A destination layer elevator group control scheduling method, device and system

Abstract

The application relates to a destination floor elevator group control scheduling method, device and system. The scheduling method comprises the following steps: obtaining a destination floor call request set; collecting the current state of each elevator in the elevator group, and obtaining an available elevator set according to the current state of the elevator and the destination floor call request set; obtaining a dispatching scheme according to the destination floor call request set, the available elevator set and historical operation data; issuing a dispatching instruction to a corresponding elevator controller and feeding back dispatching information to a destination floor call request input module according to the dispatching scheme. The destination floor elevator group control scheduling method, device and system have the advantages of improving the rationality of elevator traffic mode weight parameter setting, enabling the system to continuously adapt to various call scenarios and the like.

Description

Technical Field

[0001] This invention relates to the field of elevator control technology, and in particular to a method, device and system for group control and scheduling of elevators at the destination floor. Background Technology

[0002] With the continuous increase in pedestrian density in super high-rise office buildings, commercial complexes, and industrial parks, the operational efficiency of elevator systems and passenger experience have become important indicators of building transportation systems.

[0003] In traditional elevator group control systems, passengers typically register their direction (up or down) outside the hall and then select their destination floor after entering the elevator car. However, because the system only has directional information and not the specific destination floor or the number of passengers traveling together, problems can easily arise, such as multiple passengers going to different floors sharing the same car, frequent elevator stops, multiple elevators competing for the same floor, and passengers accidentally pressing up or down buttons, disrupting the dispatch logic. Furthermore, in actual project operation, this typical group control logic leads to increased waiting and riding times during peak hours, difficulty in controlling elevator overcrowding, increased energy consumption, and may result in complaints about registered calls not stopping as expected or malfunctioning.

[0004] To address this, existing technologies have developed destination floor group control systems. These systems use a destination floor selector outside the hall to allow passengers to input their destination floor while waiting for the elevator. This allows the system to obtain the starting floor, destination floor, number of passengers, and special service requests in advance, thereby concentrating passengers going to the same or adjacent destination floors into suitable elevators, reducing the number of stops and improving traffic efficiency.

[0005] However, existing destination-level group control solutions still have shortcomings. First, some systems rely mainly on fixed rules, preset schedules, or weight parameters obtained through offline training. These fixed parameters struggle to adapt continuously to time-varying scenarios such as meeting endings, lunch rushes, concentrated visitor arrivals, backup power activation, and sudden elevator calls on specific floors. Second, many elevator dispatch strategies primarily optimize the waiting time or shortest distance for the current request, without fully assessing the impact of the current dispatch decision on passenger flow service capacity in the next few tens of seconds to minutes. Third, the stripping, reallocation, and data closure of already assigned requests in abnormal scenarios are insufficient, making it difficult to support embedded algorithm verification and continuous optimization. Summary of the Invention

[0006] Based on this, the purpose of the present invention is to provide a method, device and system for controlling and scheduling elevator groups at the destination floor, which can improve the rationality of the elevator traffic mode weight parameter setting, enable the system to continuously adapt to various elevator calling scenarios, and take into account the elevator group service capacity for a period of time in the future, such as tens of seconds to several minutes, significantly reduce the risk of long waiting time in the future, and reduce multiple beneficial effects such as invalid return trips and start-stop energy consumption.

[0007] A method for group control and scheduling of elevators to a destination floor includes the following steps: Obtain the set of call requests for the destination floor; Collect the current status of each elevator in the elevator group, and obtain the set of available elevators based on the current status of the elevator and the set of elevator call requests for the destination floor; Based on the set of elevator call requests for the destination floor, the set of available elevators, and historical operation data, an elevator dispatch plan is obtained; According to the elevator dispatching scheme, an elevator dispatching command is sent to the corresponding elevator controller, and elevator dispatching information is fed back to the elevator call request input module of the destination floor; The step of obtaining an elevator dispatch plan based on the destination floor call request set, the available elevator set, and historical operation data includes: Based on the set of elevator call requests for the destination floor and the set of available elevators, candidate elevator dispatch schemes are obtained; Based on the set of elevator call requests for the destination floor within the current time window and the corresponding historical operation data, the current elevator traffic mode is determined, and the elevator traffic mode parameter weights are obtained according to the elevator traffic mode; wherein, different elevator traffic modes have different elevator traffic mode parameter weights. Based at least on the weights of the elevator traffic mode parameters, the dispatch generation value corresponding to each candidate dispatch scheme is obtained, and the candidate dispatch scheme corresponding to the minimum dispatch generation value is selected as the final dispatch scheme.

[0008] Compared to existing technologies, the solution of this invention identifies elevator traffic patterns based on the destination floor call request data and corresponding historical operation data within the current time window, and adjusts the corresponding elevator traffic pattern weights according to the elevator traffic pattern, rather than using fixed elevator traffic pattern weight parameters. This improves the rationality of parameter settings and enables the system to continuously adapt to various call scenarios.

[0009] Furthermore, the elevator traffic mode parameter weights include the weights for estimated waiting time, estimated riding time, estimated number of new stops, estimated number of floors to be bypassed by empty passengers, estimated capacity congestion penalty, and estimated special task penalty / reward.

[0010] Furthermore, the step of obtaining an elevator dispatch plan based on the destination floor call request set, the available elevator set, and historical operation data also includes: Based on historical operational data, a set of future passenger flow scenarios within a preset time period is obtained; each future passenger flow scenario in the set of future passenger flow scenarios includes a set of future destination floor elevator call requests and a passenger flow scenario probability. Based on the elevator traffic mode parameter weights and the set of future passenger flow scenarios, the dispatch value corresponding to each candidate dispatch scheme is obtained.

[0011] Furthermore, the dispatch value includes the current dispatch value and the expected value of future passenger flow scenarios, and the dispatch value Z is obtained through the following formula: Z = Sum(C _r_k × x _r_k ) + beta × Sum(p_omega × Q(x, omega)) Among them, C _r_ k represents the current dispatch cost of assigning the destination floor call request r to elevator k; x _r_k This represents the binary elevator allocation decision variable, where x is the variable that determines when a call request r to the destination floor is allocated to elevator k. _r_k =1, otherwise x _r_k =0; beta represents the weight of the expected cost of the future passenger flow scenario; p_omega represents the probability of the future passenger flow scenario omega; Q(x, omega) represents the remaining service cost generated by the future passenger flow scenario omega under the current elevator dispatch decision x.

[0012] Furthermore, the step of obtaining an elevator dispatch plan based on the destination floor call request set, the available elevator set, and historical operation data also includes: Based on historical operational data, a set of future passenger flow scenarios within a preset time period is obtained; each future passenger flow scenario in the set of future passenger flow scenarios includes a set of future destination floor elevator call requests and a passenger flow scenario probability. Based on the set of future passenger flow scenarios and their respective elevator traffic modes, determine the elevator resource reservation strategy; Based on the elevator traffic mode parameter weights, the future passenger flow scenario set, and the elevator resource reservation strategy, the dispatch value corresponding to each candidate dispatch scheme is obtained.

[0013] Furthermore, the dispatch value includes the current dispatch value, the expected value of future passenger flow scenarios, and the penalty value for insufficient elevator resource reservation on key floors in future passenger flow scenarios, as shown in the following formula: Z = Sum(C _r_k × x _r_k ) + beta × Sum(p_omega × Q(x, omega)) + gamma ×R _loss (x) Among them, C _r_ k represents the current dispatch cost of assigning the destination floor call request r to elevator k; x _r_k This represents the binary elevator allocation decision variable, where x is the variable that determines when a call request r to the destination floor is allocated to elevator k. _r_k =1, otherwise x_r_k =0; beta represents the weight of the expected cost of future passenger flow scenarios; p_omega represents the probability of passenger flow scenario omega in the future passenger flow scenario; Q(x, omega) represents the remaining service cost generated by future passenger flow scenario omega under the current elevator dispatch decision x; gamma represents the penalty weight for insufficient elevator resource reservation on key floors; R _loss (x) represents the penalty for insufficient elevator resource reservation on key floors in future passenger flow scenarios under the current elevator dispatch decision x.

[0014] Furthermore, the current dispatch cost C _r_ k is obtained using the following formula: C _r_k = A1× W _r_k + A2× J _r_k + A3× N _r_k + A4× D _r_k + A5× H _r_k + A6× P _r_k Among them, W _r_k Indicates the estimated arrival time at the starting floor, J _r_k Indicates the estimated elevator travel time, N _r_k Indicates the number of new stops, D _r_k Indicates the number of floors to bypass when unloaded, H _r_k P indicates a penalty for overcrowding in the elevator car. _r_k The numbers represent special task penalty and reward items; A1 to A6 represent the weights of estimated waiting time, estimated riding time, estimated number of new stops, estimated number of floors to be detoured when empty, estimated capacity congestion penalty, and estimated special task penalty and reward, respectively.

[0015] Furthermore, the scheduling method also includes: The system acquires any abnormal situations that occur during the execution of the elevator dispatching scheme, and acquires an abnormal situation response elevator dispatching scheme based on these abnormal situations.

[0016] In addition, the present invention also provides a destination floor elevator group control and scheduling device for executing the aforementioned destination floor elevator group control and scheduling method.

[0017] The present invention also provides a destination floor elevator group control and scheduling system, which includes multiple elevator units and the aforementioned elevator destination floor group control and scheduling device; wherein, each elevator unit includes a car, a car drive device, an elevator controller, and a destination floor call request input module; the car is connected to the car drive device, and the car drive device is connected to the elevator controller; the destination floor call request input module and the elevator controller are respectively connected to the elevator destination floor group control and scheduling device.

[0018] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the target floor elevator group control and scheduling method described in this invention; Figure 2 This is a flowchart illustrating step S3 of obtaining the elevator dispatch scheme in the target floor elevator group control and scheduling method described in this invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention more readily understood by those skilled in the art, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] refer to Figure 1 , Figure 1 A schematic diagram of the target floor elevator group control and scheduling method described in this invention is shown.

[0022] In this invention, the scheduling method is triggered by a set allocation cycle and an event triggering method. Specifically, the scheduling method is started when a set triggering event occurs. The triggering event is selected from: receiving a new elevator call request generated by the destination floor selector, a change in elevator status, or an abnormal situation where service cannot continue. This abnormal situation includes malfunction, full load, timeout failure to respond, door malfunction, communication malfunction, entry into VIP exclusive service, entry into fire service, and limitation by backup power supply. As long as one triggering event occurs, it is sufficient. Alternatively, if no triggering event occurs, it can also be started when the allocation cycle is reached, thereby updating the traffic mode, short-term prediction results, elevator waiting layout, and resource reservation strategy, etc.

[0023] The elevator group control and scheduling method of the present invention includes the following specific steps: S1. Obtain the set of elevator call requests for the destination floor.

[0024] In this step, passengers input their elevator information in the waiting hall using a touch-screen or button-type destination floor selector. Based on this information, a corresponding destination floor call request is generated, thus forming a destination floor call request set. The destination floor call request includes the starting floor f. _start , destination floor f _dest The request includes the request time t and the number of people traveling with the same person n. Optionally, the elevator call request for the destination floor may further include one or more of the following service identification information: VIP service, disability service, access control reservation, conference elevator call, fire alarm linkage, etc.

[0025] In other embodiments, the source device of the destination floor elevator call request, namely the destination floor elevator call request input module, may come from the destination floor selector outside the lobby, or from the access control system, visitor reservation system, mobile APP, building management system, or multimodal signal acquisition device.

[0026] In this invention, compared with existing elevator group control methods that only obtain the upward or downward direction, this step can obtain multi-dimensional information such as the origin and destination and the number of people waiting before dispatching the elevator, providing a comprehensive data foundation for subsequent merging and allocation of the same or adjacent destination floors, reducing the number of stops, and controlling the congestion of the elevator car.

[0027] S2. Collect the current status of each elevator in the elevator group, and obtain the set of available elevators based on the current status of the elevator and the set of elevator call requests for the destination floor.

[0028] In this step, the current elevator state S of each elevator k in the elevator group is collected in real time. _k The elevator is currently in state S. _k This includes the elevator's current location, direction of travel, car load, remaining capacity, door status, registered floors, current service sequence to be executed, set of floors that can be served, fault status, full load status, backup power supply operation status, and whether it is in special modes such as VIP, fire protection, or maintenance.

[0029] Then, based on the current state S of the elevator _k Combine the set of elevator call requests to the destination floor to generate a set of available elevators K. _available If the elevator is in the first state, it will not be considered as a candidate elevator in the set of available elevators, and will therefore not participate in the allocation of all currently received destination floor call requests. The first state includes, but is not limited to, fault, full load, fire, and maintenance. If the elevator is in the second state, it will only be considered as a candidate elevator in the set of available elevators if the destination floor call request set includes a corresponding service request. The second state includes, but is not limited to, VIP exclusive.

[0030] S3. Obtain an elevator dispatch plan based on the destination floor elevator call request set, the available elevator set, and historical operation data.

[0031] In this invention, this step specifically includes the following steps: S31. Based on the set of elevator call requests for the destination floor, the set of available elevators, and the constraints, obtain candidate elevator dispatch schemes.

[0032] Specifically, a mixed-integer programming solver can be used to obtain candidate elevator dispatch schemes. Alternatively, when embedded resources are limited, heuristic search, genetic algorithms, co-evolutionary algorithms, ant colony algorithms, reinforcement learning value function approximations, or combinations thereof can be used for approximate solutions to adapt to different embedded hardware computing power and obtain candidate elevator dispatch schemes. Alternatively, all feasible candidate elevator dispatch schemes can be obtained through enumeration. Preferably, the elevator call requests to be assigned to the target floor can be pre-grouped according to floor proximity and direction consistency. Then, the pre-grouped elevator call requests to be assigned to the target floor can be assigned to elevators using the above methods to obtain candidate elevator dispatch schemes, thereby improving allocation efficiency and rationality, and meeting millisecond-level real-time requirements.

[0033] When obtaining candidate elevator dispatch schemes, the service path of each destination floor call request and the corresponding candidate elevator in the available elevator set can be obtained. This service path is not simply from the elevator's current position to the starting floor, but rather the estimated arrival time W at the starting floor corresponding to the destination floor call request, based on the elevator's current pending service sequence, the newly added starting floor, the newly added destination floor, the direction of travel, and the stopping order. _r_k 1. Estimated elevator travel time J _r_k , Number of new stops N _r_k Number of floors to bypass when unloaded (D) _r_k Congestion penalty H _r_k Subsequently, the service path selection constraints include, but are not limited to: each destination floor call request must be served by one and only one elevator; the sum of the number of passengers already in any elevator and the number of newly assigned passengers must not exceed the capacity limit; elevators must meet the logic of zoning, single / double floors, high / low zones, and operating direction; the allocation relationship displayed to passengers and without any abnormalities remains stable within the tolerance threshold; conditions such as VIP, fire protection, disabled, full-load direct travel, backup power restriction, and elevator service floors must be met; the waiting time for any destination floor call request must not exceed the set maximum waiting time threshold, or if the maximum waiting time threshold is exceeded, a high penalty (increased priority) or forced entry into a feasible solution will be imposed. In this invention, the setting of these constraints enables the coordinated operation of engineering functions and algorithm optimization.

[0034] In addition, other constraints can be set, including restrictions on the number of elevators, such as minimum number of elevators and maximum number of elevators. Only solutions that meet the elevator quantity requirements will be considered as candidate elevator dispatch solutions.

[0035] S32. Based on the set of elevator call requests for the destination floor within the current time window and the corresponding historical operation data, determine the current elevator traffic mode and obtain the elevator traffic mode parameter weights according to the elevator traffic mode.

[0036] In this invention, the current elevator traffic mode is first determined based on the historical operating characteristics obtained from historical operating data and the current operating characteristics obtained from the destination floor call request set.

[0037] The elevator traffic modes include peak upward movement, peak downward movement, midday balance, normal mode, off-peak energy-saving mode, and instantaneous peak mode. In specific embodiments, depending on the building type, the elevator traffic modes may further include one or more of the following: meeting mode, lunch mode, visitor mode, and parking lot linkage mode.

[0038] The historical operational data is a passenger flow dataset organized by time period, work / holiday, weekday, and the relationship between the origin and destination floors. From this dataset, multi-dimensional historical operational characteristics such as passenger arrival rate, number of elevator calls (both up and down), average waiting time, average travel time, occupancy rate, long waiting rate, and number of stops can be statistically obtained for different floors within the corresponding time period. In a specific embodiment, based on the historical operational data, the historical origin-destination passenger flow arrival rate matrix Lambda(t) corresponding to the current time window can be obtained, and the historical operational characteristics data can be obtained from this matrix. The historical origin-destination passenger flow arrival rate matrix Lambda(t) can be obtained through a non-homogeneous Poisson process, kernel density estimation, Gaussian mixture model, machine learning classification, regression model, or a combination thereof. Of course, in other embodiments, the origin-destination passenger flow arrival rate matrix Lambda(t) can also be obtained using spatiotemporal prediction models based on long short-term memory networks, Transformer time series models, or graph neural networks, as long as the origin-destination passenger flow distribution within the future time window can be output.

[0039] At the same time, based on the elevator call requests for the destination floor within the current time window, multi-dimensional current operation characteristic data such as the current floor call density, floor start-end matrix, long waiting ratio, full load rate, and elevator group spatial distribution can be obtained.

[0040] In a specific embodiment, the current elevator traffic mode is determined based on historical and current operating characteristics. This includes calculating the feature distances between the historical and current operating characteristics and the standard features of each elevator traffic mode, then weighting and fusing the two feature distances, and selecting the traffic mode with the smallest distance as the current elevator traffic mode. Alternatively, in other embodiments, other methods can be used to determine the current elevator traffic mode. For example, the distribution of destination floor call requests within a preset time period (e.g., within 30 seconds) in multiple historical periods of the same type can be statistically analyzed, arranged in ascending order, and a certain quantile can be used as a threshold. The number of requests within the current preset time window can then be determined to see if it exceeds this threshold for traffic mode identification.

[0041] After determining the current elevator traffic mode, the corresponding elevator traffic mode parameter weights are further determined based on the elevator traffic mode. These elevator traffic mode parameter weights include waiting time weight A1, riding time weight A2, number of new stops weight A3, number of floors to detour when empty weight A4, capacity congestion penalty weight A5, and special task penalty reward weight A6. In this invention, different elevator traffic modes have different weights. Specifically, the weights of each element in the normal elevator traffic mode are used as the baseline weights: waiting time weight A1, capacity congestion penalty weight A5, and special task penalty reward weight A6 are increased, while the weight of empty detour floors A4 is appropriately reduced to prioritize improving the response speed of passengers at the entrance level and the car's carrying efficiency. In the downward peak mode, due to the concentrated downward demand from higher floors to the exit level, the waiting time weight A1, riding time weight A2, additional stop weight A3, and capacity congestion penalty weight A5 are increased to reduce the waiting time of downward passengers and unnecessary stops along the way. In the midday balanced traffic mode, due to the simultaneous existence of upward and downward and inter-floor requests, the riding time weight A2 and additional stop weight A5 are increased. The weights of A3 and A5 for capacity congestion penalty are increased, and the weight of A6 for special task penalty reward is appropriately increased to balance the response priority of requests from different directions and floors. In the off-peak energy-saving mode, since there are fewer elevator requests, the weights of A1 for waiting time and A5 for capacity congestion penalty are reduced, while the weights of A3 for the number of new stops, A4 for the number of empty detour floors, and A6 for special task penalty reward are increased to reduce unnecessary starts and stops, empty detours, and the number of elevators in use. In the instantaneous peak mode, since there are sudden requests on a certain floor or some floors in a short period of time, the weights of A1 for waiting time, A5 for capacity congestion penalty, and A6 for special task penalty reward are increased, and a long waiting time penalty item is added through the special task penalty reward item P_r_k, thereby increasing the scheduling priority of sudden floor requests.

[0042] S33. Based on historical operational data, obtain a set of future passenger flow scenarios within a preset time period.

[0043] In this invention, based on historical operational data, a passenger arrival rate matrix Lambda(t) can be obtained for a future preset time period (e.g., H seconds or M minutes, which can be set to 120 seconds in a specific embodiment) starting from the current decision time. Here, the matrix element lambda(t)_ij represents the expected passenger arrival rate from floor i to floor j at future time t. Then, based on this passenger arrival rate matrix Lambda(t), multiple future passenger flow scenarios omega are obtained within the future preset time period, thus obtaining a future passenger flow scenario set Omega. The future passenger flow scenario includes a set of future destination floor elevator call requests and a passenger flow scenario probability p_omega. The future destination floor elevator call request includes the starting floor, destination floor, request time, number of companions, and optionally, one or more of the following: VIP service information, disability service information, access control reservation information, and meeting elevator call information.

[0044] Preferably, to ensure millisecond-level real-time computing capabilities, multiple representative future passenger flow scenarios can be obtained from the generated multiple future passenger flow scenarios, and these representative future passenger flow scenarios constitute the future passenger flow scenario set, thereby achieving the reduction of future passenger flow scenarios. Specifically, representative future passenger flow scenarios can be obtained based on one or more of the following: passenger flow scenario probability, origin-destination distribution similarity, elevator waiting risk, peak floor coverage, and elevator resource occupancy impact. The elevator waiting risk refers to the probability or severity of long waiting times for passengers in future passenger flow scenarios, which can be determined by weighted summation of parameters such as estimated waiting time, the number of requests exceeding the longest waiting threshold, the proportion of long-waiting passengers, and the lack of available elevators on key floors. The peak floor coverage refers to whether the retained future passenger flow scenarios cover key floors where passenger flow peaks may occur within a preset future time period. These peak floors and key floors typically include lobbies, base stations, underground parking garage entrances, conference floors, restaurant floors, and floors with concentrated descent from high-rise areas. The impact of elevator resource occupation refers to the degree of occupation of elevator operating resources in future passenger flow scenarios and the impact of this occupation on the current elevator dispatch decision. It can be determined by weighted summation of parameters such as the number of elevators expected to be occupied in the future passenger flow scenario, occupation time, number of stops, car capacity, running distance, and whether it will cause key floors to lack available elevators.

[0045] In this invention, there is no limit to the number of representative future passenger flow scenarios. In specific embodiments, 10 to 20 representative future passenger flow scenarios can be retained.

[0046] S34. Determine the elevator resource reservation strategy based on the set of future passenger flow scenarios and the corresponding elevator traffic modes.

[0047] In this step, based on the future passenger flow scenarios in the future passenger flow scenario set and the elevator traffic mode determined above, an elevator reservation resource strategy is determined. The elevator reservation resource strategy includes at least the key floor set, the number of reserved elevators, the reserved or return target floors, and the reserved time window.

[0048] Specifically, based on the current elevator traffic pattern, a set of key floors can be determined. Then, for each future passenger flow scenario, the request intensity, waiting risk, and elevator resource occupancy impact corresponding to each key floor are obtained. The request intensity, waiting risk, and elevator resource occupancy impact corresponding to each future passenger flow scenario for each key floor are weighted and summed according to the probability of the passenger flow scenario to obtain the comprehensive reserved demand intensity. Based on this, the current number of reserved elevators, the target floors for reservation or return, the reservation time window, and the penalty value R for insufficient elevator resource reservation are determined. _loss And its weights, gamma, and other parameters.

[0049] S35. Based on the elevator traffic mode parameter weights, the future passenger flow scenario set, and the elevator resource reservation strategy, obtain the dispatch generation value corresponding to each candidate dispatch scheme, and select the candidate dispatch scheme corresponding to the minimum dispatch generation value as the final dispatch scheme.

[0050] In this invention, preferably, the dispatching generation value corresponding to each candidate dispatching scheme is obtained based on the elevator traffic mode parameter weights, the future passenger flow scenario set, and the elevator resource reservation strategy, and the candidate dispatching scheme corresponding to the minimum dispatching generation value is selected as the final dispatching scheme. In this preferred scheme, the dispatching generation value includes the current dispatching generation value, the expected generation value of the future passenger flow scenario, and the penalty generation value for insufficient elevator resource reservation on key floors in the future passenger flow scenario, which can be calculated using the following formula: Z = Sum(C _r_k × x _r_k ) + beta × Sum(p_omega × Q(x, omega)) + gamma ×R _loss (x)(1) Among them, C _r_ k represents the current dispatch cost of assigning the destination floor call request r to elevator k; x _r_k This represents the binary elevator allocation decision variable, where x is the variable that determines when a call request r to the destination floor is allocated to elevator k. _r_k =1, otherwise x _r_k=0; beta represents the weight of the expected cost of future passenger flow scenarios; p_omega represents the probability of passenger flow scenario omega in the future passenger flow scenario; Q(x, omega) represents the remaining service cost generated by future passenger flow scenario omega under the current elevator dispatch decision x, i.e., candidate elevator dispatch scheme x, and is used to evaluate whether the elevator dispatch scheme will lead to situations such as insufficient available elevators on subsequent peak floors, long waiting times for elevator calls on subsequent destination floors, increased overall energy consumption, and unbalanced load; gamma represents the penalty weight for insufficient elevator resource reservation on key floors; R _loss (x) indicates that under the current elevator dispatch decision x, there is a penalty for insufficient elevator resource reservation in the future passenger flow scenario. That is, when the amount of reserved elevator resources required for key floors in the future passenger flow scenario is greater than the amount of elevator resources expected to be available to serve key floors in the future passenger flow scenario under the current elevator dispatch decision x, it is judged that there is insufficient resource reservation.

[0051] Formula (1) represents the formula for calculating the dispatch value corresponding to the current dispatch value, the expected value of future passenger flow scenarios, and the penalty value for insufficient elevator resource reservation on key floors in future passenger flow scenarios. Furthermore, from Formula (1), it can be seen that the first term Sum(C) in the formula... _r_k × x _r_k The first term, Sum(p_omega × Q(x, omega)), is used to obtain the current generation value of the passenger flow scenario. The second term, Sum(p_omega × Q(x, omega)), is used to obtain the expected generation value of the future passenger flow scenario. The third term, gamma × R, is used to obtain the expected generation value of the passenger flow scenario. _loss (x) is used to obtain the penalty value for insufficient elevator resource reservation. In other embodiments, those skilled in the art can obtain the dispatch value corresponding to each candidate dispatch scheme according to the elevator traffic mode parameter weight, or obtain the dispatch value corresponding to each candidate dispatch scheme according to the elevator traffic mode parameter weight and the future passenger flow scenario set. In this case, it is only necessary to remove the second, third or third term in formula (1) to obtain the corresponding dispatch value.

[0052] In this invention, a rapid forward evaluation is performed on each future passenger flow scenario in the set of future passenger flow scenarios, thereby avoiding local optima through the expected cost of the future scenario. Specifically, in a future passenger flow scenario, after determining the assigned elevators corresponding to all pending destination floor call requests, the available elevator locations, elevator capacity, and existing service queues are estimated when subsequent destination floor call requests arrive, and the remaining service cost Q(x, omega) of the future passenger flow scenario is calculated.

[0053] Regarding the current ladder cost C _r_k It includes the costs of waiting time, riding time, additional stops, number of floors traversed when empty, car capacity congestion, and special task penalty / reward costs, as shown below: C _r_k = A1× W _r_k + A2× J _r_k + A3× N _r_k + A4× D _r_k + A5× H _r_k + A6× P _r_k Among them, W _r_k Indicates the estimated waiting time, J _r_k Indicates the estimated elevator travel time, N _r_k Indicates the number of new stops, D _r_k Indicates the number of floors to bypass when unloaded, H _r_k P indicates a penalty for overcrowding in the elevator car. _r_k The special task penalty and reward items are represented by A1 to A6, which respectively represent the weights of the aforementioned elevator traffic mode parameters: estimated waiting time weight, estimated riding time weight, estimated number of new stops weight, number of floors to be detoured when empty weight, capacity congestion penalty weight, and special task penalty and reward weight.

[0054] Based on the current elevator dispatch cost C _r_k As can be seen from the calculation method, this invention incorporates multiple factors such as waiting time and number of start-stop cycles into the calculation of the dispatch cost Z, and can further consider future passenger flow scenarios and elevator resource reservation strategies. In the off-peak energy-saving mode, it can first obtain candidate schemes that meet the longest waiting time threshold, then calculate the total cost Z among the candidate schemes, including the current dispatch cost, the expected cost of future scenarios, and the penalty for insufficient resource reservation, and select the scheme with the smallest Z. This enables the elevator to accept waiting or resource reservation control, and move or stay at the base station, the waiting floor in the zone, or the predicted peak start floor, etc., to reduce the waiting time and empty return trip of future short-term requests, so as to balance energy saving and subsequent response capabilities. In addition, in determining the dispatch scheme from the candidate dispatch schemes, constraints can be added, such as the minimum number of elevators and the maximum number of elevators. Only the scheme with the required number of elevators is used as the candidate dispatch scheme, and the candidate dispatch scheme with the smallest dispatch cost is selected as the final dispatch scheme.

[0055] S4. According to the elevator dispatching scheme, send an elevator dispatching command to the corresponding elevator controller, and send elevator dispatching information back to the elevator call request input module of the destination floor.

[0056] After obtaining the elevator dispatch plan, for each destination floor call request to be assigned, the corresponding elevator number, direction or location guidance is returned to the destination floor call request input module, such as the destination floor selector. Passengers can then go to the designated elevator to wait according to the displayed results. At the same time, an elevator dispatch command is sent to the corresponding elevator controller, so that the elevator responds to the corresponding destination floor call request according to the elevator dispatch plan.

[0057] S5. Obtain the abnormal situation that occurs during the execution of the elevator dispatching scheme, and obtain the abnormal situation response elevator dispatching scheme based on the abnormal situation.

[0058] During the execution of the elevator dispatching plan, if an assigned elevator experiences an anomaly such as malfunction, overload, timeout failure, door malfunction, communication anomaly, entry into VIP exclusive service, entry into fire service, or backup power limitation, the elevator is removed from the set of available elevators, and the set of incomplete destination floor call requests in the dispatching plan is identified. Then, for the set of incomplete destination floor call requests, for those incomplete destination floor call requests where passengers have not yet boarded the elevator, steps S1 to S4 are re-executed to obtain a new dispatching plan; for those incomplete destination floor call requests where passengers have already boarded the elevator, if the elevator is currently in an operational state, service continues according to the predetermined destination floor in the car; if the elevator is currently in an inoperable state, for all incomplete destination floor call requests of that elevator, steps S1 to S4 are re-executed based on all incomplete destination floor call requests and the elevator's current floor to obtain a new dispatching plan.

[0059] Compared to existing technologies, the present invention identifies elevator traffic patterns based on destination floor call request data within the current time window and corresponding historical operational data. It then adjusts the corresponding elevator traffic pattern weights based on these patterns, rather than using fixed weight parameters. This improves the rationality of parameter settings, enabling the system to continuously adapt to various call scenarios. Furthermore, the present invention models the elevator allocation for current destination floor call requests in a unified manner with the expected costs of future short-term passenger flow scenarios. This ensures that current dispatch decisions not only pursue the shortest possible wait time but also consider the elevator group service capacity over a future period, such as tens of seconds to several minutes. This significantly reduces the risk of long waiting times and minimizes ineffective return trips and start-stop energy consumption. By acquiring elevator anomalies during the dispatch process and implementing an anomaly identification and corresponding destination floor call request reassignment mechanisms, and by setting the allocation results to be changed only in cases of full load, malfunction, timeout, or special tasks where service is unavailable, the invention achieves a balance between passenger guidance stability and system robustness.

[0060] Furthermore, the present invention also provides an elevator destination floor group control and scheduling device, which is used to execute the above-described elevator destination floor group control and scheduling method. The device includes a destination floor call request acquisition module, an elevator current status acquisition module, an elevator dispatch plan acquisition module, an elevator dispatch plan feedback module, an elevator abnormality acquisition and reallocation module, and a storage module.

[0061] The destination floor elevator call request acquisition module is used to acquire the destination floor elevator call request set.

[0062] The elevator current status acquisition module is used to collect the current status of each elevator in the elevator group, and obtain the set of available elevators based on the current status of the elevator and the set of elevator call requests for the destination floor.

[0063] The elevator dispatching scheme acquisition module includes a candidate elevator dispatching scheme acquisition submodule, an elevator traffic pattern recognition and parameter weight determination submodule, a future passenger flow scenario acquisition submodule, an elevator resource reservation strategy acquisition submodule, and an elevator dispatching scheme determination submodule. Among them: The candidate elevator dispatch scheme acquisition submodule is used to acquire candidate elevator dispatch schemes based on the destination floor call request set and the current elevator status set; the elevator traffic pattern recognition and parameter weight determination submodule is used to determine the current elevator traffic pattern based on the destination floor call request set and corresponding historical operation data within the current time window, and acquire elevator traffic pattern parameter weights based on the elevator traffic pattern; the future passenger flow scenario acquisition submodule is used to acquire the future passenger flow scenario set within a preset time period based on historical operation data; the elevator resource reservation strategy acquisition submodule is used to determine the elevator resource reservation strategy based on the future passenger flow scenario set and the elevator traffic pattern; the elevator dispatch scheme determination submodule is used to acquire the dispatch generation value corresponding to each candidate elevator dispatch scheme based on the elevator traffic pattern parameter weights, and select the candidate elevator dispatch scheme with the minimum dispatch generation value as the final elevator dispatch scheme, or, based on the elevator traffic pattern parameter weights and the elevator resource reservation strategy, acquire the dispatch generation value corresponding to each candidate elevator dispatch scheme, and select the candidate elevator dispatch scheme with the minimum dispatch generation value as the final elevator dispatch scheme.

[0064] The elevator dispatching scheme feedback module: based on the dispatching scheme, sends an elevator dispatching command to the corresponding elevator controller and sends dispatching information back to the destination floor call request input module. The elevator controller controls the elevator operation according to the dispatching command.

[0065] The elevator anomaly acquisition and redistribution module is used to acquire anomalies that occur during the execution of the elevator dispatching scheme, and to acquire an anomaly response scheme based on the anomalies.

[0066] The storage module is used to store various types of data, such as destination floor call requests, elevator dispatch schemes, actual arrival time at the starting floor, actual arrival time at the destination floor, actual elevator ride time, current elevator status, whether the elevator has been abnormally disconnected, whether the elevator dispatch scheme has been reassigned, and energy consumption estimates, thereby achieving data closure. This data is used for continuous updates of related data such as future elevator traffic pattern recognition, future passenger flow scenario sets, and passenger flow scenario probabilities, supporting embedded algorithm verification and continuous optimization.

[0067] In addition, during the embedded testing phase, each input and output data can be recorded as a scheduling case and stored in the storage module. This input and output data includes the algorithm name, configuration parameters, passenger requests, elevator status, and dispatch results, so that developers can export the scheduling cases as text files for algorithm playback, sample elevator verification, and performance report analysis.

[0068] Furthermore, this invention also provides an elevator destination floor group control and scheduling system, which includes multiple elevator units and the elevator destination floor group control and scheduling device. Each elevator unit includes a car, a car drive unit, an elevator controller, and a destination floor call request input module; the car is connected to the car drive unit, and the car drive unit is connected to the elevator controller; the destination floor call request input module and the elevator controller are respectively connected to the elevator destination floor group control and scheduling device. In a specific embodiment, the elevator destination floor group control and scheduling device and the elevator controller can be connected via CAN, 485, RS422, or other industrial communication interfaces.

[0069] To facilitate understanding of the present invention, the following embodiments are provided to illustrate the present invention.

[0070] Example 1 Assume an office building has 16 floors and 6 elevators from A to F, each with a capacity of 10 people. The inter-floor travel time is 1.5 seconds, and a single stop, including deceleration, door opening, closing, and acceleration losses, takes a total of 8 seconds. What is the maximum waiting time threshold T? _max =45 seconds, current time is 08:30 on a weekday.

[0071] Currently, a total of 5 destination floor call requests have been received, specifically: R1 is 4 people from 1F to 15F, R2 is 3 people from 1F to 16F, R3 is 2 people from 8F to 1F, R4 is 1 person from 6F to 12F, and R5 is 3 people from 14F to 1F.

[0072] Table 1 Current destination floor call requests

[0073] Based on the elevator call request from the destination floor, it is determined to be an upward peak mode.

[0074] Subsequently, based on the upward peak mode, the weights of the elevator traffic modes were determined as follows: waiting time weight A1=1.0, riding time weight A2=0.35, additional stop count weight A3=4.0, empty detour floor count weight A4=0.6, and capacity congestion penalty weight A5=10.0. Since the elevator call requests in this embodiment are all ordinary and there are no long waiting elevator call requests, the special task penalty reward weight A6=0. Meanwhile, based on historical operational data, a strong upward passenger flow is predicted for the first floor within the next 120 seconds. Specifically, a total of 22 people are expected to arrive from the 1F to 12F, 13F, 15F, and 16F directions, including 8 from 1F to 15F, 6 from 1F to 16F, 5 from 1F to 12F, and 3 from 1F to 13F. Therefore, the expected cost weight beta for this future passenger flow scenario is 0.8. If, after dispatching an elevator for the current call request, there are no available elevators in the lobby or no elevators that can return to the lobby within 20 seconds, then R... _loss (x) is set to 34.0; if at least one lobby elevator is reserved, the penalty term for insufficient resource reservation on key floors is R_loss(x) = 6.0; the penalty weight for resource reservation on key floors is gamma = 1.0.

[0075] The current elevator status of each elevator obtained in this embodiment is shown in the following table: Table 2 Current Status of Elevator

[0076] In this embodiment, based on the destination floor elevator call request, the set of available elevators, the elevator traffic mode weight set, and the resource reservation strategy, feasible combinations of elevator dispatching are enumerated and the total target value Z is calculated. The allocation scheme corresponding to the minimum total target value Z is selected as the final elevator dispatching scheme. R1 and R2 are both on the same floor going up and their destination floors are adjacent, which meets the condition of merging elevators. After merging, the number of people is 7, which does not exceed the single elevator capacity of 10 people. Elevator dispatching is carried out after merging.

[0077] Table 3. Ladder output scheme of the method of the present invention.

[0078] It should be noted that the estimated arrival time at the starting floor in Table 3 includes the estimated response / waiting time from when a passenger registers their request to board the elevator at the destination floor until the corresponding elevator completes its dispatch response and reaches a state where passengers can board. This includes the time for dispatch calculation, instruction issuance, controller response, door operator action, safety confirmation, or standby wake-up. The estimated arrival time at the destination floor is based on the estimated elevator travel time from when the passenger boards the elevator to when they arrive at the destination floor, rather than a rough estimate that includes all losses from a single stop.

[0079] According to Table 3, the method of the present invention combines R1 and R2 and assigns them to elevator A, while keeping elevator D on standby on the 1st floor to respond to future predicted passenger flow. Although R2's stop on the 15th floor results in an 8-second increase in the elevator travel time compared to directly reaching its destination floor, the above-mentioned elevator dispatching scheme of the present invention can significantly reduce the average waiting time of lobby passengers within the next 120 seconds.

[0080] To further demonstrate the advantages of the proposed solution over the existing dispatching strategies, the following comparison is made between the proposed solution and the current nearest-distance dispatching strategy, as detailed in Tables 4 and 5.

[0081] Table 4 shows the data of some candidate combinations and the current nearest-distance elevator dispatch strategy in the method of this invention.

[0082] Table 5 Comparison of simulation metrics between the present invention's scheme G1 and the current closest distance strategy G2

[0083] As shown in Tables 3 to 5, the present invention does not simply select the elevator that is currently closest, but rather weighs the current dispatch cost against the future peak passenger flow reservation. In scheme G1, although its current dispatch cost is 6.4 higher than that of scheme G2, its future / reservation cost is reduced by 28.0. Therefore, the dispatch cost Z is reduced from 93.4 to 71.8. This result shows that in the continuous upward peak scenario in this embodiment, retaining one available elevator in the lobby can significantly reduce the risk of subsequent long waiting times and reduce ineffective return trips and start-stop energy consumption.

[0084] Example 2 When a meeting floor receives more requests for its destination floor than the threshold calculated from historical operational data within 30 seconds, it is identified as a peak-hour mode. In this mode, a higher weight is assigned to waiting time (A1), capacity congestion penalty (A5), and special task penalty reward (A6) compared to normal traffic mode. A special task penalty reward item (P_r_k) is used to increase the penalty for long waiting times, and a penalty item (R_loss(x)) with its weight (gamma) for insufficient resource reservation on critical floors is used to increase the cost of elevator shortages on that floor. A preset number of elevators serving that floor can be set as a constraint, ensuring that elevators with the lowest dispatch cost and a quantity meeting the preset limit are dispatched to that floor. In this scheme, the optimal dispatching scheme is ultimately selected from the candidate dispatching schemes through Z-score calculation, using elevator traffic mode parameter weights, penalty reward items, and further constraints.

[0085] Example 3 When the elevator's operating mode is switched to standby power mode, the system determines the available elevator set based on the limit on the number of elevators allowed to operate simultaneously. Within this limit, it then obtains candidate dispatch schemes that meet the elevator quantity constraints and further re-obtains dispatch schemes for the destination floor requests. If already assigned requests are affected by the elevator operating mode adjustment, the system will re-obtain the corresponding dispatch schemes for those affected destination floor requests as needed and send the relevant information to the corresponding floor selector.

[0086] Compared to existing technologies, the solution of this invention identifies elevator traffic patterns based on destination floor call request data within the current time window and corresponding historical operation data. It then adjusts the corresponding elevator traffic pattern weights based on these patterns, rather than using fixed weight parameters. This improves the rationality of parameter settings, enabling the system to continuously adapt to various call scenarios. Furthermore, this invention models the elevator allocation for current destination floor call requests in a unified manner with the expected costs of future short-term passenger flow scenarios. This ensures that current dispatch decisions do not only pursue the closest or shortest waiting time but also consider the elevator group service capacity over a future period, such as tens of seconds to several minutes, based on future short-term passenger flow predictions. This significantly reduces the risk of long waiting times and minimizes ineffective return trips and start-stop energy consumption. By acquiring elevator anomalies during the dispatch process and implementing an anomaly identification and corresponding destination floor call request reassignment mechanism, and by setting the allocation results to be changed only in cases of full load, malfunction, timeout, special tasks, or other unserviceable anomalies, this invention achieves a balance between passenger guidance stability and system robustness.

[0087] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and the present invention also intends to include these modifications and variations.

Claims

1. A method for group control and scheduling of elevators at a destination floor, characterized in that, Includes the following steps: Obtain the set of call requests for the destination floor; Collect the current status of each elevator in the elevator group, and obtain the set of available elevators based on the current status of the elevator and the set of elevator call requests for the destination floor; Based on the set of elevator call requests for the destination floor, the set of available elevators, and historical operation data, an elevator dispatch plan is obtained; According to the elevator dispatching scheme, an elevator dispatching command is sent to the corresponding elevator controller, and elevator dispatching information is fed back to the elevator call request input module of the destination floor; The step of obtaining an elevator dispatch plan based on the destination floor call request set, the available elevator set, and historical operation data includes: Based on the set of elevator call requests for the destination floor and the set of available elevators, candidate elevator dispatch schemes are obtained; Based on the set of elevator call requests for the destination floor within the current time window and the corresponding historical operation data, the current elevator traffic mode is determined, and the elevator traffic mode parameter weights are obtained according to the elevator traffic mode. Among them, different elevator traffic modes have different elevator traffic mode parameter weights. Based at least on the weights of the elevator traffic mode parameters, obtain the dispatch generation value corresponding to each candidate dispatch scheme, and select the candidate dispatch scheme corresponding to the minimum dispatch generation value as the final dispatch scheme.

2. The scheduling method according to claim 1, characterized in that, The elevator traffic mode parameter weights include the weights for estimated waiting time, estimated riding time, estimated number of new stops, estimated number of floors to be bypassed by empty passengers, estimated capacity congestion penalty, and estimated special task penalty reward.

3. The scheduling method according to claim 2, characterized in that, The step of obtaining the elevator dispatch plan based on the destination floor elevator call request set, the available elevator set, and historical operation data also includes: Based on historical operational data, a set of future passenger flow scenarios within a preset time period is obtained; each future passenger flow scenario in the set of future passenger flow scenarios includes a set of future destination floor elevator call requests and a passenger flow scenario probability. Based on the elevator traffic mode parameter weights and the set of future passenger flow scenarios, the dispatch value corresponding to each candidate dispatch scheme is obtained.

4. The scheduling method according to claim 3, characterized in that, The elevator dispatch value includes the current elevator dispatch value and the expected value of future passenger flow scenarios. The elevator dispatch value Z is obtained through the following formula: Z = Sum(C _r_k × x _r_k ) + beta × Sum(p_omega × Q(x, omega)) Among them, C _r_ k represents the current dispatch cost of assigning the destination floor call request r to elevator k; x _r_k This represents the binary elevator allocation decision variable, where x is the variable that determines when a call request r to the destination floor is allocated to elevator k. _r_k =1, otherwise x _r_k =0; beta represents the weight of the expected cost of the future passenger flow scenario; p_omega represents the probability of the future passenger flow scenario omega; Q(x, omega) represents the remaining service cost generated by the future passenger flow scenario omega under the current elevator dispatch decision x.

5. The scheduling method according to claim 2, characterized in that, The step of obtaining the elevator dispatch plan based on the destination floor elevator call request set, the available elevator set, and historical operation data also includes: Based on historical operational data, a set of future passenger flow scenarios within a preset time period is obtained; each future passenger flow scenario in the set of future passenger flow scenarios includes a set of future destination floor elevator call requests and a passenger flow scenario probability. Based on the set of future passenger flow scenarios and their respective elevator traffic modes, determine the elevator resource reservation strategy; Based on the elevator traffic mode parameter weights, the future passenger flow scenario set, and the elevator resource reservation strategy, the dispatch value corresponding to each candidate dispatch scheme is obtained.

6. The scheduling method according to claim 5, characterized in that, The elevator dispatch value includes the current elevator dispatch value, the expected value of future passenger flow scenarios, and the penalty value for insufficient elevator resource reservation on key floors in future passenger flow scenarios, as shown in the following formula: Z = Sum(C _r_k × x _r_k ) + beta × Sum(p_omega × Q(x, omega)) + gamma ×R _loss (x) Among them, C _r_ k represents the current dispatch cost incurred when assigning the destination floor call request r to elevator k; x _r_k This represents the binary elevator allocation decision variable, where x is the variable that determines when a call request r to the destination floor is allocated to elevator k. _r_k =1, otherwise x _r_k =0; beta represents the weight of the expected cost of future passenger flow scenarios; p_omega represents the probability of passenger flow scenario omega in the future passenger flow scenario; Q(x, omega) represents the remaining service cost generated by future passenger flow scenario omega under the current elevator dispatch decision x; gamma represents the penalty weight for insufficient elevator resource reservation on key floors; R _loss (x) represents the penalty for insufficient elevator resource reservation on key floors in future passenger flow scenarios under the current elevator dispatch decision x.

7. The scheduling method according to claim 4 or 6, characterized in that, The current elevator cost C _r_ k is obtained using the following formula: C _r_k = A1 × W _r_k + A2 × J _r_k + A3 × N _r_k + A4 × D _r_k + A5 × H _r_k + A6 ×P _r_k Among them, W _r_k Indicates the estimated arrival time at the starting floor, J _r_k Indicates the estimated elevator travel time, N _r_k Indicates the number of new stops, D _r_k Indicates the number of floors to bypass when unloaded, H _r_k P indicates a penalty for overcrowding in the elevator car. _r_k The numbers represent special task penalty and reward items; A1 to A6 represent the weights of estimated waiting time, estimated riding time, estimated number of new stops, estimated number of floors to be detoured when empty, estimated capacity congestion penalty, and estimated special task penalty and reward, respectively.

8. The scheduling method according to any one of claims 1-7, characterized in that, The scheduling method further includes: obtaining abnormal situations that occur during the execution of the elevator dispatching scheme, and obtaining an abnormal situation response elevator dispatching scheme based on the abnormal situation.

9. A destination floor elevator group control and scheduling device, characterized in that, The device is used to execute the destination floor elevator group control and scheduling method as described in any one of claims 1-8.

10. A destination floor elevator group control and dispatching system, the system comprising multiple elevator units and an elevator destination floor group control and dispatching device; wherein, Each elevator unit includes a car, a car drive unit, an elevator controller, and a destination floor call request input module; the car is connected to the car drive unit, and the car drive unit is connected to the elevator controller; the destination floor call request input module and the elevator controller are respectively connected to the elevator destination floor group control and scheduling device, characterized in that the elevator destination floor group control and scheduling device is the elevator destination floor group control and scheduling device as described in claim 9.

Images