevator operation system load and elevator resources.

[0037] In an embodiment of the present invention, the elevator dispatching system also includes a data analysis module, which is an auxiliary module of the elevator dispatching engine, and is mainly responsible for obtaining various data during the operation of the elevator, analyzing and processing them, identifying bottlenecks in elevator dispatching, providing optimization suggestions, and generating detailed analysis reports. The data analysis module not only supports the monitoring of elevator system performance, but also provides decision support for elevator managers. Through the big data analysis algorithm, the collected data is deeply analyzed to find potential problems and optimization space. For example, the frequency of use of elevators on different floors, the flow patterns of passengers, etc. are analyzed to provide a basis for the optimization of dispatching strategies.

[0038] In an embodiment of the present invention, the elevator dispatching system also includes a report generation module. According to the analysis results, the module will generate a detailed operation report, including statistical data of various key indicators, dispatching efficiency analysis, energy efficiency report, etc. The system will also provide optimization suggestions based on the data analysis results to help managers adjust dispatching strategies or upgrade equipment. Through real-time analysis of multi-dimensional data such as elevator operation data, passenger flow data, energy consumption, etc., it can provide valuable optimization suggestions for decision makers. The analysis results can generate detailed reports to help users understand the bottlenecks of system operation and adjust the elevator dispatching strategy according to the analysis results.

[0039] Most traditional dispatching algorithms are static dispatching methods based on rules. For example, the shortest distance priority algorithm, first-come-first-served algorithm, and dynamic programming algorithm are all based on static rules or mathematical model calculations. They cannot be dynamically adjusted according to the real-time environment and passenger needs, and are difficult to handle complex scenarios with multiple elevators, multiple floors, and multiple passenger requests. In addition, these algorithms cannot effectively handle emergencies (such as emergency needs, system failures, etc.) and cannot achieve self-optimization.

[0040] The elevator dispatching system provided by the embodiment of the present invention introduces deep reinforcement learning to optimize dispatching decisions, combines reinforcement learning algorithms with classical dispatching algorithms, realizes adaptive and real-time optimized elevator dispatching decisions, and improves the dispatching efficiency and flexibility of the system; realizes multi-scenario adaptation and dispatching strategy switching. Through the multi-scenario adaptation module, the system can automatically switch dispatching strategies according to actual needs to adapt to complex scenarios such as peak hours and emergency evacuation. By optimizing dispatching decisions through reinforcement learning, the dispatching strategy can be effectively adjusted under dynamically changing needs to improve the overall efficiency of the system. Dispatching according to real-time conditions can greatly reduce the average waiting time of passengers and improve user experience. Through the scenario adaptation module, the system can automatically switch dispatching strategies according to different application scenarios (such as emergency evacuation, peak hours, etc.) to ensure the efficient operation of the elevator system in complex environments. Each user can independently configure simulation parameters and dispatching strategies, and multiple simulation instances can run in parallel to ensure data isolation and improve the system concurrency capability. The system modules are independent and interoperable, which is convenient for expanding or updating functions according to needs, and enhancing the adaptability and maintainability of the system.

[0041] Figure 2 The functional structure diagram of the elevator dispatching engine provided by the embodiment of the present invention is as follows: Figure 2 As shown, the elevator dispatching engine provided by the embodiment of the present invention includes the elevator dispatching system 201 as described in the above embodiment, and a simulation unit 202 for simulating the elevator operation conditions under different scenarios of the current dispatching strategy; The strategy adjustment module is also used to optimize the current scheduling strategy according to the elevator operation conditions in different scenarios of the current scheduling strategy simulated by the simulation engine.

[0042] The example of the present invention improves multi-mode simulation and flexible parameter configuration. The elevator simulation engine supports multiple simulation modes such as step simulation and double-speed simulation, which can meet the needs of different users for simulation accuracy and speed. At the same time, the simulation engine supports highly customized parameter configuration. Users can freely set parameters such as building structure, passenger flow, and scheduling algorithm according to the needs of different buildings and different time periods, so as to perform accurate simulation in different scenarios.

[0043] In an embodiment of the present invention, the elevator dispatching engine further includes: Step-by-step simulation mode for analyzing the process of each scheduling decision step by step; Double-speed simulation mode: used to quickly verify the effect of the scheduling scheme under high load conditions; Real-time simulation mode: used to display scheduling status and waiting time in real time.

[0044] In the implementation of the present invention, the simulation engine supports multiple modes of simulation, simulates the operation of elevators in different scenarios, and provides real feedback data for the scheduling algorithm. It supports different operation modes such as step simulation, double-speed simulation, and real-time simulation, and supports dynamic changes in the elevator system and different passenger behavior models. Step simulation is suitable for gradually analyzing the process of each scheduling decision. Users can observe the specific details of each elevator scheduling and understand the optimization logic behind each decision. Double-speed simulation: It is suitable for quickly verifying the effect of the scheduling scheme under high load conditions, and can complete a large number of scheduling tasks in a short time, which is suitable for large-scale testing. Real-time simulation: Real-time simulation provides users with scheduling feedback similar to reality by simulating the actual elevator scheduling process. Users can view important data such as real-time scheduling status and waiting time. The simulation engine module is closely integrated with the scheduling algorithm module. Whenever the simulation engine performs a new scheduling operation, it will pass real-time data to the scheduling algorithm module, including passenger demand, floor request, elevator load, etc., to help the algorithm make the best decision.

[0045] In an embodiment of the present invention, the elevator dispatching engine further includes: An instance creation unit, used to independently configure simulation parameters, scheduling strategies and scenario settings according to the needs of each user, and create a multi-user independent simulation instance according to the simulation parameters, scheduling strategies and scenarios; In the embodiment of the present invention, when each user starts the simulation, the system will create an independent simulation instance for him. Each simulation instance has independent elevator configuration, scheduling strategy, scene settings, etc. Users can freely configure the simulation environment according to their own needs without affecting the simulation process of other users. Support multiple users to run independent simulation instances at the same time. Each user can independently configure simulation parameters, scheduling strategy, scene settings, etc. according to their own needs, and they will not interfere with each other.

[0046] The user management and resource allocation unit is used to independently manage the simulation instance of each user.

[0047] In the embodiment of the present invention, the system manages each user's simulation instance independently to ensure that data and resources in the simulation process do not interfere with each other. The system automatically allocates computing resources according to resource requirements and ensures that the simulation processes of multiple users can run efficiently in parallel.

[0048] In some embodiments of the present invention, after the simulation is completed, the system will generate an independent simulation report for each user, which includes key performance indicators of elevator scheduling, optimization suggestions and other data, and ensures that each user's data is isolated from other users' simulation data to prevent data leakage.

[0049] In traditional simulation systems, it is often difficult to ensure independence when multiple users perform simulations at the same time, resulting in scheduling interference or data conflicts. The embodiments of the present invention use independent simulation instance technology to enable each user to have an independent elevator scheduling simulation environment, supporting multi-user parallel operation without mutual interference. Each user can configure independent simulation parameters and scheduling strategies according to their own needs, and perform personalized optimization.

[0050] The elevator dispatch engine operation steps include: 1. Parameter configuration: Users can configure the parameters required for simulation through a graphical interface or API interface, including the number of floors in the building, the number of elevators, the passenger flow demand on each floor, and the physical parameters of elevator operation (such as acceleration, maximum speed, stop time, etc.). The system automatically loads these configurations and generates the corresponding simulation model. Users can also select simulation modes as needed, such as step simulation or double-speed simulation, so as to perform different simulation verifications in different scenarios.

[0051] 2. Scheduling algorithm selection: You can choose a classic rule-based scheduling algorithm (such as shortest response time first) or a reinforcement learning algorithm. If you choose reinforcement learning, the system will train based on historical data, or users can choose an existing training model. Users can also adjust the scheduling strategy according to specific needs, such as giving priority to energy efficiency, reducing waiting time, etc.

[0052] 3. Start the simulation service: After starting the simulation, the system performs scheduling calculations and displays the elevator's operating status, passenger location, waiting time and other information in real time on the graphical interface. Users can view the elevator's operating status in real time during the simulation process and dynamically adjust the scheduling parameters or algorithms based on the simulation results to further optimize the system's performance.

[0053] 4. Data analysis and result output: When the simulation is completed, the system will automatically generate a detailed analysis report, and users can view data such as elevator utilization efficiency, average waiting time of passengers, energy consumption, etc. The report will also provide optimization suggestions, and users can adjust the scheduling strategy based on the data in the report to ensure that the system achieves the best operating effect in the real scenario.

[0054] 5. Network service call: If integration with external systems is required, users can submit simulation tasks and query results remotely through the API.

[0055] The elevator dispatching engine provided by the embodiment of the present invention supports multi-user parallel simulation. The multi-user independent simulation module can ensure that multiple users can independently and parallelly run simulations, avoid data conflicts, and improve concurrent processing capabilities. It has flexible multi-mode simulation functions, and modes such as step simulation, double-speed simulation, and real-time simulation enable the system to adapt to different testing needs and provide a diverse simulation experience. By building an elevator dispatching simulation platform that supports multiple simulation modes and dynamic parameter configuration, it can efficiently simulate and optimize elevator operation plans, support multi-user concurrent simulation instances, and output optimization results through data analysis functions, providing an efficient and reliable tool for the research and development, verification, deployment, and improvement of elevator dispatching systems. It is suitable for elevator management systems in multi-level, multi-functional large buildings such as modern buildings, commercial complexes, office buildings, hospitals, and shopping malls, especially in high-rise buildings and high passenger flow density scenarios.

[0056] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The memory 330 includes a computer program, an operating system and acquired data, and the processor 310 can call the logic instructions in the memory 330. The electronic device includes an elevator dispatch engine.

[0057] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0058] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An elevator dispatching system, characterized in that: include: Elevator data acquisition module, used to collect real-time elevator operation data; An elevator status recognition module, used to recognize the current elevator status according to the real-time operation data of the elevator; Resource acquisition module, used to obtain the elevator request on each floor as well as the elevator operation system load and elevator resources; The strategy adjustment module is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the elevator request at each floor, the elevator operation system load and the elevator resources.

2. The elevator dispatching system according to claim 1, characterized in that: The policy adjustment module includes: A strategy initial selection unit, including at least two scheduling algorithms, for outputting a current scheduling strategy according to different scheduling algorithms; A reward acquisition unit, used to obtain the reward obtained by each scheduling algorithm after executing an action; The scheduling strategy optimization unit is used to optimize the current scheduling strategy according to the rewards obtained by each scheduling algorithm after executing the action.

3. The elevator dispatching system according to claim 2, characterized in that: The strategy preliminary selection unit includes: The reinforcement learning framework is used for the elevator system to select an action based on the current state at each moment. After executing the action, the system enters a new state. The execution result of each action will give a reward signal, and the scheduling strategy is optimized by continuously iterating and updating the Q value. The action space corresponding to the reinforcement learning framework includes the running direction of the elevator and the scheduling order of the elevator.

4. The elevator dispatching system according to claim 2, characterized in that: The scheduling strategy selection unit also includes: A Q network model, wherein the Q network model includes a neural network, an experience replay mechanism, and a target network; The input of the neural network is the current state of the elevator system, and the output is the Q value of each possible action; Experience replay mechanism: used to store multivariate data sets in an experience pool, where the multivariate data sets consist of historical state-action-reward-next state data. The data in the experience pool is used to provide random training samples during the training of the Q network model. The weight update module is used to periodically update the weights of the target network.

5. The elevator dispatching system according to claim 2, characterized in that: The reward acquisition unit comprises: A positive reward obtaining unit, used to give positive rewards when the dispatching strategy selected by the elevator reduces the waiting time of passengers, or when resources can be used more efficiently within a certain time period; The negative reward acquisition unit is used to give negative rewards if the scheduling strategy causes the passenger to wait too long, the resources are unevenly used, or the elevator runs idle for a time greater than a preset threshold.

6. The elevator dispatching system according to claim 1, characterized in that: The resource acquisition module further includes: acquiring the request priority of the elevator on each floor; The strategy adjustment module is used to dynamically adjust the elevator scheduling strategy according to the current elevator state, the request priority of the elevator on each floor, the request of the elevator on each floor, and the elevator operation system load and elevator resources.

7. An elevator dispatching engine, characterized in that: The elevator dispatching system comprises the elevator dispatching system as claimed in any one of claims 1 to 6, and a simulation unit for simulating the elevator operation under different scenarios of the current dispatching strategy; The strategy adjustment module is also used to optimize the current scheduling strategy according to the elevator operation conditions in different scenarios of the current scheduling strategy simulated by the simulation engine.

8. The elevator dispatch engine according to claim 7, characterized in that: The elevator dispatch engine comprises: Step-by-step simulation mode for analyzing the process of each scheduling decision step by step; Double-speed simulation mode: used to quickly verify the effect of the scheduling scheme under high load conditions; Real-time simulation mode: used to display scheduling status and waiting time in real time.

9. The elevator dispatch engine according to claim 7, characterized in that: The elevator dispatch engine also includes: An instance creation unit, used to independently configure simulation parameters, scheduling strategies and scenario settings according to the needs of each user, and create a multi-user independent simulation instance according to the simulation parameters, scheduling strategies and scenarios; The user management and resource allocation unit is used to independently manage the simulation instance of each user.

10. An electronic device, characterized in that: Comprising an elevator dispatching engine as described in any one of claims 7 to 9.

Images