Multi-dimensional passenger flow intelligent regulation and control distribution method and system for scenic region operation, medium and processor

Abstract

The invention discloses a multi-dimensional passenger flow intelligent regulation and control distribution method and system for scenic area operation, a medium and a processor, and relates to the technical field of scenic area datamation management. The method comprises the steps of obtaining passenger flow data collected by multi-source sensing equipment in real time; according to the real-time updated facility state of the passenger flow data and the calculated path passing weight, obtaining a real-time road network state; the method comprises the following steps: training a personalized route recommendation model based on tourist historical behavior data to obtain tourist behavior characteristics; according to the real-time road network state and the tourist behavior characteristics, an LSTM-Attention hybrid model is adopted to predict the passenger flow distribution of each region in the future time, and a prediction result of a potential congestion point is identified; and according to the prediction result, generating a tourist regulation and control distribution strategy. According to the method, the passenger flow distribution of each region in the future is predicted, and the scenic spot passenger flow is regulated and controlled through a reasonable regulation and control strategy, so that the digital intelligent operation of the scenic spot is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of digital management of scenic spots, and in particular to a method, system, medium and processor for intelligently controlling and allocating multi-dimensional passenger flows in scenic spot operations. Background Art

[0002] With rising living standards and shifting tourism concepts, the tourism industry is booming, with the number of visitors to scenic spots steadily increasing. However, this has also, to a certain extent, increased pressure on scenic areas. Tourists are overly concentrated in popular scenic spots or certain areas within scenic areas, leading to frequent traffic congestion, long queues, and a diminished tourist experience. Against this backdrop, how to rationally regulate and allocate tourist flow within scenic areas has become a key issue in scenic area management.

[0003] Traditional scenic spot visitor flow management methods rely primarily on manual statistics and empirical judgment, resulting in inaccurate data, poor real-time performance, and insufficient forecasting capabilities. These methods struggle to meet the complex and ever-changing visitor flow management needs of modern scenic spots. While some existing scenic spot visitor flow monitoring systems can collect basic data and perform simple statistics, they estimate visitor numbers based solely on ticket sales data or entrance gate data. These systems are unable to accurately grasp the distribution of visitor flow within each area of ​​the scenic spot in real time, nor can they predict congestion in advance and implement effective preventive measures. Existing visitor monitoring methods lack the ability to dynamically predict data and implement intelligent control. Summary of the Invention

[0004] To address the problem that existing tourist monitoring methods lack dynamic data prediction and intelligent regulation, the present invention provides a multi-dimensional passenger flow intelligent regulation and distribution method, system, medium, and processor for scenic area operations. These methods can predict the future passenger flow distribution in each area and regulate the passenger flow of the scenic area through reasonable regulation strategies, thus realizing digital intelligent operation of the scenic area. The specific technical solutions are as follows:

[0005] A multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations, comprising:

[0006] Obtain passenger flow data collected in real time by multi-source sensing devices;

[0007] The real-time road network status is obtained based on the real-time updated facility status and calculated path traffic weights based on passenger flow data;

[0008] By training a personalized route recommendation model based on tourists’ historical behavior data, we can obtain tourist behavior characteristics;

[0009] Based on the real-time road network status and tourist behavior characteristics, the LSTM-Attention hybrid model is used to predict the passenger flow distribution in each area in the future and identify potential congestion points.

[0010] Generate tourist control and allocation strategies based on the prediction results.

[0011] Preferably, a multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation also includes:

[0012] Multi-agent simulation of the control allocation strategy is performed in a digital twin environment, and the control effectiveness is evaluated by comparing the KL divergence of the actual passenger flow distribution with the predicted results.

[0013] Preferably, the acquiring of passenger flow data collected in real time by multi-source sensing devices includes:

[0014] The passenger flow data collected by gate identification equipment, video surveillance equipment, positioning equipment and mobile terminals is obtained, de-identified by edge computing nodes, and uploaded to the cloud big data platform.

[0015] Preferably, the real-time road network status obtained by updating the facility status and path traffic weight in real time based on the passenger flow data includes:

[0016] According to the types, uses and correlations of various facilities in the scenic area with passenger flow data, a real-time updated facility operation status matrix is ​​obtained;

[0017] Calculate the path traffic weight assigned to each path in the scenic area based on the spatial characteristics, key locations and passenger flow data of the path;

[0018] According to the updated facility operation status matrix and the calculated path traffic weight data, combined with the geographic spatial information of the scenic area, the road network topology structure is dynamically generated to obtain the real-time road network status.

[0019] Preferably, the tourist behavior characteristics obtained by training the personalized route recommendation model based on the tourist historical behavior data include:

[0020] Obtain historical tourist behavior data at scenic spots, including historical tour routes, length of stay at scenic spots, consumption preferences, and number of travelers;

[0021] The model uses tourists' historical tour routes, duration of stay at scenic spots, consumption preferences, and number of companions as input features, and tourists' satisfaction with the routes as output labels. A deep learning-based neural network model is trained to perform personalized route recommendation, resulting in a trained personalized route recommendation model.

[0022] Based on the trained personalized route recommendation model, the real-time behavior data of tourists in the scenic area is analyzed in real time, and the tourist behavior characteristics are dynamically generated.

[0023] Preferably, the prediction results of using the LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future time based on the real-time road network status and tourist behavior characteristics and identifying potential congestion points include:

[0024] The traffic weights of each path in the real-time road network data, the matrix data of the facility operation status, and the characteristics of tourist behavior are input into the trained LSTM-Attention hybrid model. The LSTM layer processes the time series information, and the Attention mechanism focuses on key historical information. The model outputs the predicted passenger flow distribution values ​​for each area in each time period in the future, including the number of tourists, population density, and passenger flow direction.

[0025] The predicted passenger flow distribution values ​​of each area are compared with the preset congestion threshold. When the predicted passenger flow index of a certain area exceeds the threshold, it is identified as a potential congestion point.

[0026] Preferably, generating a tourist control and allocation strategy based on the prediction results includes:

[0027] Based on the prediction results, a multi-objective optimization model combining minimum congestion cost and best tourist experience is constructed;

[0028] The multi-objective optimization model is solved based on the path induction strategy to obtain the tourist regulation and allocation results.

[0029] A multi-dimensional passenger flow intelligent control and allocation system for scenic area operations, applying the aforementioned multi-dimensional passenger flow intelligent control and allocation method for scenic area operations, comprises:

[0030] A data acquisition unit, used to acquire passenger flow data collected in real time by multi-source sensing devices;

[0031] The road network construction unit is used to obtain the real-time road network status by updating the facility status in real time based on passenger flow data and calculating the path traffic weight;

[0032] A behavior feature generation unit, configured to generate tourist behavior features by training a personalized route recommendation model based on tourist historical behavior data;

[0033] The congestion prediction unit uses the LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future based on the real-time road network status and tourist behavior characteristics, and identifies the prediction results of potential congestion points;

[0034] The control and allocation unit is used to generate a control and allocation strategy for tourists based on the prediction results.

[0035] A computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the aforementioned multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations.

[0036] A processor is used to run a program, wherein when the program is running, the multi-dimensional passenger flow intelligent control and allocation method for scenic area operations is executed.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention provides a multi-dimensional intelligent passenger flow control and allocation method for scenic area operations. The method obtains passenger flow data collected in real time by multi-source sensing devices; obtains the real-time road network status based on the real-time updated facility status and calculated path traffic weights of the passenger flow data; obtains tourist behavior characteristics by training a personalized route recommendation model based on historical tourist behavior data; uses an LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future based on the real-time road network status and tourist behavior characteristics, and identifies the predicted results of potential congestion points; and generates a tourist control and allocation strategy based on the predicted results. The present invention ensures the safety of tourists by predicting the passenger flow distribution in each area in the future. It also achieves precise control of the passenger flow in the scenic area through a reasonable passenger flow control strategy, thereby improving the efficiency of the digital intelligent operation of the scenic area. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0040] Figure 1 This is a flow chart of a multi-dimensional passenger flow intelligent control and allocation method for scenic area operations according to the present invention.

[0041] Figure 2 This is a schematic diagram of a multi-dimensional passenger flow intelligent control and distribution system for scenic area operations according to the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0043] It should be understood that when used in this specification, the terms "include" and "comprising" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0044] It should also be understood that the terms used in the present specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0045] It should be further understood that the term “and / or” used in the description of the present invention refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0046] Please refer to the following examples Figure 1 and Figure 2 .

[0047] The present application provides a multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations, including:

[0048] Step S1: Obtain passenger flow data collected in real time by multi-source sensing devices;

[0049] Passenger flow data collected by gate identification equipment, video surveillance equipment, positioning equipment and mobile terminals is obtained, de-identified by edge computing nodes, and uploaded to the cloud big data platform.

[0050] By deploying cameras at key points within the scenic area, such as entrances, popular attractions, and rest areas, these cameras can be used to identify visitors and facial recognition cameras can be used to roughly calculate real-time visitor traffic at each location. Infrared bidirectional counters are also installed at ticket gates and entrances to various venues. As visitors pass through the gates, the counters accurately record the number of people entering and exiting, enabling precise statistics on the number of visitors entering the scenic area and each venue. Positioning devices can utilize WiFi probes, which track the MAC addresses of visitors' phones. By strategically placing WiFi probes throughout the scenic area, when a visitor's phone's WiFi function is enabled, the probes can capture the MAC addresses of their phones, allowing analysis of their movements within the area and understanding their preferred routes. Shared vehicles, such as shuttle buses and shared bicycles, are equipped with GPS / Beidou positioning modules to provide real-time location information. Mobile point-of-sale (POS) terminals are installed at commercial outlets within the scenic area to collect real-time consumption data, understand visitor spending patterns, and identify consumer trends in different areas.

[0051] Step S2: obtaining the real-time road network status based on the facility status updated in real time by the passenger flow data and the calculated path traffic weights;

[0052] Specifically, the real-time road network status obtained by updating the facility status and path traffic weights in real time based on passenger flow data includes:

[0053] According to the types, uses and correlations of various facilities in the scenic area with passenger flow data, a real-time updated facility operation status matrix is ​​obtained; the facility operation status matrix is ​​expressed as

[0054] S=[s1,s2,...,s n ],s i ∈{00,10,11}

[0055] Among them, "00" means the device is closed; "10" means the device is open and in idle state; "11" means the device is open and in busy state.

[0056] Based on the spatial characteristics, key locations and passenger flow data of the path, the path traffic weight assigned to each path in the scenic area is calculated; the path traffic weight assigned to each path is calculated as follows:

[0057] W ij (t) = T ij ×[1+β×(α i +α j ) / 2]+γ×D(t)

[0058]

[0059] Among them, W ij (t) represents the communication weight of the path from node i to node j at time t; β is the congestion impact factor (taken as 0.5 in this embodiment); γ is the weather correction coefficient, and D(t) is the real-time weather impact function; C i is the visitor capacity of node i; N i is the number of visitors at the current node.

[0060] Nodes are mainly set up at key locations in the scenic area, such as entrances, scenic spots, rest areas and toilets.

[0061] According to the updated facility operation status matrix and the calculated path traffic weight data, combined with the geographic spatial information of the scenic area, the road network topology structure is dynamically generated to obtain the real-time road network status.

[0062] Set the key locations of the scenic area as nodes, and define the node attribute set based on the passenger flow data:

[0063] V={v i |v i =(coordinate, node capacity Ci , the current number of nodes N i )}

[0064] Set the paths of the scenic area as edges, and define the edge attribute set based on the spatial characteristics of the paths of the scenic area:

[0065] E={e ij |e ij =(length L ij , benchmark travel time T ij , real-time path weight W ij )}

[0066] Based on the path network in the GIS map of the scenic area, the path network G=(V, E) in the GIS map is extracted.

[0067] Step S3: obtaining tourist behavior characteristics by training a personalized route recommendation model based on tourist historical behavior data; specifically, including:

[0068] Obtain historical tourist behavior data at scenic spots, including historical tour routes, length of stay at scenic spots, consumption preferences, and number of travelers;

[0069] The model uses tourists' historical tour routes, duration of stay at scenic spots, consumption preferences, and number of companions as input features, and tourists' satisfaction with the routes as output labels. A deep learning-based neural network model is trained to perform personalized route recommendation, resulting in a trained personalized route recommendation model.

[0070] Based on the trained personalized route recommendation model, the real-time behavior data of tourists in the scenic area is analyzed in real time, and the tourist behavior characteristics are dynamically generated.

[0071] Step S4: Based on the real-time road network status and tourist behavior characteristics, the LSTM-Attention hybrid model is used to predict the passenger flow distribution in each area in the future and identify the prediction results of potential congestion points. Specifically, the following are performed:

[0072] The traffic weights of each path in the real-time road network data, the matrix data of the facility operation status, and the characteristics of tourist behavior are input into the trained LSTM-Attention hybrid model. The LSTM layer processes the time series information, and the Attention mechanism focuses on key historical information. The model outputs the predicted passenger flow distribution values ​​for each area in each time period in the future, including the number of tourists, population density, and passenger flow direction.

[0073] The predicted passenger flow distribution values ​​of each area are compared with the preset congestion threshold. When the predicted passenger flow index of a certain area exceeds the threshold, it is identified as a potential congestion point.

[0074] Features such as each area's geographic location, facility status, and path weights are extracted from the road network topology and converted into numerical form suitable for model input. Furthermore, key features relevant to passenger flow distribution prediction are extracted based on tourist behavior feature vectors, such as the real-time density of tourists within the scenic area, the inflow and outflow rates of tourists in different areas, and the weights of tourists' preferences for various attractions. These extracted feature data are normalized, and the preprocessed input data is fed into a hybrid LSTM + Attention model. Based on the learned time series features and attention weights, the model predicts passenger flow for each area of ​​the scenic area within the next hour. Based on the predicted passenger flow data and combined with the spatial capacity and facility carrying capacity of each area, potential congestion points are identified.

[0075] Step S5: Generate a tourist control and allocation strategy based on the prediction results, specifically including:

[0076] Based on the prediction results, a multi-objective optimization model combining minimum congestion cost and best tourist experience is constructed;

[0077] (1) Minimizing the congestion cost is:

[0078]

[0079] There are V nodes and E paths in the scenic area. The congestion cost of node i is defined as A i =w i t i , where w i is the node weight, t i is the average waiting time of tourists at node i.

[0080] The congestion cost of path ij is B ij =ρ ij L ij , where ρ i is the path congestion; L ij is the path length.

[0081] Since tourists are mainly concentrated at the node locations, the path congestion ρ i A similar estimate can be made:

[0082] Path congestion ρ i = Total estimated number of people at each node along the route / tourist capacity at each node along the route

[0083] (2) Maximizing the visitor experience:

[0084]

[0085] Where S i For comfort, S i=1-ρ i ;E ij Indicates convenience, E ij =1 / L ij ; V K represents the satisfaction of node k; ω1, ω2 and ω3 are the corresponding weights respectively.

[0086] At the same time, corresponding constraints are set, including the diversion number limit constraint and the maximum detour distance limit constraint.

[0087] The diversion number limit constraint is expressed as:

[0088]

[0089] Among them, X ij Indicates the number of diverted people; represents the remaining capacity of path ij.

[0090] The maximum detour distance constraint is expressed as:

[0091] Alternative path distance - original path distance ≤ maximum detour distance

[0092] The multi-objective optimization model is solved based on the path induction strategy to obtain the tourist regulation and allocation results.

[0093] By pushing 1-3 optimal alternative routes to tourists. The alternative routes are determined based on factors such as the current tourist flow of the scenic spot, the congestion of the route, and the personalized needs of tourists. A score is calculated for each recommended route. The scoring formula is:

[0094] Score = 0.5 × time + 0.3 × crowding + 0.2 × attraction value

[0095] By giving a larger weight to time factors, this approach helps guide tourists quickly around congestion and save them time. Congestion reflects the degree of congestion along a route, minimizing the need to recommend relatively relaxed and comfortable routes. Scenic spot value considers the appeal of attractions along the route, ensuring that even detours provide a positive experience. Visitors can choose the route that best suits them based on route scores, achieving autonomous route optimization.

[0096] In step S3 of this embodiment, by constructing a historical behavior database, multi-dimensional tourist behavior data is collected, including user ID, tour route, hot spots (attractions or areas where tourists stay for a long time), consumption records (consumption details in the scenic area, such as dining, shopping, etc.) and device interaction logs (interaction status between tourists and smart devices in the scenic area, such as querying attraction information, purchasing tickets, etc.).

[0097] By analyzing the variance of tourists' movement speed and acceleration, we can understand their walking rhythm and speed changes. We can also calculate directional persistence to determine whether tourists' directional preferences are stable throughout their visit, such as whether they tend to visit multiple attractions in a certain direction or frequently change their directions. Based on data such as tourists' itineraries, stopover hotspots, and spending records, we construct an attraction type preference matrix. Each element represents a tourist's preference for a specific attraction type. This matrix is ​​calculated by analyzing metrics such as the length of time a tourist spends at each attraction type and the amount of money spent. This matrix identifies tourists' interests and enables personalized recommendations.

[0098] The personalized recommendation model is trained using deep reinforcement learning methods to construct a state space, including the current location of tourists and the environmental status of the scenic area (such as congestion conditions of surrounding attractions, weather conditions, etc.); the action space is defined as a set of recommended routes, that is, the model can select different routes from this set and recommend them to tourists.

[0099] In step S4 of this embodiment, a three-layer stacked LSTM (Long Short-Term Memory) network structure is constructed to process time series features. The number of cells in the first LSTM layer is 256, the second layer is 128, and the third layer is 64. The number of cells decreases as the number of layers increases. A dropout rate (0.2) is set after each LSTM layer to randomly discard the output of some neurons.

[0100] The spatial attention of the Attention mechanism is used to calculate the association weights between different areas within the scenic area and determine which areas have a strong correlation in passenger flow migration. For example, there may be a high degree of correlation between popular attractions and their surrounding dining areas and rest areas. Tourists may go to these areas to rest and consume after visiting the attractions. The temporal attention of the Attention mechanism is used to capture the cyclical patterns of passenger flow changes, including daily cycles (such as the differences in passenger flow peaks and troughs during the day and night) and hourly cycles (such as fluctuations in passenger flow at different time periods every hour). Considering that the passenger flow in scenic areas usually has obvious diurnal changes and differences in different activity periods within a day, the temporal attention mechanism can help the model better understand the cyclical characteristics of passenger flow.

[0101] The final prediction result is output in the form of a regional passenger flow density matrix, where each element represents the passenger flow density of each region at a future moment.

[0102] The present invention provides a multi-dimensional intelligent passenger flow control and allocation method for scenic area operations. The method obtains passenger flow data collected in real time by multi-source sensing devices; obtains the real-time road network status based on the real-time updated facility status and calculated path traffic weights of the passenger flow data; obtains tourist behavior characteristics by training a personalized route recommendation model based on historical tourist behavior data; uses an LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future based on the real-time road network status and tourist behavior characteristics, and identifies the predicted results of potential congestion points; and generates a tourist control and allocation strategy based on the predicted results. The present invention ensures the safety of tourists by predicting the passenger flow distribution in each area in the future. It also achieves precise control of the passenger flow in the scenic area through a reasonable passenger flow control strategy, thereby improving the efficiency of the digital intelligent operation of the scenic area.

[0103] Specifically, in a preferred embodiment of the present application, a multi-agent simulation of the control allocation strategy is performed in a digital twin environment, and the control effectiveness is evaluated by comparing the KL divergence of the actual passenger flow distribution with the predicted results.

[0104] In this embodiment, a digital twin model of a scenic spot is constructed to simulate visitor flow distribution and tourist behavior within the scenic spot. Multi-agent simulation is introduced within the digital twin environment. Agents can represent elements such as tourists and scenic spot facilities, each of which is assigned different behavioral rules and decision-making logic to better reflect the complex dynamics within the scenic spot. For example, the tourist agent can select the optimal tour route based on real-time path access weights, while the facility agent can adjust its accessibility based on tourist flow. In scenic spot passenger flow control, the effectiveness of the control strategy can be quantitatively evaluated by comparing the KL divergence between the actual passenger flow distribution and the predicted results. When the control strategy is effective, the difference between the predicted and actual passenger flow distribution decreases, that is, the KL divergence decreases. Conversely, a large KL divergence indicates that the control strategy needs further optimization. Using the KL divergence to quantitatively evaluate the effectiveness of the control strategy enables refined management and optimization of scenic spot passenger flow, enhances the tourist experience, and ensures the efficiency of scenic spot operations.

[0105] The present application also provides a multi-dimensional passenger flow intelligent control and allocation system for scenic spot operations, which applies the aforementioned multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations, including:

[0106] A data acquisition unit, used to acquire passenger flow data collected in real time by multi-source sensing devices;

[0107] The road network construction unit is used to obtain the real-time road network status by updating the facility status in real time based on passenger flow data and calculating the path traffic weight;

[0108] A behavior feature generation unit, configured to generate tourist behavior features by training a personalized route recommendation model based on tourist historical behavior data;

[0109] The congestion prediction unit uses the LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future based on the real-time road network status and tourist behavior characteristics, and identifies the prediction results of potential congestion points;

[0110] The control and allocation unit is used to generate a control and allocation strategy for tourists based on the prediction results.

[0111] The functional explanation of each unit in this embodiment is the same as that of a multi-dimensional passenger flow intelligent control and distribution method for scenic spot operations, and the technical effect is the same, so it will not be repeated here.

[0112] An embodiment of the present application also provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the aforementioned multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations.

[0113] The technical effect of this embodiment is the same as the technical effect of a multi-dimensional passenger flow intelligent control and distribution method for scenic spot operations in an embodiment, and will not be repeated here.

[0114] The present invention can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations, such as personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments that include any of the above.

[0115] An embodiment of the present application further provides a processor, which is used to run a program, wherein when the program is running, the aforementioned multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations is executed.

[0116] The technical effect of this embodiment is the same as the technical effect of the multi-dimensional passenger flow intelligent control and distribution method for scenic spot operation in embodiment 1, and will not be repeated here.

[0117] The processor in this embodiment may be a central processing unit (CPU), a controller, a microcontroller, or other data processing chips.

[0118] Those skilled in the art will appreciate that the units of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0119] In the embodiments provided by the present invention, it should be understood that the division of units is merely a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0120] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0121] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nly Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program code.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the specification of the present invention.

Claims

1. A multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation, characterized in that: include: Obtain passenger flow data collected in real time by multi-source sensing devices; Real-time road network status is obtained based on the real-time updated facility status and calculated path traffic weights based on passenger flow data; By training a personalized route recommendation model based on tourists’ historical behavior data, we can obtain tourist behavior characteristics; Based on the real-time road network status and tourist behavior characteristics, the LSTM-Attention hybrid model is used to predict the passenger flow distribution in each area in the future and identify potential congestion points. Generate tourist control and allocation strategies based on the prediction results.

2. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 1 is characterized in that: Also includes: Multi-agent simulation of the control allocation strategy is performed in a digital twin environment, and the control effectiveness is evaluated by comparing the KL divergence of the actual passenger flow distribution with the predicted results.

3. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 1 is characterized in that: The acquisition of passenger flow data collected in real time by multi-source sensing devices includes: The passenger flow data collected by gate identification equipment, video surveillance equipment, positioning equipment and mobile terminals is obtained, de-identified by edge computing nodes, and uploaded to the cloud big data platform.

4. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 3 is characterized in that: The real-time road network status obtained by the facility status and path traffic weight updated in real time based on passenger flow data includes: According to the types, uses and correlations of various facilities in the scenic area with passenger flow data, a real-time updated facility operation status matrix is ​​obtained; Calculate the path traffic weight assigned to each path in the scenic area based on the spatial characteristics, key locations and passenger flow data of the path; According to the updated facility operation status matrix and the calculated path traffic weight data, combined with the geographic spatial information of the scenic area, the road network topology structure is dynamically generated to obtain the real-time road network status.

5. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 3 is characterized in that: The personalized route recommendation model is trained based on the tourists’ historical behavior data to obtain the following tourist behavior characteristics: Obtain historical tourist behavior data at scenic spots, including historical tour routes, length of stay at scenic spots, consumption preferences, and number of travelers; The model uses tourists' historical tour routes, duration of stay at scenic spots, consumption preferences, and number of companions as input features, and tourists' satisfaction with the routes as output labels. A deep learning-based neural network model is trained to perform personalized route recommendation, resulting in a trained personalized route recommendation model. Based on the trained personalized route recommendation model, the real-time behavior data of tourists in the scenic area is analyzed in real time, and the tourist behavior characteristics are dynamically generated.

6. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 4 is characterized in that: Based on the real-time road network status and tourist behavior characteristics, the LSTM-Attention hybrid model is used to predict the passenger flow distribution in each area in the future. The prediction results of identifying potential congestion points include: The traffic weights of each path in the real-time road network data, the matrix data of the facility operation status, and the characteristics of tourist behavior are input into the trained LSTM-Attention hybrid model. The LSTM layer processes the time series information, and the Attention mechanism focuses on key historical information. The model outputs the predicted passenger flow distribution values ​​for each area in each time period in the future, including the number of tourists, population density, and passenger flow direction. The predicted passenger flow distribution values ​​of each area are compared with the preset congestion threshold. When the predicted passenger flow index of a certain area exceeds the threshold, it is identified as a potential congestion point.

7. The multi-dimensional passenger flow intelligent control and allocation method for scenic spot operation according to claim 1 is characterized in that: The control and allocation strategy for tourists generated according to the prediction results includes: Based on the prediction results, a multi-objective optimization model combining minimum congestion cost and best tourist experience is constructed; The multi-objective optimization model is solved based on the path induction strategy to obtain the tourist regulation and allocation results.

8. A multi-dimensional passenger flow intelligent control and distribution system for scenic area operations, characterized by: The multi-dimensional passenger flow intelligent control and distribution method for scenic spot operation according to any one of claims 1 to 7 is applied, comprising: A data acquisition unit, used to acquire passenger flow data collected in real time by multi-source sensing devices; The road network construction unit is used to obtain the real-time road network status by updating the facility status in real time based on passenger flow data and calculating the path traffic weight; A behavior feature generation unit, configured to generate tourist behavior features by training a personalized route recommendation model based on tourist historical behavior data; The congestion prediction unit uses the LSTM-Attention hybrid model to predict the passenger flow distribution in each area in the future based on the real-time road network status and tourist behavior characteristics, and identifies the prediction results of potential congestion points; The control and allocation unit is used to generate a control and allocation strategy for tourists based on the prediction results.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored program, wherein, when the program is running, the device where the computer-readable storage medium is located is controlled to execute the multi-dimensional passenger flow intelligent control and distribution method for scenic spot operations as described in any one of claims 1 to 7.

10. A processor, characterized in that: The processor is used to run a program, wherein when the program is running, the multi-dimensional passenger flow intelligent control and allocation method for scenic spot operations described in any one of claims 1 to 7 is executed.

Images