Sightseeing vehicle dispatching method and system based on multi-source data fusion

Abstract

The application discloses a sightseeing vehicle scheduling method and system based on multi-source data fusion, the method comprising the following steps: acquiring vehicle operation data, tourist distribution data, station card swiping data, reservation data and scenic road data, constructing a scenic service graph, and determining an associated scenic spot set according to the walking time from the scenic spot area to the station; combining the number of tourists, the departure rate, the time slice length, the shunting proportion, the number of reservations and the number of people who have been waiting to obtain an estimated value of the comprehensive demand of each station; correcting the remaining demand according to the actual number of passengers after the vehicle arrives at the station and picks up passengers, and converting the remaining demand into the number of service vehicles required; constructing a time expansion network on the basis of the scenic service graph, and solving the time expansion network in combination with the vehicle, the seat, the power, the road and the station capacity constraints to generate a scheduling instruction.

Description

Technical Field

[0001] This application relates to the field of vehicle scheduling technology, specifically to a method and system for scheduling sightseeing vehicles based on multi-source data fusion. Background Technology

[0002] Sightseeing vehicles are a crucial mode of transportation for tourists within scenic areas. They typically operate between fixed stops, connecting scenic areas, entrances / exits, parking areas, and other service facilities. Tourist flow within scenic areas is significantly influenced by factors such as tour routes, attraction attractiveness, dwell time, reservation arrangements, and weather, exhibiting marked temporal variability and spatial dispersion. Demand for vehicles varies considerably between stops at different times of day. Current sightseeing vehicle dispatching methods largely rely on current queue lengths at stops, human experience, or historical passenger flow statistics. This approach struggles to promptly reflect the potential demand generated by tourists moving from scenic areas to stops on foot, easily leading to problems such as delayed vehicle allocation and a mismatch between capacity and actual demand.

[0003] Meanwhile, scenic area roads are typically narrow, have limited capacity, and are prone to congestion in certain sections. Some bus stops are also limited by space constraints, allowing only a limited number of vehicles to stop at any one time. If constraints such as road capacity, stop capacity, remaining seats, and remaining battery power in electric vehicles are not fully considered during scheduling, multiple vehicles may converge on the same stop or section of road, causing congestion, increased empty runs, and longer waiting times for tourists. Furthermore, relying solely on static statistical parameters for demand forecasting lacks a mechanism to correct forecast deviations based on actual visitor numbers, making it easy for errors to accumulate over consecutive scheduling cycles. Summary of the Invention

[0004] This application provides a method and system for scheduling sightseeing vehicles based on multi-source data fusion, in order to at least solve some of the technical problems existing in the related technologies described above.

[0005] According to a first aspect of the embodiments of this application, a method for scheduling sightseeing vehicles based on multi-source data fusion is provided, including: Acquire vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data to construct a scenic area service map; establish walking accessibility relationships based on walking time from each scenic area to each station, and determine the set of associated scenic spots for each station; For each station, the demand contribution is calculated based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each related attraction. The demand contribution of all related attractions, the number of reserved passengers and the number of waiting passengers are summed to obtain the comprehensive demand estimate. Once the vehicle arrives at the station and completes passenger pick-up, the actual number of passengers is obtained. The remaining demand is then corrected by subtracting the actual number of passengers from the overall demand estimate. The remaining demand after correction at each station is converted into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Constraints such as vehicle flow conservation, seat capacity, power consumption, road segment traffic capacity, and station stop capacity are set. The goal is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and a scheduling instruction is generated.

[0006] As an optional solution, the scenic area service map includes scenic spot nodes, station nodes, parking area nodes, and charging point nodes; the attributes of the station nodes include geographical location and station parking capacity, and the attributes of the charging point nodes include charging power and the number of available charging positions; the edges of the scenic area service map represent road segments connecting each node, and each edge has road segment length, travel time, and road segment travel capacity.

[0007] As an optional approach, the calculation of demand contribution based on the number of tourists, tourist departure rate, time slot length, and diversion ratio of each associated attraction includes: for each associated attraction of the station, determining an effective departure time window based on the walking time from the attraction to the station and the start and end times of the target time slot; the start time of the effective departure time window is the start time of the target time slot minus the corresponding walking time, and the end time is the end time of the target time slot minus the corresponding walking time; obtaining the number of tourists at the start time of the effective departure time window; multiplying the number of tourists by the tourist departure rate, time slot length, and diversion ratio from the attraction to the station to obtain the demand contribution of the attraction to the station.

[0008] As an optional approach, the diversion ratio is determined based on the inverse relationship between the walking time from the associated attractions to each station, with stations having shorter walking times corresponding to higher diversion ratios; the initial value of the diversion ratio is determined through historical card-swipe data statistics.

[0009] As an optional approach, after obtaining the actual number of guests, the actual diversion ratio of each related attraction to the station is inferred from the actual number of guests in the most recent time slots. The actual diversion ratio is then weighted and summed with the diversion ratio before the update using an exponential smoothing method to obtain the updated diversion ratio, which is used to calculate the demand contribution in subsequent time slots. The smoothing coefficient of the exponential smoothing is an open interval between 0 and 1, and the smoothing coefficient is pre-calibrated according to the fluctuation characteristics of the scenic area's passenger flow.

[0010] As an optional solution, the nodes of the time-extended network are tuples composed of node numbers and time slice numbers in the scenic area service map; the arcs of the time-extended network include driving arcs, waiting arcs, and charging arcs; driving arcs connect time-extended nodes of different nodes in the corresponding time layer of road segments, waiting arcs connect time-extended nodes of the same node in adjacent time layers, and charging arcs connect time-extended nodes of the charging point node in the starting time layer and the corresponding time layer after charging is completed.

[0011] As an optional solution, the station stopping capacity constraint is: within the same time slot, the total number of vehicles stopping at the same station does not exceed the station's maximum stopping capacity; the road segment traffic capacity constraint is: within the same time slot, the total number of vehicles passing through the same road segment does not exceed the road segment's maximum traffic capacity.

[0012] As an optional approach, a rolling time-domain strategy is adopted to perform the solution: after each time slice ends, the time-extended network is reconstructed and solved based on the latest corrected remaining demand and vehicle status data. Only the scheduling instructions corresponding to the current time slice are executed, and the scheduling schemes for subsequent time slices are reserved as pre-planned plans.

[0013] As an optional approach, the scheduling instructions include a go instruction, a wait instruction, a continue service instruction, and a charging instruction; each scheduling instruction includes the vehicle number, instruction type, target node, and planned execution time slice.

[0014] According to a second aspect of the embodiments of this application, a sightseeing vehicle scheduling system based on multi-source data fusion is also provided, comprising: The data acquisition and service map construction module is used to acquire vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data to construct a scenic area service map; and to establish walking accessibility relationships based on walking time from each scenic area to each station, thereby determining the set of associated scenic spots for each station. The demand estimation module is used to calculate the demand contribution for each station based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each related attraction. The demand contribution of all related attractions, the number of reserved passengers and the number of waiting tourists are summed to obtain a comprehensive demand estimate. The demand correction module is used to obtain the actual number of passengers after a vehicle arrives at the station and completes passenger pick-up, and then subtract the actual number of passengers from the comprehensive demand estimate to obtain the corrected remaining demand. The scheduling optimization module is used to convert the revised remaining demand of each station into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Vehicle flow conservation constraints, seat capacity constraints, power constraints, road segment traffic capacity constraints, and station stop capacity constraints are set. The objective is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and then generate scheduling instructions.

[0015] This application integrates vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data into the scenic area service map. It also estimates the passenger demand at each station within subsequent time slots by considering the walking accessibility from attractions to stations. This expands the scheduling basis from the current queuing status at stations to a proactive assessment of potential arriving passenger flow. The actual number of passengers after vehicle arrival is used to adjust remaining demand and update the diversion ratio, reducing the impact of discrepancies between historical statistical parameters and actual passenger flow on subsequent scheduling. Furthermore, the adjusted demand is converted into service vehicle trips. In the extended time network, constraints such as vehicle flow conservation, seating capacity, battery power, road capacity, and station parking capacity are comprehensively considered. Scheduling instructions are generated with the combined cost of tourist waiting time and vehicle idle time as the objective, making vehicle allocation more aligned with the actual operating conditions of the scenic area, reducing ineffective dispatching and station congestion, and improving the accuracy and feasibility of sightseeing vehicle scheduling.

[0016] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Furthermore, no embodiment in this disclosure is required to achieve all the effects described above. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0018] Figure 1 A flowchart of a sightseeing vehicle scheduling method based on multi-source data fusion is provided for an embodiment of this disclosure.

[0019] Figure 2 A flowchart for obtaining demand estimates provided in embodiments of this disclosure.

[0020] Figure 3 A flowchart illustrating the requirement modification process provided in this embodiment of the disclosure.

[0021] Figure 4 A flowchart illustrating the scheduling optimization provided in this embodiment of the disclosure.

[0022] Figure 5 This is a schematic block diagram of a sightseeing vehicle scheduling system based on multi-source data fusion, provided as an embodiment of the present disclosure.

[0023] Figure 6 This is a schematic block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0025] The sightseeing vehicle scheduling method provided in this disclosure is applicable to transportation scenarios where fixed boarding stations are located within a scenic area, and vehicles operate along roads between these stations to transport tourists. The scenic area includes several attraction areas, boarding stations, parking areas, and charging facilities, all connected by internal roads. The sightseeing vehicles are electric. During operation, tourists walk to nearby stations after visiting the attraction areas. Due to the time-varying and spatially dispersed nature of tourist movement, the actual demand at each station is difficult to obtain before tourists arrive at the station. Traditional scheduling methods often rely on the current number of people queuing at a station or historical statistics to allocate vehicles, lacking foresight regarding the demand of tourists on the walk. This can easily lead to vehicles being dispatched to stations where demand has subsided or overlooking stations where demand is increasing.

[0026] Furthermore, some stations within the scenic area have limited space, accommodating only a small number of vehicles at a time. If the scheduling scheme does not consider this physical condition, multiple vehicles may be dispatched to the same station simultaneously, causing congestion. The method described in this embodiment acquires multi-source operational data to construct a scenic area service map. Based on the distribution of tourists in the scenic area and the walking access time relationship, it estimates the travel demand for each station in future time slices. It uses actual observation data after vehicles arrive at the station to correct demand deviations and update the estimated parameters. Under the condition of considering resource constraints such as road capacity and station parking capacity, it completes scheduling optimization through a time-extended network and generates scheduling instructions for each vehicle.

[0027] The implementation process of the method described in this application will be described in detail below with reference to specific embodiments. It should be noted that this embodiment is only used to explain this application and is not intended to limit the scope of protection of this application. Conventional adjustments or substitutions of each step by those skilled in the art without departing from the concept of this application should be included in the scope of protection of this application.

[0028] Please see Figure 1 , Figure 1 This is a flowchart of a sightseeing vehicle scheduling method based on multi-source data fusion according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes steps S1-S4: In step S1, vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data are obtained to construct a scenic area service map.

[0029] In some embodiments, vehicle operation data includes the current location, speed, number of remaining seats, and remaining battery power of each vehicle, used to determine the real-time status and dispatchability of each vehicle; tourist distribution data reflects the real-time number of tourists in each scenic area, which can be obtained through scenic area entrance / exit counting devices, regional video surveillance counting, or mobile terminal location statistics; station card swipe data includes historical and real-time card swipe records for each station; reservation data is the ride reservation information submitted by tourists through the scenic area reservation system, which records the reservation station, reservation time, and number of passengers; scenic area road data includes the start and end nodes, road length, speed limits, and capacity of each road segment, used to construct the road network topology within the scenic area and constrain vehicle travel paths and traffic flow in scheduling optimization. Among the above five types of data, vehicle operation data and tourist distribution data need to be updated at the beginning of each time slice to reflect the latest vehicle and tourist status; station card swipe data and reservation data are continuously collected; scenic area road data can remain unchanged when the scenic area road network structure does not change.

[0030] According to embodiments of this disclosure, a scenic area service map is constructed based on the aforementioned data. The scenic area service map is a graph structure with attributes. The node set includes: scenic spot nodes representing scenic areas within the scenic area, with attributes including the number of tourists and the area; station nodes representing bus stops, with attributes including geographical location and station capacity, where station capacity refers to the maximum number of vehicles allowed to stop at a station simultaneously, determined by the station's physical space and safety distance; parking area nodes representing vehicle parking areas; and charging point nodes representing the location of charging facilities, with attributes including charging power and the number of available charging spots. Edges in the scenic area service map represent road segments connecting the nodes. Each edge has attributes of segment length, travel time, and segment capacity, where segment capacity represents the maximum number of vehicles allowed to pass through the segment in a single time slice.

[0031] In one embodiment, a walkable reachability relationship is established between scenic spot nodes and station nodes in the scenic area service map. Specifically, for each station, the walking time from the boundaries of surrounding scenic spot areas to the station is calculated. The walking time is calculated based on the length of the walking path between the scenic spot area boundary and the station and the average walking speed. Scenic spots whose walking time is within a preset threshold range are included in the associated scenic spot set of the station. This threshold can be configured, for example, 10 to 20 minutes. Scenic spots exceeding the threshold are not associated with the station, and it is assumed that tourists in the scenic spot area will not walk to the station to take a ride. The associated scenic spot set determines which scenic spot areas' tourist flow will affect the ride demand of the station. In some embodiments, the walking time can also be obtained based on the actual walking route markings within the scenic area, which is not limited in this disclosure.

[0032] In step S2, for each station, the demand contribution is calculated based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each associated attraction. The demand contribution of all associated attractions, the number of reserved passengers and the number of waiting tourists are summed to obtain a comprehensive demand estimate.

[0033] In scenic area operation scenarios, the currently visible queue number at a station only reflects tourists who have arrived at the station. Tourists walking from the scenic area to the station constitute unobserved potential transportation demand. Since the walking time from different scenic spots to different stations varies, this unobserved demand will arrive at various stations at different future times. If the dispatch system only allocates vehicles based on the instantaneous queue information at the station, it cannot arrange transportation capacity in advance for the upcoming passenger flow, and tourists will face long waiting times after arriving at the station. To address this, this embodiment of the disclosure uses the current number of tourists in each scenic area, combined with the walking time of tourists from the scenic area to the station, to calculate how many tourists will walk to each station in the future time slice, thereby pre-allocating vehicles to the corresponding stations while tourists are still en route.

[0034] The scheduling time is divided into consecutive time slices, and the duration of each time slice is denoted as . The time slice can be configured to, for example, ten to fifteen minutes; the length of the time slice should be on the same order of magnitude as the walking time of tourists from the attraction to the station. Too short a time slice will increase the computational frequency of scheduling optimization, while too long a time slice will reduce the temporal resolution of demand estimation; the following uses the target time slice as an example. Taking this as an example, we will explain the specific process of demand estimation for each site.

[0035] like Figure 2 As shown, Figure 2 A flowchart for obtaining demand estimates provided in an embodiment of this disclosure is shown, wherein in block 201, for each site, the demand contribution is calculated based on the number of visitors, visitor departure rate, time slice length, and diversion ratio of each associated attraction.

[0036] Specifically, for the site Extract all scenic spot nodes that are within walking distance of the station from the scenic area service map to form a set of related scenic spots. For each attraction in the collection Get information from attractions Area to Station walking time The walking time can be calculated and stored during the scenic area service map construction phase; the distance from each attraction to the station in the associated attraction set. Walking times vary; closer attractions take less time to walk, while farther attractions take longer.

[0037] Tourists from the scenic spot Start walking to the station It takes time to walk. Therefore, a tourist at a certain moment from the attraction Departure will occur after the departure time. Arrival Station Conversely, this leads to the calculation that within the target time slice... Internal arrival stations Tourists from the attraction The departure time must fall within a specific interval, which is the effective departure time window; the start time of the effective departure time window is the start time of the target time slice. Subtract the corresponding walking time The end time is the end time of the target time slice. Subtract walking time .

[0038] Because the walking times from different related attractions to the same station vary, the effective departure time windows for each attraction also differ on the timeline: attractions with longer walking times have their effective departure time windows shifted further forward on the timeline, meaning the scheduling system needs to monitor visitor activity at those attractions earlier. For example, if attraction A is located at station... The walking time is 5 minutes from attraction B to the station. If the walking time is 15 minutes, then for the same target time slot, the effective departure time window for attraction B is shifted forward by 10 minutes compared to attraction A. Although tourists departing from attraction B have a longer walking time, if they are already on their way, they will arrive at the station within the target time slot. This generates travel demand; the time correspondence allows the scheduling system to predict which time slot a tourist will arrive at which station while they are still on foot.

[0039] For related attractions Obtain the number of tourists at the start time of the valid departure time window. This number comes from tourist distribution data and reflects the actual number of tourists in the scenic area at that moment; when the start time of the effective departure time window is not later than the current decision time, Use the observed or recorded values ​​from the tourist distribution data at that starting time; when the starting time of the effective departure time window is later than the current decision time, The tourist distribution data already acquired in the current time slice is used as the estimate of the number of tourists at the starting moment, and then recalculated based on the latest tourist distribution data during subsequent rolling time-domain updates.

[0040] Determine the attractions Tourist departure rate This represents the proportion of tourists leaving the scenic area within a unit of time, relative to the total number of tourists currently in the area. The value ranges from 0 to 1 in an open interval, and is expressed as a time slice length. When calculating the contribution of demand, This represents the percentage of tourists who leave the attraction area within a given time frame, and the percentage value does not exceed 1.

[0041] The visitor departure rate is influenced by factors such as attraction attractiveness, average visit duration, and weather conditions. Its initial value is calibrated based on historical entrance and exit count data for the attraction area: that is, by dividing the number of visitors leaving the attraction area within a certain historical period by the average number of visitors and the duration of that period, the calibrated value is obtained. Visitor departure rates can be calibrated separately for different attractions and different time periods. For example, the visitor departure rate at popular attractions is often higher during peak hours than during off-peak hours. Optionally, during operation, if actual observations show a persistent deviation between the visitor departure rate and the calibrated value, the visitor departure rate can be corrected. The correction method is similar to the diversion ratio correction described later and will not be elaborated further here.

[0042] At the same time, determine the diversion ratio. , indicating from the tourist attraction Departing tourists choose to go to the station The proportion of passengers using public transport. The passenger diversion ratio reflects the differences in tourist appeal among various stops around the attraction. For attractions... The sites that are within walking distance of the attraction will form a set of associated sites. The determination of the diversion ratio is based on the inverse relationship between walking time and the number of attractions. To related site collection The reciprocal of the walking time to each station within the area is used as the weight, and the weights are allocated proportionally. Specifically, for each station... The diversion ratio is equal to the walking time. The reciprocal of the number of attractions The ratio of traffic diversion to the destination is the sum of the reciprocals of the walking time to all its associated sites. Sites with shorter walking times receive a higher traffic diversion ratio, while sites with longer walking times receive a lower traffic diversion ratio.

[0043] The sum of the diversion ratios from the same attraction to all its associated stations does not exceed 1. The unassigned portion corresponds to tourists who left the attraction but did not choose the sightseeing vehicle service. The initial value of the diversion ratio is determined through historical card-swipe data statistics. When there are insufficient historical card-swipe records between some attractions and stations, the inverse normalized result of walking time can be used as the initial value of the diversion ratio for that part, or the historical statistical values ​​can be corrected using the inverse relationship of walking time so that stations with shorter walking times correspond to higher diversion ratios. Specifically, based on the card-swipe records of each station in historical operational data, the number of passengers from different directions of the attraction is counted, and the proportion of passengers from each direction to the total number of passengers leaving the attraction is calculated, which is used as the initial calibration value of the diversion ratio.

[0044] After determining the above parameters, the scenic spot Within the target time slice, the site The demand contribution is calculated as follows:

[0045] in For tourist attractions The number of tourists at the start of the valid departure time window. For tourist departure rate, The time slice length, For the diversion ratio, Multiply and Get from the attraction within a time slice Multiply the estimated number of departing tourists by Later, we obtained the destination station. The number of people, i.e., the demand contribution of the attraction to the site.

[0046] Optionally, in some embodiments, when the reservation data includes information about the tourist's location within the attraction area, the same group of tourists may be counted as both walking-source demand and reservation demand. To eliminate duplicate counting, when calculating the attraction... Before the demand contribution, The number of pre-booked tourists whose departure time falls within the valid departure time window is subtracted, and the number of tourists after deduction is limited to not less than zero. That is, the non-negative number of tourists after deduction is used to calculate the demand contribution, so that the demand from walking sources no longer includes tourists with clear reservations.

[0047] The number of tourists after deduction can be recorded as follows: ,in The number of reserved visitors whose departure time falls within the valid departure time window and the corresponding scenic spot is ; Used to replace Participate in the calculation of demand contribution.

[0048] In box 202, sum up the demand contributions of all associated scenic spots, the number of reserved passengers, and the number of waiting visitors to obtain the comprehensive demand estimate. Sum up the demand contributions of all associated scenic spots of the station to obtain the total value of the demand from walking sources. Specifically, extract the number of reserved passengers with the station as the boarding station within the target time slice , and at the same time obtain the number of visitors waiting at the station at the start time of the target time slice . This number can be directly collected by the station counting device or accumulated from the unmet demand at the station at the end of the previous time slice; the number of waiting visitors reflects the visitors who have arrived at the station but have not been served in the previous time slice, and this part of the demand will continue to exist in the current time slice; after summing up the three items, obtain the [[ID=2|1]]comprehensive demand estimate within the target time slice for the station:

[0049] Where is the set of associated scenic spots of the station, is the demand contribution of each scenic spot, is the number of reserved passengers, is the number of waiting visitors; the comprehensive demand estimate integrates three sources of unobserved demand during walking, confirmed reserved demand, and existing waiting demand at the station; among them, the demand from walking sources maps the number of visitors in the scenic spot to the future arrival volume at the station through the valid departure time window, the number of reserved passengers provides demand information with relatively high certainty, and the number of waiting visitors reflects the unmet demand accumulated in the previous time slice; after combining the three, the dispatching system makes a quantitative prediction of the boarding pressure at each station in the future time slice, providing a basis for early vehicle allocation. <00002|1>

[0050] In step S3, when the vehicle arrives at the station and finishes boarding passengers, obtain the actual number of boarding passengers, and subtract the actual number of boarding passengers from the comprehensive demand estimate to obtain the corrected remaining demand.

[0051] As mentioned earlier, the comprehensive demand estimate is based on tourist distribution data and statistical calibration parameters. Due to the randomness of tourist behavior and the accuracy of parameter calibration, the estimated results inevitably deviate from the actual situation. In actual operation, tourists at some attractions may leave earlier or later than expected, or may prefer stations that deviate from statistical patterns. These factors can cause the estimated value to deviate from the actual demand. Without correction, the error will gradually accumulate as time slices progress, making the scheduling of subsequent time slices based on inaccurate demand information. When a vehicle arrives at a station and completes the passenger pick-up operation, the actual number of passengers at that station during the vehicle's stop can be obtained. This data is directly obtainable ground observation information during the scheduling process, reflecting the actual demand situation at that station at that moment. This method uses the actual number of passengers to both correct the remaining demand in the current time slice and update the diversion ratio parameters used for demand estimation in subsequent time slices.

[0052] like Figure 3 As shown, Figure 3 A flowchart of the demand correction process provided in an embodiment of this disclosure is shown, wherein in block 301, after a vehicle arrives at the station and completes passenger pick-up, the actual number of passengers is obtained, and the revised remaining demand is obtained by subtracting the actual number of passengers from the comprehensive demand estimate.

[0053] Vehicle arrives at the station After completing the check-in process, obtain the actual number of guests. That is, the number of passengers who actually boarded the train at the station. If the station is equipped with card-swiping devices, the number of people recorded by card swiping can also be used as a reference or verification source for the actual number of passengers.

[0054] In reality, the number of passengers is constrained by the number of remaining seats on the vehicle: if the number of remaining seats when the vehicle arrives at the station is greater than or equal to the number of passengers currently waiting at the station, then all passengers who have arrived and are waiting can board the vehicle. The actual number of passengers is approximately equal to the total number of passengers currently waiting at the station, indicating that the waiting demand for the vehicle's arrival time has been met. Whether the overall demand for the current time slot at the station has been fully met can be determined based on the adjusted remaining demand. Whether it is zero or not is determined; if the number of remaining seats is less than the number of waiting tourists, then the actual number of passengers is equal to the number of remaining seats in the vehicle, and there is still unmet demand at the station.

[0055] The remaining demand for this site after corrections within the current time slice can be represented as:

[0056] in This is the overall demand estimate before the revision. The calculation uses the actual number of passengers, taking the maximum value to ensure that the adjusted remaining demand does not result in a negative value. If multiple vehicles stop at the same station within the same time slot, the actual passenger counts for each stop are summed sequentially and then substituted into the calculation. The adjusted remaining demand is then calculated. Replace the original comprehensive demand estimate The remaining demand for services at the current time slice will be included in subsequent scheduling optimizations. If the remaining demand is zero after the adjustment, it means that the demand for the current time slice at the current station has been fully met, and the station will no longer need additional vehicle services in the current time slice during subsequent scheduling optimizations. If the remaining demand is still greater than zero after the adjustment, the scheduling system will continue to arrange vehicles for the station in the next round of optimizations.

[0057] In box 302, the diversion ratio is updated based on the corrected remaining demand and actual observation data. The difference between the actual number of guests and the estimated comprehensive demand not only affects the judgment of the remaining demand for the current time slot, but also reflects, to some extent, the deviation between the estimated parameters such as the diversion ratio and the actual passenger flow distribution. In some embodiments, when the actual number of guests at a certain station continuously deviates from the estimated value within several consecutive time slots, the diversion ratio of the related attractions at that station should be adjusted to make the demand estimate for subsequent time slots closer to the actual situation.

[0058] Specifically, the adjustment process is as follows: Summarize the actual number of visitors to the site within the most recent time slots; the number of these recent time slots is denoted as... , A pre-defined positive integer greater than 1, and a smoothing coefficient. The calibration process is determined synchronously.

[0059] By combining the total number of tourists leaving each related attraction and the actual passenger distribution to each station during the same period, the actual diversion ratio from each related attraction to that station can be deduced. The actual passenger distribution at each station is determined by the information on the tourist's location within the attraction area contained in the station's card swipe data and reservation data, or by the change in tourist distribution data within adjacent time slots; for attractions where the source attraction can be identified, the most recent... Within a certain time frame, the attractions Head to the station The actual number of guests And statistics on attractions during the same period Total number of tourists leaving ,according to Calculate the actual diversion ratio; when If the passenger origin is zero or the source attraction cannot be determined, the passenger diversion ratio from that attraction to that station will not be updated; the previous diversion ratio will be used.

[0060] Then, an exponential smoothing method is used to weight and sum the actual traffic splitting ratio with the original traffic splitting ratio to obtain the updated traffic splitting ratio:

[0061] in To calculate the actual diversion ratio, The splitting ratio used before the update. This represents the smoothing coefficient. Exponential smoothing is a commonly used time series smoothing method. Its principle is to mix new observations with historical estimates in a fixed proportion. The larger the value, the higher the proportion of new observations, and the more sensitive the updated results are to the latest changes; The smaller the value, the better the historical values ​​are preserved, and the more stable the update results.

[0062] Among them, the smoothing coefficient The value range is an open interval between 0 and 1. It can be pre-calibrated according to the fluctuation characteristics of tourist flow in the scenic area, and can be configured to, for example, 0.2-0.5. For peak periods or popular attractions where tourist flow changes frequently, a larger smoothing coefficient value can be selected to make the update follow the actual changes more quickly; for periods or attractions with stable tourist flow patterns, a smaller value can be selected to maintain the stability of the estimate.

[0063] Updated traffic splitting ratio Replace original value This is used to calculate the demand contribution of this attraction to the site in subsequent time slices; for those due to For associations where the passenger originating attraction is zero or cannot be determined and therefore no update has been performed, the existing associations will continue to be used. It participates in subsequent time-slice calculations; optionally, when the deviation between the actual observed data and the estimated value exceeds a preset proportional threshold, the visitor departure rate of the relevant attractions can also be calculated using a similar exponential smoothing method. Make corrections.

[0064] Through the above correction process, the actual observation information is fed back to the demand estimation stage, so that the estimation parameters are continuously adjusted according to the actual passenger flow in operation; thereby avoiding the continuous transmission and amplification of demand estimation errors caused by parameter calibration deviations between consecutive time slices.

[0065] In step S4, the remaining demand after correction at each station is converted into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Vehicle flow conservation constraints, seat capacity constraints, power constraints, road segment traffic capacity constraints, and station stop capacity constraints are set. The goal is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and a scheduling instruction is generated.

[0066] According to the embodiments of this disclosure, after completing the demand estimation and correction for each station, the remaining demand after correction needs to be converted into a demand representation for vehicle services, and all available vehicles are uniformly scheduled by taking into account the current status of vehicles and the physical resource constraints of scenic roads and stations. In this embodiment, a time-extended network is constructed along the time dimension based on the scenic service map, and the vehicle scheduling problem is transformed into an optimization problem on the network for solution.

[0067] Specifically, such as Figure 4 As shown, Figure 4 A flowchart of the scheduling optimization provided in an embodiment of this disclosure is shown, wherein in block 401, the corrected remaining demand of each station is converted into the required number of service vehicles.

[0068] For the site In time slice Remaining demand after correction Calculate the minimum number of service vehicles required for this station in this time slice using the following formula:

[0069] in The average number of available seats per vehicle is calculated based on the remaining seats of vehicles eligible for scheduling in the current time slot; if the set of vehicles eligible for scheduling in the current time slot is... , For the number of vehicles, vehicles The number of remaining seats is ,but Can be taken as When the number of seats in the scenic area vehicles is the same and the vehicles were empty before being dispatched, The vehicle's rated seating capacity can be used. This indicates the rounding up operation, representing the number of service vehicles required. The demand, which is based on the number of people, is transformed into a service target based on the number of train trips. The subsequent optimization model then allocates vehicles to each station accordingly.

[0070] In box 402, based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. The time-extended network expands along the time dimension based on the spatial topology of the scenic area service map, assuming that the time range of the scheduling plan covers the current time slice. To the future A time slice Planning time domain length It can be configured to, for example, 4 to 8 time slices.

[0071] In some embodiments, nodes in the time-extended network are defined as node numbers in the scenic area service graph. With time slice number The binary pair formed In vehicle routing planning, stations, parking areas, and charging points that participate in vehicle passage, parking, or charging generate time extension nodes for vehicle route selection at each time layer; scenic spot nodes participate in demand estimation as demand source nodes, but generally do not participate in vehicle routing planning as vehicle parking service nodes; each station, parking area, and charging point generates a time extension node at each time layer, while scenic spot nodes do not participate in vehicle routing planning and do not generate nodes in the time extension network.

[0072] Specifically, the arc is divided into a driving arc, a waiting arc, and a charging arc. Among them, the driving arc extends from the time extension node. point to ,in For the node Drive to the node Number of time slices spanned , For road section to The passage time, The time slice length; the travel arc is only established between node pairs that have direct road segments connected in the scenic area service map, indicating that the vehicle travels from the node... In time slice Departure, in time slice Reaching the node .

[0073] Waiting arc from time extension node Pointing to the next time layer of the same node This indicates that the vehicle remains at this node for one time slice, and the charging arc starts from the charging point node. point to , The number of time slices required to complete a charging unit is determined by rounding up the ratio of the time required to complete the charging unit to the length of the time slice.

[0074] After expansion, a path from the initial node to the destination for a vehicle on the extended time network corresponds to its complete behavioral trajectory throughout the entire planning time domain, including where it travels to, waits at, and charges in each time slice. The initial node for each vehicle is determined by its current actual location and current time slice; specifically, the tuple consisting of the scenic area service map node corresponding to the current location in the vehicle's operational data and the current time slice number serves as the vehicle's starting node in the extended time network. The scheduling problem is thus transformed into simultaneously selecting paths for all available vehicles on the extended time network, maximizing the number of service trips required for each station and time slice while adhering to various resource constraints.

[0075] In box 403, constraints such as vehicle flow conservation, seat capacity, power consumption, road segment traffic capacity, and station stop capacity are set. The objective is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and a scheduling instruction is generated.

[0076] In one embodiment, the decision variable for scheduling optimization is the passage choice of each vehicle on each arc of the time-extended network. For example, the constraints cover the following five aspects.

[0077] The vehicle flow conservation constraint requires that the number of inflow arcs for each vehicle at each time extension node equals the number of outflow arcs, ensuring the continuity and integrity of the vehicle path; the starting node of each vehicle in the initial time slice is determined by its current actual position.

[0078] Seating capacity constraints limit the number of passengers a vehicle can carry after picking up passengers at any station to no more than the total number of seats in the vehicle. After a vehicle arrives at a station, the number of passengers to pick up is determined based on the number of service trips required for that station and the current number of seats remaining in the vehicle. When the vehicle arrives at the tourist's destination station, the number of passengers to pick up is reduced accordingly after the passenger disembarks.

[0079] The power constraint stipulates that when a vehicle moves along a driving arc, it consumes power according to the length of the driving segment, and when it passes through a charging arc, it replenishes power according to the charging power and charging time. The remaining power of the vehicle at any time extension node shall not be lower than the preset minimum power threshold. The setting of this threshold shall ensure that the vehicle can still drive to the nearest charging point for replenishment at its current position.

[0080] The road segment capacity constraint limits the total number of vehicles passing through the same road segment within the same time slot to no more than the upper limit of the road segment's capacity. Some road segments in scenic areas have narrow roads or complex road conditions, and the number of vehicles allowed to pass at the same time is limited. This constraint ensures that the scheduling plan does not cause traffic congestion at the road segment level.

[0081] The station's parking capacity constraint limits the total number of vehicles stopping at the same station within the same time slot to no more than the station's maximum parking capacity. This maximum capacity is determined by the station's physical space and safety clearance. Parked vehicles include those picking up / dropping off passengers and those waiting. If the number of dispatched vehicles arriving at a high-demand station within the same time slot exceeds its parking capacity, later arriving vehicles will be unable to park, leading to station congestion and preventing dispatch instructions from being executed. This constraint allows the optimization model to allocate vehicles in batches across time slots for stations with high demand but limited space. For example, the first batch of vehicles might arrive and pick up passengers in the current time slot before departing, and the second batch might arrive in the next time slot, rather than concentrating all vehicles in the same time slot. This batch scheduling ensures that the number of vehicles arriving at the station in each time slot does not exceed its physical carrying capacity.

[0082] The optimization objective is to minimize the weighted sum of tourist weighted waiting time and vehicle idle time while satisfying all constraints. Tourist weighted waiting time refers to the time difference between the moment demand is generated at each station and the moment a vehicle arrives at that station to provide service, multiplied by the number of passengers required at that station, and then summed. Vehicle idle time refers to the total time all vehicles travel without passengers. The weighting coefficients between the two are... The value range is an open interval between 0 and 1, which is preset by the operation and management team according to the scenic area's service strategy. When the value is too large, the optimization results tend to shorten the waiting time for tourists; When the value is relatively small, it tends to reduce empty runs of vehicles to save operating costs. Understandably, the specific value can be set according to the actual situation, and this disclosure does not impose any restrictions.

[0083] The aforementioned optimization problem can be modeled as a mixed integer programming problem, and is solved using the branch and bound method according to the embodiments of this disclosure. The branch and bound method first relaxes the integer constraints into continuous variables to obtain the lower bound of the objective function, and then branches the integer variables, with each branch corresponding to a set of subproblems for variable values. By comparing the lower bound of each subproblem with the current optimal feasible solution, branches inferior to the optimal solution are pruned, and the search space is gradually reduced until convergence. According to the embodiments of this disclosure, the decision variable is the zero- or one-pass choice of each vehicle on each arc, and the lower bound can be quickly obtained through linear programming after relaxation.

[0084] Specifically, in the time-extended network, the decision of whether each sightseeing vehicle passes through each candidate arc is set as a zero-to-one decision variable; for example, let... Indicates vehicle Do you want to select an arc? At that time, it indicates that the vehicle performed the action represented by that arc during the corresponding time slice, such as driving, waiting, or charging; when When the action is not selected, it indicates that the action is not selected. By uniformly solving the zero and one variables of each vehicle, each time layer, and each type of arc, the complete running path of the vehicle in the planning time domain can be determined, and constraints such as flow conservation, seating, power consumption, road segment capacity and station stopping capacity can be further satisfied to achieve comprehensive optimization of tourist waiting time and vehicle empty running time.

[0085] In the branch and bound solution, the restriction that zero and one variables must be 0 or 1 is temporarily removed, and the passage selection variables of vehicles on each arc are relaxed to continuous variables between 0 and 1. At this time, the model is transformed into a linear programming problem, and the theoretical lower bound of the objective function can be quickly obtained. This is used to determine whether the current branch may still produce a better scheduling scheme, thereby pruning invalid search branches and improving the solution efficiency.

[0086] After optimization, the optimal path for each vehicle within the entire planning time domain is obtained. The arcs that each vehicle should execute in the current time slice are extracted and converted into scheduling instructions. The arc selected by each vehicle in the current time slice corresponds to the action that the vehicle should perform within that time slice. In one embodiment, for example, the scheduling instructions include: a "go" instruction instructing the vehicle to travel from its current location along a designated road segment to the target station; a "wait" instruction instructing the vehicle to remain at the current station or parking area until the next time slice; a "continue service" instruction instructing the vehicle to continue operating along the predetermined route and pick up and drop off passengers at stations along the way; and a "charge" instruction instructing the vehicle to proceed to a designated charging point for charging. Each scheduling instruction includes the executing vehicle's number, instruction type, target node, and planned execution time slice, and is issued to the vehicle's onboard terminal for execution via the scenic area scheduling management system.

[0087] In some embodiments, a rolling time-domain strategy can be used to perform the above scheduling optimization in actual operation. Specifically, at the end of each time slice, the scheduling system reacquires the latest vehicle status data such as the location, remaining seats, and remaining battery power of each vehicle, updates the tourist distribution data, and redetermines the number of tourists corresponding to each valid departure time window in the subsequent target time slice based on the updated tourist distribution data. Demand corrections are performed on the stations where existing vehicles stop. On this basis, a time extension network covering several subsequent time slices is reconstructed and solved. The scheduling system only executes the scheduling instructions corresponding to the current time slice, and the scheduling schemes for subsequent time slices are reserved as pre-plannings but not issued immediately.

[0088] In one embodiment, the window length of the rolling time domain can be configured to, for example, 3 to 6 time slices. The rolling replanning method on a time slice basis allows the scheduling system to optimize based on the latest corrected remaining demand and vehicle status data at each decision moment. The pre-planned scheme generated in the previous time slice solution is replaced by the new scheme based on the latest data in the new round of optimization. In actual implementation, if the corrected remaining demand of a certain station increases significantly in the new round of correction, the scheduling system will allocate more service trips to that station accordingly during re-optimization; if a road segment cannot pass the required number of vehicles in the same time slice due to the upper limit of traffic capacity, the scheduling system will arrange alternative routes or adjust the departure time slice for some vehicles; when the remaining battery power of a vehicle is close to the minimum battery power threshold, the optimization solution will arrange a charging arc for the vehicle to ensure that it does not interrupt operation due to battery depletion.

[0089] Therefore, this embodiment of the disclosure, through the coordination of demand estimation, demand correction, and time-extended network scheduling optimization, generates a scheduling scheme that simultaneously considers multiple constraints such as station stop capacity, road traffic capacity, vehicle seating capacity, and battery power in scenic area scenarios where tourist demand information is delayed and uncertain. The distribution of tourists in the scenic area is mapped to the future travel demand of each station through walkability relationships and effective departure time windows, ensuring that vehicle dispatch precedes tourist arrivals. Demand estimates are corrected using the actual number of passengers, and the diversion ratio parameter is updated exponentially to reduce the cumulative propagation of estimation errors. Simultaneously, road segment traffic capacity constraints and station stop capacity constraints are introduced into the time-extended network to ensure that the generated scheduling instructions are physically executable, reducing misscheduling caused by delayed passenger flow data and limited station capacity, and improving the accuracy and operational efficiency of scenic area sightseeing vehicle scheduling.

[0090] Please see Figure 5 , Figure 5 This is a structural block diagram of a sightseeing vehicle scheduling system based on multi-source data fusion, provided in an embodiment of this application. Figure 5 As shown, the system includes: The data acquisition and service map construction module 501 is used to acquire vehicle operation data, tourist distribution data, station card swipe data, reservation data and scenic road data, and construct a scenic service map; establish walking accessibility relationships based on walking time from each scenic spot area to each station, and determine the set of associated scenic spots for each station; The demand estimation module 502 is used to calculate the demand contribution for each station based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each related attraction, and to sum the demand contribution of all related attractions, the number of reserved passengers and the number of waiting tourists to obtain a comprehensive demand estimate. The demand correction module 503 is used to obtain the actual number of passengers after the vehicle arrives at the station and completes the passenger pick-up, and to subtract the actual number of passengers from the comprehensive demand estimate to obtain the corrected remaining demand. The scheduling optimization module 504 is used to convert the revised remaining demand of each station into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Vehicle flow conservation constraints, seat capacity constraints, power constraints, road segment traffic capacity constraints, and station stop capacity constraints are set. The objective is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and a scheduling instruction is generated.

[0091] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0092] Please see Figure 6It shows a schematic block diagram of an electronic device according to an embodiment of this application, which can be used to implement... Figure 1 The method in the illustrated embodiment. (As shown) Figure 6 As shown, the electronic device may include: The system includes at least one processor 601, at least one network interface 604, a user interface 603, a memory 605, and at least one communication bus 602. The communication bus 602 is used to enable connection and communication between the components. The user interface 603 may include buttons, and optionally include a standard wired or wireless interface. The network interface 604 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.

[0093] The processor 601 may include one or more processing cores and connect to various parts within the electronic device through various interfaces and lines. It implements various functions and data processing of the electronic device by running or executing instructions, programs, code sets, or instruction sets stored in the memory 605, and by accessing data in the memory 605. Optionally, the processor 601 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The processor 601 may also integrate one or more combinations of CPU, GPU, and modem.

[0094] Memory 605 may include random access memory (RAM) or read-only memory (ROM). Optionally, memory 605 includes a non-transitory computer-readable medium for storing instructions, programs, code, code sets, or instruction sets. Memory 605 may be divided into a program storage area and a data storage area, wherein the program storage area can be used to store instructions for implementing an operating system and instructions for implementing the foregoing method embodiments; the data storage area can be used to store data related to the relevant method embodiments. Memory 605 may also be at least one storage device located remotely from processor 601. Figure 6 As shown, the memory 605, which serves as a computer storage medium, may contain an operating system, a network communication module, a user interface module, and program instructions.

[0095] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by processor 601, it performs the functions defined in the methods of this application.

[0096] Another embodiment of this application provides a storage medium storing computer program instructions thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0097] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. A sightseeing vehicle scheduling method based on multi-source data fusion, characterized in that, include: Acquire vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data to construct a scenic area service map; Establish walking accessibility relationships based on walking time from each scenic area to each station, and determine the set of associated scenic spots for each station; For each station, the demand contribution is calculated based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each related attraction. The demand contribution of all related attractions, the number of reserved passengers and the number of waiting passengers are summed to obtain the comprehensive demand estimate. Once the vehicle arrives at the station and completes passenger pick-up, the actual number of passengers is obtained. The remaining demand is then corrected by subtracting the actual number of passengers from the overall demand estimate. The remaining demand after correction at each station is converted into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Constraints such as vehicle flow conservation, seat capacity, power consumption, road segment traffic capacity, and station stop capacity are set. The goal is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and a scheduling instruction is generated.

2. The method according to claim 1, characterized in that, The scenic area service map includes scenic spot nodes, station nodes, parking area nodes, and charging point nodes; the attributes of station nodes include geographical location and station parking capacity, and the attributes of charging point nodes include charging power and number of available charging positions; the edges of the scenic area service map represent road segments connecting each node, and each edge has road segment length, travel time, and road segment travel capacity.

3. The method according to claim 1, characterized in that, The calculation of demand contribution based on the number of tourists, tourist departure rate, time slot length, and diversion ratio of each related attraction includes: For each associated attraction at a station, an effective departure time window is determined based on the walking time from the attraction to the station and the start and end times of the target time slice. The start time of the effective departure time window is the start time of the target time slice minus the corresponding walking time, and the end time is the end time of the target time slice minus the corresponding walking time. The number of tourists at the attraction at the start time of the effective departure time window is obtained. This number of tourists is multiplied by the tourist departure rate of the attraction, the length of the time slice, and the diversion ratio from the attraction to the station to obtain the demand contribution of the attraction to the station.

4. The method according to claim 1, characterized in that, The diversion ratio is determined based on the inverse relationship between the walking time from the associated attractions to each station, with stations having a higher diversion ratio corresponding to shorter walking times. The initial value of the diversion ratio is determined through historical card swipe data statistics.

5. The method according to claim 1, characterized in that, After obtaining the actual number of visitors, the actual diversion ratio of each related attraction to the station is inferred from the actual number of visitors in the most recent time slots. The actual diversion ratio is weighted and summed with the diversion ratio before the update using an exponential smoothing method to obtain the updated diversion ratio, which is used to calculate the demand contribution of subsequent time slots. The smoothing coefficient of the exponential smoothing is an open interval between 0 and 1, and the smoothing coefficient is pre-calibrated according to the fluctuation characteristics of the scenic area's visitor flow.

6. The method according to claim 1, characterized in that, The nodes of the time-extended network are tuples composed of node numbers and time slice numbers in the scenic area service map; the arcs of the time-extended network include driving arcs, waiting arcs, and charging arcs; driving arcs connect time-extended nodes of different nodes in the corresponding time layer of road segments, waiting arcs connect time-extended nodes of the same node in adjacent time layers, and charging arcs connect time-extended nodes of charging point nodes in the starting time layer and the corresponding time layer after charging is completed.

7. The method according to claim 1, characterized in that, The station parking capacity constraint is: within the same time slot, the total number of vehicles stopping at the same station shall not exceed the station's parking capacity limit; the road segment traffic capacity constraint is: within the same time slot, the total number of vehicles passing through the same road segment shall not exceed the road segment's traffic capacity limit.

8. The method according to claim 1, characterized in that, The solution is executed using a rolling time-domain strategy: after each time slice ends, the time-extended network is reconstructed and solved based on the latest corrected remaining demand and vehicle status data. Only the scheduling instructions corresponding to the current time slice are executed, and the scheduling schemes for subsequent time slices are reserved as pre-planned plans.

9. The method according to claim 1, characterized in that, The scheduling instructions include heading instructions, waiting instructions, continuing service instructions, and charging instructions; each scheduling instruction includes the vehicle number, instruction type, target node, and planned execution time slice.

10. A sightseeing vehicle scheduling system based on multi-source data fusion, characterized in that, include: The data acquisition and service map construction module is used to acquire vehicle operation data, tourist distribution data, station card swipe data, reservation data, and scenic area road data to construct a scenic area service map; and to establish walking accessibility relationships based on walking time from each scenic area to each station, thereby determining the set of associated scenic spots for each station. The demand estimation module is used to calculate the demand contribution for each station based on the number of tourists, tourist departure rate, time slot length and diversion ratio of each related attraction. The demand contribution of all related attractions, the number of reserved passengers and the number of waiting tourists are summed to obtain a comprehensive demand estimate. The demand correction module is used to obtain the actual number of passengers after a vehicle arrives at the station and completes passenger pick-up, and then subtract the actual number of passengers from the comprehensive demand estimate to obtain the corrected remaining demand. The scheduling optimization module is used to convert the revised remaining demand of each station into the required number of service vehicles. Based on the required number of service vehicles, a time-extended network is constructed along the time dimension on the basis of the scenic area service map. Vehicle flow conservation constraints, seat capacity constraints, power constraints, road segment traffic capacity constraints, and station stop capacity constraints are set. The objective is to minimize the weighted sum of tourist waiting time and vehicle empty running time, and then generate scheduling instructions.

Images