Target floor reservation type elevator dispatching optimization method and related equipment

Abstract

The embodiment of the invention provides a target floor reservation type elevator dispatching optimization method and related equipment, and belongs to the field of intelligent elevator dispatching control. The method comprises the steps of generating training data of an elevator state and an estimated arrival time index by simulating an elevator system to run Monte Carlo simulation; a neural network model is constructed, supervised learning is carried out with the elevator state as input and the time index as output, and a state-to-time mapping relation is established; performing iterative optimization on the weight of the micro-adaptation layer in the neural network by adopting a genetic algorithm and taking the comprehensive time index as a fitness function to obtain an optimal network parameter; and the optimized lightweight neural network is deployed to edge computing equipment, passenger requests are received in real time, the optimal scheduling scheme is predicted, and elevator operation is controlled. According to the method, the problem that real data scarcity and online calculation real-time performance are difficult to consider at the same time is effectively solved, the average waiting time of passengers is remarkably shortened, and the scheduling efficiency and the response performance of the high-rise building multi-elevator group control system are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent elevator dispatching control, in particular to a destination floor reservation type elevator dispatching optimization method and related equipment. BACKGROUND

[0002] In modern high-rise buildings, the operation efficiency of the elevator system is a key factor affecting user experience and building energy consumption. Traditional elevator dispatching algorithms, such as Group Control System, Shortest Search Floor Time Priority (SCAN) algorithm, etc., are mostly based on fixed rules or simple heuristic strategies. These methods often perform rigidly when facing dynamic, random, and multi-objective passenger flow demands, making it difficult to achieve global optimal scheduling, resulting in long average waiting time for passengers and low efficiency of elevator operation.

[0003] In recent years, with the development of artificial intelligence technology, machine learning methods, especially reinforcement learning (RL) and neural networks (NN), have been tried to be applied in the field of elevator dispatching. However, these methods usually have two major limitations: 1) highly dependent on a large amount of real operation data for model training, while the collection of high-quality data in actual scenarios is costly and involves privacy issues; 2) the computational complexity of many intelligent algorithms is high when performing online inference, making it difficult to meet the real-time response requirements of the elevator dispatching system. SUMMARY

[0004] The main purpose of the embodiments of the present application is to propose a destination floor reservation type elevator dispatching optimization method and related equipment based on genetic algorithm and neural network, which realizes efficient and low-consumption intelligent elevator dispatching through simulation of elevator system simulation and offline optimization.

[0005] To achieve the above purpose, one aspect of the embodiments of the present application proposes a destination floor reservation type elevator dispatching optimization method, which comprises the following steps: Data collection and simulation running step: processing OD data of passenger reservation through simulating the elevator system, simulating the elevator operation, and recording elevator state parameters and estimated arrival time indicators calculated based on a preset dispatching strategy; Neural network training step: constructing a neural network model, taking the recorded elevator state parameters as input features, and taking the estimated arrival time indicators as output labels for supervised learning training, to establish a nonlinear mapping relationship from elevator state to time; Genetic algorithm optimization step: taking the weight parameters of the micro-adaptive layer in the neural network model as optimization variables, taking a comprehensive time indicator function composed of average waiting time and average boarding time as fitness function, and using genetic algorithm for iterative optimization to find the optimal combination of neural network weight parameters; Online scheduling step: deploying the optimized neural network model to the elevator group control system, receiving new passenger requests and current elevator states in real time, predicting the estimated arrival time of each candidate scheduling scheme through the neural network model, and selecting the optimal scheduling instruction to send to the elevator controller for execution.

[0006] In some embodiments, the data collection and simulation running step comprises: constructing a simulation environment containing elevator physical parameters and building structure parameters; inputting historical OD data and running Monte Carlo simulation to simulate the elevator operation process under the nearest distance scheduling strategy; recording the elevator state parameters S in each simulation, including: passenger load, running direction, internal call floor, assigned external call floor, current target floor, external call departure floor, and destination floor; recording the estimated arrival time index t of the external call floor calculated based on the current state through the nearest distance formula.

[0007] In some embodiments, the estimated arrival time index t is a weighted sum function, and its expression is:

[0008] wherein and are weight coefficients.

[0009] In some embodiments, the neural network model is a fully connected neural network structure, the number of input layer nodes corresponds to the dimension of the elevator state features, and the output layer is a single node for outputting the estimated arrival time.

[0010] In some embodiments, the neural network model contains at least one hidden layer and one micro-adaptation layer, and the weight parameters of the micro-adaptation layer are the objects of genetic algorithm optimization.

[0011] In some embodiments, the genetic algorithm optimization step comprises: encoding the weight parameters of the micro-adaptation layer into a real number vector to form a chromosome; initializing the population and setting the evolution generation number, crossover rate, and mutation rate parameters; using the scheduling objective function as the fitness function to evaluate the fitness of each individual in the population; performing selection, crossover, and mutation operations according to the fitness to generate a child population, and using an elite reservation strategy to reserve the best individual of the current generation; iteratively performing the above process until the termination condition is met, and outputting the optimal weight parameter combination.

[0012] In some embodiments, the online scheduling step comprises: newly arrived passenger request with the current state of each elevator combined into a new input feature vector; input the input feature vector into the optimized neural network model, and forward calculate to obtain the estimated arrival time corresponding to the allocation of each elevator for the current new request; compare the estimated arrival times of each candidate scheme, and select the scheme with the smallest time as the optimal scheduling instruction.

[0013] In some embodiments, the specific architecture of the deployment is: the simulation training and optimization process runs on a high-performance computing server; the optimized lightweight neural network model is deployed in the edge computing device of the elevator group control system; the generated scheduling instruction is transmitted to the PLC controller of each elevator through an industrial communication protocol for execution.

[0014] To achieve the above-mentioned purpose, another aspect of the embodiment of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned method when executing the computer program.

[0015] To achieve the above-mentioned purpose, another aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the above-mentioned method.

[0016] To achieve the above-mentioned purpose, another aspect of the embodiment of the present application provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the above-mentioned method The embodiment of the present application at least includes the following beneficial effects: the present application provides a destination layer reservation type elevator scheduling optimization method, electronic device, storage medium and program product, which constructs a simulated elevator system to generate a large amount of simulation data, and optimizes the neural network parameters offline by using a genetic algorithm, and finally deploys a lightweight network for online real-time prediction, so as to significantly reduce the passenger waiting time and improve the scheduling efficiency and system response speed without a large amount of real data. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 a step flow chart of a destination layer reservation type elevator scheduling optimization method in the embodiment of the present application; Figure 2 a genetic algorithm optimization neural network algorithm example diagram in the embodiment of the present application; Figure 3Fig. 1 is an example diagram of an elevator state and estimated time database collected in an embodiment of the present application; Figure 4 Fig. 2 is an example diagram of a neural network structure design in an embodiment of the present application; Figure 5 Fig. 3 is an example diagram of a neural network progress learning process in an embodiment of the present application; Figure 6 Fig. 4 is an example diagram of a simulated elevator system built in an embodiment of the present application; Figure 7 Fig. 5 is an example diagram of a genetic algorithm optimizing neural network process in an embodiment of the present application; Figure 8 Fig. 6 is an example diagram of a comparison of the effect after optimization and the traditional method in an embodiment of the present application; Figure 9 Fig. 7 is a hardware structure schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer and more apparent, the following further describes the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementation described in the following exemplary embodiments does not represent all the implementations consistent with the embodiments of the present application, but is only an example of devices and methods consistent with some aspects of the embodiments of the present application as described in the appended claims.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0020] Before the embodiments of the present application are described in detail, first, some nouns and terms involved in the embodiments of the present application are described, and the nouns and terms involved in the embodiments of the present application are applicable to the following explanations.

[0021] 1) Genetic Algorithm (GA) is an optimization search method that simulates the biological evolution process, and its core idea is derived from Darwin's natural selection theory, which finds the optimal solution of a problem by simulating the genetic, mutation, crossover and selection mechanisms in the biological world 2) Neural network algorithm is a machine learning model that simulates the structure and function of biological neural systems, which realizes complex pattern recognition and prediction functions by simulating the information transmission and weight adjustment mechanism between neurons.

[0022] For example,Figure 1 The embodiment shown provides a destination layer reservation type elevator scheduling optimization method, including the following steps: S1, data acquisition and simulation running step. Collect the OD (Origin-Destination) data of passenger reservation through the man-machine interaction terminal deployed on each floor. Build a simulated elevator system containing elevator physical parameters (such as speed, acceleration, and load capacity) and building structure parameters (such as the number of floors and floor height). Input the historical OD data into the system, run Monte Carlo simulation, simulate the elevator running process under the nearest distance scheduling strategy, and record the real-time state parameters S of the elevator in each simulation and the estimated arrival time index t of the outer call floor calculated by the nearest distance formula.

[0023] The elevator state parameters S include but are not limited to: passenger capacity, running direction, inner call floor, allocated outer call floor, current target floor, outer call departure floor, and destination floor. The estimated arrival time index t is a comprehensive performance index, and its expression is:

[0024] Where α and β are adjustable weight coefficients for balancing the emphasis of the scheduling strategy on different indicators.

[0025] S2, neural network training step: build a deep neural network model, preferably a fully connected network structure. Use the large amount of data pairs (S, t) recorded in step S1 as the training data set, with the elevator state parameters S as the input features and the estimated arrival time index t as the output label, to supervise the learning and training of the neural network. This step aims to establish an accurate nonlinear mapping relationship from high-dimensional elevator state to comprehensive time index .

[0026] S3, genetic algorithm optimization step: extract the weight parameters of the "micro-adaptation layer" in the trained neural network as optimization variables. Define the fitness function as the scheduling objective function . Use genetic algorithm (GA) to perform iterative global search in the weight parameter space, and through selection, crossover, and mutation operations, find the optimal weight parameter combination that minimizes the fitness function J. This step further improves the accuracy and generalization ability of the neural network model.

[0027] S4, online scheduling step: deploy the lightweight neural network model optimized by the genetic algorithm and with the best performance to the edge computing device of the elevator group control system. When the system is running in real time, the newly arrived passenger request and the state of each elevator The new input feature vectors are combined and input into the neural network for forward calculation to quickly obtain the estimated arrival times corresponding to different elevators allocated for the request. The system compares the estimated times and selects the candidate scheme with the minimum time as the optimal scheduling instruction, which is sent to the corresponding elevator PLC controller for execution through an industrial communication protocol (such as Modbus, Profibus).

[0028] Next, the scheme of the embodiments of the present application will be described in detail in conjunction with the drawings and specific application examples. The embodiment provides a destination reservation type elevator scheduling optimization method based on a genetic algorithm and a neural network, and a specific implementation manner comprises the following steps: Step 1, data acquisition and preprocessing: OD data (departure floor-destination floor) of passenger reservation is collected through a man-machine interaction terminal on each floor in the building, and the data format is , wherein is the arrival time of the passenger, is the departure floor, is the destination floor.

[0029] Step 2, simulate the elevator system: A simulated elevator system containing elevator physical parameters (such as load, speed, acceleration) and building structures (such as floor height, elevator shaft number) is constructed.

[0030] Step 3, record data: After inputting the historical OD data, run the Monte Carlo simulation to simulate the elevator running state under different scheduling strategies (including the number of passengers, running direction, internal call floor, allocated external call floor, current target floor, external call departure floor and destination floor), and record the estimated arrival time of each floor calculated by the nearest distance formula .

[0031] Step 4, neural network training: A fully connected neural network containing two hidden layers and a micro-adaptive layer is designed, the input layer is the elevator state feature , and the output layer is the estimated arrival floor time . The simulation generated data set is used for supervised learning.

[0032] Step 4, genetic algorithm optimization: The weights in the neural network are used as optimization variables, and the optimization variables are limited in the “micro-adaptive layer”, and the optimization process is as follows Figure 2 ​As shown, the genetic algorithm (GA) is used to optimize the neural network parameters, the core of which is to encode the network weights into a real vector (i.e. "chromosome"), each such vector represents a "solution" to be evaluated. The goal of the genetic algorithm is to find the individual with the best performance in the solution space, and its merits are measured by the fitness value, and the fitness function is defined as the passenger average time index. The algorithm steps are as follows: first, initialize the maximum evolution generation, population size, crossover rate and mutation rate, etc. Parameters and randomly generate an initial population containing multiple individuals; then enter the iterative evolution process, in each generation, first calculate the fitness of each individual in the population, then select the parent individual according to the fitness, and randomly apply crossover (exchange part of the gene) or mutation (random disturbance to the gene) operation to produce offspring, form a new population, while recording and updating the global optimal solution of the current generation; After the loop ends, the final optimal weight combination is decoded into neural network parameters, thus obtaining the optimized model.

[0033] Step 6, online scheduling decision: The optimized neural network is deployed to the edge computing device of the elevator group control system. Real-time reception of new passenger requests , combined with the current elevator state , quickly predict the performance indicators of each candidate scheduling scheme through the neural network, and select the optimal scheduling instruction. Through the communication protocol, the scheduling instruction is transmitted to the PLC controller of each elevator, and the elevator operation strategy (such as stopping floor, running direction) is dynamically adjusted.

[0034] (1) Specific test This embodiment is tested in the elevator group control system of a 20-story office building, Figure 6 is the visual interface built in the simulation elevator system operation, the specific configuration is as follows: 1) Number of elevators: 4, rated load 1000 kg, speed 2 m / s.

[0035] 2) Simulated elevator system: running on a high-performance server (Intel Xeon Gold 6248R).

[0036] 3) Neural network model: input layer 12 nodes (elevator state features), hidden layer 24 nodes, output layer single node.

[0037] The implementation process is as follows: 1) Collect 16 sets of OD data to construct the training set, simulate 9.6 million sets , ) data, as shown in Figure 3 ; 2) Train the initial neural network, the Loss loss value is close to 0, as shown in Figure 5as shown.

[0038] 3) Run the simulation elevator system, use the neural network as the scheduling module, like Figure 4 as shown.

[0039] 4) Optimize the neural network using the genetic algorithm, and the optimal network is constantly optimized iteratively, like Figure 7 as shown.

[0040] 5) Optimize the neural network using the genetic algorithm, and the average waiting time of passengers is reduced by 12.2% compared with before optimization, and the average waiting time of passengers is reduced by 22.9% compared with the traditional scheduling algorithm, like Figure 8 as shown.

[0041] In summary, the present application generates training data by simulating an elevator system, and realizes the scheduling optimization of the external call signal by combining the neural network and the genetic algorithm, which significantly improves the efficiency and energy saving of the elevator system while ensuring real-time. This method can be flexibly adapted to different sizes of elevator group control scenes, and has wide engineering application value.

[0042] (2) Advantages Compared with the prior art, the method of the embodiment has the following advantages and beneficial effects: 1) Training data is generated by a simulation system, perfectly solving the problem of dependence of intelligent scheduling algorithms on real historical data, with low data cost, large scale, and full coverage of scenarios, which can provide comprehensive simulation data support for elevator scheduling, and ensure stable operation of the system in different actual scenarios.

[0043] 2) Offline optimization ensures high precision of the elevator scheduling model, and online real-time prediction guarantees the real-time response ability of the system, perfectly balancing the scheduling quality and real-time requirements.

[0044] 3) A multi-objective optimization framework is adopted, which can flexibly balance key indicators such as waiting time and boarding time, optimize system performance, and meet the scheduling needs of different application scenarios to improve the comprehensive benefits of elevator scheduling. Specifically, the architecture of "offline optimization (GA+NN), online prediction (lightweight NN)" takes into account the high precision of the scheduling model and the high real-time performance of online decision-making.

[0045] 4) The comprehensive time index and genetic algorithm optimization framework support multi-objective optimization, and can flexibly adapt to different scheduling requirements of different building scenarios (such as office buildings, hospitals, and residences) by adjusting the weight coefficients (α, β). Whether it is a small commercial building or a large office building, the elevator scheduling system of the present application can be flexibly deployed according to actual needs, and has good scalability and adaptability.

[0046] 5) Deployment advantage: the system architecture design is reasonable, the training optimization process is completed on a high-performance server, and finally only a lightweight neural network needs to be deployed to the edge side, the modification of the existing elevator group control system is small, the implementation is convenient, and good engineering application prospects are provided.

[0047] 6) Compared with the traditional method, the passenger waiting time can be reduced by 22.9% or more, and the scheduling time is within 200ms: the method greatly improves the speed and accuracy of elevator scheduling through intelligent optimization and efficient algorithm, and significantly improves the waiting experience of passengers.

[0048] The embodiment of the application further provides an electronic device, which comprises a memory and a processor. The memory stores a computer program, and the processor implements the method described above when executing the computer program. The electronic device can be any intelligent terminal, such as a tablet computer or a vehicle-mounted computer.

[0049] It can be understood that the contents in the above method embodiments are applicable to the device embodiments, the device embodiments specifically implement the same functions as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0050] Please refer to Figure 9 , Figure 9 The hardware structure of the electronic device of another embodiment is illustrated, and the electronic device comprises: The processor 901 can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute related programs to implement the technical solutions provided by the embodiments of the application. The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the application are implemented by software or firmware, the related program codes are stored in the memory 902 and called and executed by the processor 901 to implement the above method of the embodiments of the application. The input / output interface 903 is used to realize information input and output. The communication interface 904 is used to realize the communication interaction between the device and other devices. The communication can be realized in a wired manner (for example, USB, network cable, etc.), or in a wireless manner (for example, mobile network, WIFI, Bluetooth, etc.). a bus 905 that transmits information between various components (e.g., the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device; The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are communicatively connected with each other within the device through the bus 905.

[0051] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method.

[0052] It can be understood that the contents in the above method embodiments are all applicable to the present storage medium embodiment, the present storage medium embodiment specifically implements the functions of the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0053] The memory, as a non-transitory computer readable storage medium, can be used to store non-transitory software programs and non-transitory computer executable programs. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0054] The embodiment of the present application further provides a computer program product, which includes a computer program, and the computer program is executed by a processor to implement the method.

[0055] It can be understood that the contents in the above method embodiments are all applicable to the present program product embodiment, the present program product embodiment specifically implements the functions of the above method embodiments, and achieves the same beneficial effects as the above method embodiments. The executable computer program code or "code" for executing various embodiments can be written in a high-level programming language such as C, C++, Python, Smalltalk, Java, JavaScript, Visual Basic, Structured Query Language (for example, Transact-SQL), Perl, or in various other programming languages.

[0056] The embodiments described in the specification are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0057] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0058] The device embodiments described above are merely illustrative, and the units described as separate components can or can not be physically separated, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0059] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0060] The terms "first", "second", "third", "fourth" and the like (if any) in the specification of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0061] It should be understood that, in the application, "at least one" refers to one or more, and "multiple" refers to two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases of only A, only B, and A and B existing at the same time, wherein A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, and c can be single or multiple.

[0062] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the above units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed mutual units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0063] The units described above as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0064] In addition, the functional units in each embodiment of the application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0065] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0066] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A destination floor reservation type elevator dispatching optimization method characterized by comprising: The method comprises the following steps: A data collection and simulation running step: processing OD data of passenger reservations by simulating an elevator system, simulating elevator operation, recording elevator state parameters and estimated arrival time indicators calculated based on a preset scheduling strategy; A neural network training step: constructing a neural network model, taking the recorded elevator state parameters as input features, and taking the estimated arrival time indicators as output labels for supervised learning training to establish a nonlinear mapping relationship between elevator states and time; A genetic algorithm optimization step: taking the weight parameters of the micro-adaptive layer in the neural network model as optimization variables, taking a comprehensive time indicator function composed of average waiting time and average boarding time as a fitness function, and using a genetic algorithm for iterative optimization to find the optimal combination of neural network weight parameters; An online scheduling step: deploying the optimized neural network model to an elevator group control system, receiving new passenger requests and current elevator states in real time, predicting estimated arrival times of each candidate scheduling scheme through the neural network model, and selecting the optimal scheduling instruction to send to the elevator controller for execution.

2. The method of claim 1, wherein, The data collection and simulation running step comprises: Constructing a simulation environment containing elevator physical parameters and building structure parameters; Inputting historical OD data, running Monte Carlo simulation, and simulating elevator operation process under the nearest distance scheduling strategy; Recording elevator state parameters S in each simulation, including passenger capacity, running direction, internal call floor, assigned external call floor, current target floor, external call departure floor, and destination floor; Recording the time indicator t of the estimated arrival external call floor calculated based on the current state through the nearest distance formula.

3. The method of claim 2, wherein, The estimated arrival time indicator t is a weighted sum function, and its expression is: wherein and are weight coefficients.

4. The method of claim 1, wherein, The neural network model is a fully connected neural network structure, the number of input layer nodes corresponds to the dimension of elevator state features, and the output layer is a single node for outputting the estimated arrival time.

5. The method according to claim 1 or 4, characterized in that, The neural network model contains at least one hidden layer and one micro-adaptive layer, and the weight parameters of the micro-adaptive layer are the objects of genetic algorithm optimization.

6. The method of claim 1, wherein, The genetic algorithm optimization step comprises: Encoding the weight parameters of the micro-adaptive layer into a real number vector to form a chromosome; Initializing the population, setting the evolution generation number, crossover rate, and mutation rate parameters; with the scheduling objective function evaluating the fitness of each individual in the population as a fitness function; Selecting, crossing, and mutating according to the fitness to generate a child population, and using an elite reservation strategy to reserve the best individual of the current generation; Iteratively executing the above process until the termination condition is met, and outputting the optimal weight parameter combination.

7. The method of claim 1, wherein, The online scheduling step comprises: combining the new arriving passenger request with the current state of each elevator into a new input feature vector; Inputting the input feature vector into the optimized neural network model to forward calculate the estimated arrival time corresponding to the allocation of each elevator for the current new request; Comparing the estimated arrival times of each candidate scheme to select the scheme with the smallest time as the optimal scheduling instruction.

8. The method of claim 1, wherein, The specific architecture of the deployment is: The simulation training and optimization process runs on a high-performance computing server; The optimized lightweight neural network model is deployed in the edge computing device of the elevator group control system; The generated scheduling instruction is transmitted to the PLC controller of each elevator for execution through an industrial communication protocol.

9. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method in any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1-9. The computer program is executed by the processor to implement the method in any one of claims 1 to 8.

Images