Regular bus optimal scheduling method and system based on equilibrium constraint mathematical programming

Abstract

The invention discloses a regular bus optimal scheduling method and system based on equilibrium constraint mathematical programming, and belongs to optimal scheduling, and the method comprises the steps: receiving travel request data of a user; clustering and grouping the users based on the destinations and the time; for each passenger group, traffic flow data and real-time monitoring images of candidate get-off points are collected, region features are identified based on image data, and a target get-off point is determined in combination with passenger destination positions; based on the traffic flow data, calculating the prediction cost of each candidate path to the target get-off point through a path prediction algorithm; and by using the path prediction cost, constructing an equilibrium constraint mathematical programming model with the purpose of minimizing the driving cost, and solving to obtain an optimal scheduling path scheme, thereby performing optimal scheduling on the target vehicle. According to the invention, through data processing, algorithm analysis and a mathematical planning model, a plurality of passengers with the same requirements can be quickly sent to the destination.

Description

Technical Field

[0001] This invention relates to the field of vehicle scheduling, and in particular to a method and system for optimizing bus scheduling based on equilibrium constraint mathematical programming. Background Technology

[0002] In urban commuting scenarios, office buildings attract a large number of passengers with similar destinations and time requirements. Traditional shuttle buses, with their fixed routes and stops, result in long journey times, uneven passenger waiting times, and high vehicle vacancy rates, increasing operating costs and reducing passenger experience. Therefore, achieving intelligent and optimized shuttle bus scheduling is crucial for improving the utilization rate of transportation resources, reducing operating costs, and meeting users' needs for fast and direct transportation.

[0003] To improve dispatching efficiency, existing technologies largely focus on using computers for demand analysis and route planning. Common methods include: using clustering algorithms to roughly group passengers' spatiotemporal demands, or introducing real-time traffic data to plan better routes. These methods improve the targeting of dispatching to some extent. However, they typically treat "demand grouping" and "route planning" as two separate, sequential steps, which directly affects the overall efficiency of bus dispatching. Summary of the Invention

[0004] This invention provides a shuttle bus optimization scheduling method and system based on equilibrium constraint mathematical programming, which can quickly transport multiple passengers with the same needs to their destination and improve the efficiency of shuttle bus scheduling.

[0005] An embodiment of the present invention provides a method for optimizing bus scheduling based on equilibrium-constrained mathematical programming, comprising: Receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time. Based on the destination area to which the destination location belongs and the expected arrival time period, users are grouped using a clustering algorithm to form several target passenger groups; For each target passenger group, traffic flow data from the office building to each destination area during the expected arrival time period is collected, as well as real-time monitoring images of several candidate drop-off points in each destination area. The real-time monitoring images are analyzed using image analysis algorithms to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, the target drop-off point corresponding to the target passenger group is determined. Based on the traffic flow data and the regional feature information, the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point is calculated by the path prediction algorithm. The equilibrium constraint mathematical programming model is constructed using the path prediction costs. The scheduling path scheme is obtained by solving the scheme with the goal of minimizing the driving cost. The target vehicle is determined based on the vehicle status data.

[0006] This invention provides a precise and original data foundation for subsequent intelligent demand aggregation and route planning by acquiring users' personalized travel needs. Through intelligent clustering of spatiotemporal demands, it can automatically and efficiently integrate scattered, individual travel requests into several target passenger groups with a common destination area and similar time windows. This achieves large-scale demand aggregation, transforming the problem of "serving multiple people" into "serving several groups," fundamentally creating conditions for designing efficient and intensive shuttle bus routes. By introducing and analyzing real-time monitoring images of the destination area, it can dynamically identify the actual regional characteristics of each candidate drop-off point, thereby intelligently selecting a route for each passenger group that comprehensively considers walking convenience and the suitability of the real-time drop-off environment. The optimal drop-off point not only avoids local delays or safety hazards caused by improper drop-off point selection, but also provides a more reasonable and feasible destination node for subsequent global route planning, ensuring the efficiency and reliability of the last mile in the rapid delivery process. By estimating the future cost of each alternative route through a route prediction algorithm, and using this as the basis, an equilibrium-constrained mathematical programming model with the goal of minimizing system operating costs is constructed for solution. This method can systematically weigh the needs of all passenger groups, the costs of all available routes, and the status of vehicle resources, generating a scheduling route plan from a globally optimal perspective. This scientifically directs target vehicles to serve each passenger group in the most efficient order and route, ultimately maximizing the improvement of bus scheduling efficiency at the system level.

[0007] Furthermore, based on the destination location's destination area and the expected arrival time, users are grouped using a clustering algorithm to form several target passenger groups, including: The destination area to which each of the stated destination locations belongs is determined using a geographic information system; Users with the same destination region are clustered using a first clustering algorithm to divide the users into several initial passenger groups; The second clustering algorithm is used to cluster users in each initial passenger group according to their expected arrival time periods, so as to aggregate users whose expected arrival time periods overlap or have an interval of less than a preset threshold into a subgroup, thereby obtaining several target passenger groups.

[0008] By intelligently clustering spatiotemporal demand, scattered and individual travel requests can be automatically and efficiently integrated into several target passenger groups with a common destination area and similar time windows. This achieves large-scale aggregation of demand, transforming the problem of "serving multiple people" into the problem of "serving several groups," fundamentally creating conditions for designing efficient and intensive shuttle bus routes.

[0009] Furthermore, the step of analyzing the real-time monitoring image using an image analysis algorithm to identify the regional feature information of each of the candidate drop-off points includes: The real-time monitoring images are preprocessed to obtain preprocessed images; The preprocessed image is subjected to target detection by an image analysis algorithm to obtain several target detection boxes. Semantic segmentation is then performed on each target detection box to extract regional feature information of each candidate drop-off point. The regional feature information includes at least one of the following: pedestrian density, traffic sign status, distribution of stationary obstacles, and road surface condition.

[0010] By introducing and analyzing real-time monitoring images of the destination area, the actual regional characteristics of each candidate drop-off point can be dynamically identified, thereby intelligently selecting the optimal target drop-off point for each passenger group that comprehensively considers walking convenience and the suitability of the real-time drop-off environment.

[0011] Further, determining the target drop-off point corresponding to the target passenger group based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information includes: Calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; The estimated walking time is corrected based on real-time weather information to obtain the target walking time; The regional risk coefficient of each candidate drop-off point is calculated using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. The overall suitability of each candidate drop-off point is calculated by combining the target walking time and the regional risk coefficient. The candidate drop-off point with the highest overall suitability is selected as the target drop-off point for the target passenger group.

[0012] By determining the target drop-off point, local delays or safety hazards caused by improper drop-off point selection are avoided. It also provides a more reasonable and feasible destination node for subsequent global route planning, ensuring the efficiency and reliability of the last mile in the rapid delivery process.

[0013] Further, the step of calculating the path prediction cost of each candidate driving path for the target vehicle to travel from the office building to the target drop-off point using a path prediction algorithm based on the traffic flow data and the regional feature information includes: The target vehicle is determined based on the vehicle status data of available shuttle buses in the office building; Based on traffic control rules, several candidate driving routes for the target vehicle from the office building to the target drop-off point are determined; Historical traffic flow data, the traffic flow data, and reported traffic incident information are input into a preset fusion model to predict the baseline travel time and congestion probability of each candidate travel route within the expected arrival time period. Based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, the baseline travel time and the estimated energy consumption cost are corrected to obtain the target travel time and the target energy consumption cost. Based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path, the path risk coefficient of each path is calculated. The path prediction cost for each candidate travel path is obtained by integrating the target travel time, the target energy consumption cost, and the path risk coefficient.

[0014] This allows for the estimation of future costs for each alternative path using a path prediction algorithm, facilitating the subsequent determination of the scheduling path scheme.

[0015] Furthermore, the determination of the target vehicle based on the vehicle status data of available office building shuttle buses includes: Obtain vehicle status data for all available shuttle buses in the office building, wherein the vehicle status data includes remaining driving range and rated passenger capacity; Based on the rated passenger capacity of each available shuttle bus, a set of candidate vehicles whose rated passenger capacity is greater than the total number of people in the target passenger group is selected. Vehicles in the candidate vehicle set whose remaining driving range is greater than a preset threshold are identified as the target vehicles.

[0016] Furthermore, the step of constructing an equilibrium-constrained mathematical programming model using the predicted costs of each path, and solving it with the objective of minimizing travel costs to obtain a scheduling route scheme, includes: Based on each candidate driving path and the corresponding path prediction cost, an equilibrium constrained mathematical programming model is constructed with the goal of minimizing driving costs. The equilibrium constrained mathematical programming model includes coverage constraints, resource constraints, and decision variable constraints. The equilibrium constraint mathematical programming model is solved, and the set of candidate driving paths corresponding to the decision variables with a value of 1 in the solution results is determined as the scheduling path scheme.

[0017] By using path prediction costs as the basic data, a balanced constrained mathematical programming model is constructed with the goal of minimizing the system's operating costs. This method can systematically weigh the needs of all passenger groups, the costs of all available paths, and the status of vehicle resources, generating a scheduling route plan from a globally optimal perspective. This allows the target vehicles to be scientifically directed to serve each passenger group in the most efficient order and route, ultimately maximizing the improvement of bus scheduling efficiency at the system level.

[0018] Another embodiment of the present invention provides a bus optimization scheduling system based on equilibrium constraint mathematical programming, comprising: The receiving module is used to receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time period; The processing module is used to group users into several target passenger groups based on the destination area to which the destination location belongs and the expected arrival time period, using a clustering algorithm. The determination module is used to collect traffic flow data from the office building to each destination area within the expected arrival time period for each target passenger group, as well as real-time monitoring images of several candidate drop-off points in each destination area. The real-time monitoring images are analyzed using an image analysis algorithm to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, the target drop-off point corresponding to the target passenger group is determined. The scheduling module is used to calculate the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point based on the traffic flow data and the regional feature information, and to construct an equilibrium constraint mathematical programming model using the path prediction costs, and solve the scheduling path scheme with the goal of minimizing the driving cost, for optimizing the scheduling of the target vehicle, wherein the target vehicle is determined according to vehicle status data.

[0019] Furthermore, the determining module includes: The first calculation unit is used to calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; The first correction unit is used to correct the estimated walking time based on real-time weather information to obtain the target walking time. The second calculation unit is used to calculate the regional risk coefficient of each of the candidate drop-off points using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. The third calculation unit is used to calculate the overall suitability of each of the candidate drop-off points by combining the target walking time and the regional risk coefficient. The first determining unit is used to select the candidate drop-off point with the highest overall suitability as the target drop-off point corresponding to the target passenger group.

[0020] Furthermore, the scheduling module includes: The second determining unit is used to determine the target vehicle based on the vehicle status data of available shuttle buses in the office building; The third determining unit is used to determine several candidate driving routes for the target vehicle from the office building to the target drop-off point in combination with traffic control rules; The prediction unit is used to input historical traffic flow data, the traffic flow data and the reported traffic incident information into a preset fusion model to predict the baseline travel time and the probability of congestion for each candidate travel route within the expected arrival time period. The second correction unit is used to correct the baseline travel time and the estimated energy consumption cost based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, so as to obtain the target travel time and the target energy consumption cost. The fourth calculation unit is used to calculate the path risk coefficient of each path based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path. The fifth calculation unit is used to integrate the target travel time, the target energy consumption cost, and the path risk coefficient to obtain the path prediction cost of each candidate travel path. Attached Figure Description

[0021] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating an embodiment of the bus optimization scheduling method based on equilibrium constraint mathematical programming provided in this application. Figure 2 This is a flowchart illustrating one embodiment of steps S201 to S203 provided in this application; Figure 3 This is a flowchart illustrating one embodiment of steps S301 to S305 provided in this application; Figure 4 This is a flowchart illustrating one embodiment of steps S401 to S406 provided in this application; Figure 5 This is a schematic diagram of an embodiment of the bus optimization scheduling system based on equilibrium constraint mathematical programming provided in this application. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0025] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0026] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0027] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0028] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0029] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0030] In urban commuting scenarios, office buildings attract a large number of passengers with similar destinations and time requirements. Traditional shuttle buses, with their fixed routes and stops, result in long journey times, uneven passenger waiting times, and high vehicle vacancy rates, increasing operating costs and reducing passenger experience. Therefore, achieving intelligent and optimized shuttle bus scheduling is crucial for improving the utilization rate of transportation resources, reducing operating costs, and meeting users' needs for fast and direct transportation.

[0031] See Figure 1 To quickly transport multiple passengers with the same needs to their destination, an embodiment of the present invention provides a bus optimization scheduling method based on equilibrium constraint mathematical programming, including steps S101 to S104. Step S101: Receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time. In some embodiments, this is accomplished in conjunction with a backend service system deployed on a cloud server or local server. Specifically, firstly, when a user submits a travel request through an interactive interface deployed on the user's mobile terminal (such as a mobile APP, WeChat mini-program, or web page frontend), they need to input or select the destination location and the expected arrival time period. After the user submits the request, these two pieces of data are encapsulated into a structured request message and sent to the specified API interface of the backend server via the HTTPS protocol. When the backend server receives this message, it parses and verifies the message to ensure that the data format is correct, the coordinates are valid, and the time period is reasonable. If the verification passes, the valid travel request data is stored in a temporary request pool or database.

[0032] It should be noted that the destination location can be selected by selecting a point on the map, entering a specific address, or selecting from a list of preset popular locations. The selection result will be converted into a standard geographic information format containing latitude and longitude coordinates. The desired arrival time period can be specified by the time selector, which specifies a time window with a clear start and end time (e.g., "17:30-18:00").

[0033] Step S102: Based on the destination area to which the destination location belongs and the expected arrival time period, the users are grouped using a clustering algorithm to form several target passenger groups. Please refer to Figure 2 In some embodiments, step S102 includes steps S201 to S203: Step S201: Use a geographic information system to determine the destination area to which each of the destination locations belongs; In some embodiments, an integrated geographic information system (GIS) service or library (such as a web-based map API or a local GIS engine) is invoked to quickly match each destination location obtained from the user's request with predefined polygon boundaries in the geographic information system, thereby assigning a unique destination area identifier to each user. At this point, it can be determined which key area each user's destination location belongs to.

[0034] Step S202: Cluster users with the same destination area using the first clustering algorithm to divide the users into several initial passenger groups; In some embodiments, a classification clustering algorithm is used to traverse all user requests that have been assigned destination area identifiers. The identifier is used as the clustering key, and a hash table or dictionary data structure is used for fast grouping. That is, users with the same identifier are grouped into the same set. Once the classification is complete, several initial passenger groups can be obtained.

[0035] Step S203: Using the second clustering algorithm, clustering is performed based on the expected arrival time periods of each user in each initial passenger group, so as to aggregate users whose expected arrival time periods overlap or whose intervals are less than a preset threshold into a subgroup, thereby obtaining several target passenger groups.

[0036] In some embodiments, after determining several initial passenger groups, since each user's expected arrival time is different, if two users' expected arrival times overlap, or if they do not overlap but the interval between the end time of the earlier time period and the start time of the later time period is less than a system-preset threshold (e.g., 10 minutes), it indicates that their expected arrival times are similar. In this case, the two users are considered time-compatible and grouped into the same subgroup. This process will iterate until all time-compatible users within each initial passenger group have been aggregated. Finally, each initial passenger group is refined into one or more target passenger groups, where users not only have similar destinations but also highly similar expected arrival times.

[0037] By intelligently clustering spatiotemporal demand, scattered and individual travel requests can be automatically and efficiently integrated into several target passenger groups with a common destination area and similar time windows. This achieves large-scale aggregation of demand, transforming the problem of "serving multiple people" into the problem of "serving several groups," fundamentally creating conditions for designing efficient and intensive shuttle bus routes.

[0038] Step S103: For each target passenger group, collect traffic flow data from the office building to each destination area within the expected arrival time period, as well as real-time monitoring images of several candidate drop-off points in each destination area. Analyze the real-time monitoring images using an image analysis algorithm to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, determine the target drop-off point corresponding to the target passenger group. In some embodiments, for each target passenger group, traffic flow data from the office building to each of the destination areas during the expected arrival time period is collected, along with real-time monitoring images of several candidate drop-off points in each of the destination areas. Specifically, after determining the target passenger group, a real-time traffic data service interface (such as a route planning API or a data platform of a traffic management department) is invoked. Using the office building coordinates as the starting point and the geographic center or outline of the destination area of ​​the passengers in the target passenger group as the endpoint, estimated or real-time traffic flow data for the expected arrival time period is requested. This traffic flow data typically includes average vehicle speed, travel time, congestion index, and event information. Simultaneously, real-time monitoring image streams or snapshots of several pre-recorded candidate drop-off points (such as bus stops, sections of road allowing temporary parking, etc.) located near the same target time period are retrieved from the urban traffic monitoring system or the security system API of a specific area.

[0039] In some embodiments, the step of analyzing the real-time monitoring image using an image analysis algorithm to identify the regional feature information of each candidate drop-off point includes: preprocessing the real-time monitoring image to obtain a preprocessed image; performing target detection on the preprocessed image using an image analysis algorithm to obtain several target detection boxes, and performing semantic segmentation on each target detection box to extract the regional feature information of each candidate drop-off point, wherein the regional feature information includes at least one of the following: pedestrian density, traffic sign status, distribution of stationary obstacles, and road surface conditions. Specifically, firstly, the received real-time monitoring image of each candidate drop-off point is filtered using methods including Gaussian filtering or median filtering, and the contrast of the filtering result is enhanced by histogram equalization or adaptive contrast enhancement algorithms to obtain a preprocessed image. Subsequently, the preprocessed image is input into an integrated image analysis algorithm model (e.g., the YOLO model) to identify and select various targets in the preprocessed image, generating several target detection boxes containing category and location information, wherein the categories include stationary obstacles such as pedestrians and traffic signs, as well as road potholes and water accumulation areas. Next, the model performs pixel-level semantic segmentation on the regions within each target detection box and the background region of the image, accurately distinguishing semantic categories such as pedestrians, passable road surfaces, obstacles, traffic signs, and abnormal road surfaces, and calculates region feature information based on the pixel-level label map output by semantic segmentation. It should be noted that pedestrian density is calculated by statistically analyzing the proportion of the number of pixels of different pedestrian types to the total area of ​​the image or region of interest; traffic sign status is judged by recognizing the detection boxes of traffic signs and combining them with their visual integrity; the distribution of stationary obstacles is evaluated by statistically analyzing the spatial clustering and occupied area of ​​obstacle category pixels; and road surface conditions are determined by analyzing the distribution and morphology of abnormal road surface category pixels.

[0040] By introducing and analyzing real-time monitoring images of the destination area, the actual regional characteristics of each candidate drop-off point can be dynamically identified, thereby intelligently selecting the optimal target drop-off point for each passenger group that comprehensively considers walking convenience and the suitability of the real-time drop-off environment.

[0041] Please refer to Figure 3 In some embodiments, determining the target drop-off point corresponding to the target passenger group based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information includes steps S301 to S305: Step S301: Calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; In some embodiments, each candidate drop-off point is first calculated using Geographic Information System (GIS) services. To the destination location of each passenger within the target passenger group The actual walkable pedestrian path length is denoted as . Next, a normal average walking speed is preset (e.g., 80 meters per minute), and the walking distance is divided by this average walking speed. This allows for a preliminary estimate of the estimated walking time for each passenger. .

[0042] Step S302: Based on the real-time acquired weather information, the estimated walking time is corrected to obtain the target walking time; In some embodiments, since certain weather information (such as rain, snow, strong winds, etc.) can directly affect pedestrian walking speed, a preset speed correction coefficient is applied in advance based on the weather type and intensity. (For example, during light rain) (This indicates a 20% increase in walking time), and this speed correction factor is applied to the estimated walking time. Make corrections, that is Therefore, the target walking time can be obtained.

[0043] Step S303: Calculate the regional risk coefficient of each candidate drop-off point using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. In some embodiments, regional feature information (including pedestrian density) obtained from real-time monitoring images is utilized. Traffic sign status rating Stationary obstacle distribution score Road surface condition rating The regional risk coefficient of each candidate drop-off point is obtained by weighting and summing various features using a pre-defined safety scoring model. The calculation formula is: ,in, , , and The preset weights for each feature reflect their impact on disembarkation safety.

[0044] Step S304: Calculate the overall suitability of each candidate drop-off point by combining the target walking time and the regional risk coefficient. In some embodiments, when the target walking time and regional risk coefficient Next, both are normalized to eliminate the influence of dimensions; then, adjusting weights are used. and (satisfy + =1) Target walking time and regional risk coefficient The scores are weighted to obtain a comprehensive score. In the formula, where The maximum tolerable walking time set for the system, where a higher value indicates a more suitable drop-off point.

[0045] Step S305: Select the candidate drop-off point with the highest overall suitability as the target drop-off point corresponding to the target passenger group.

[0046] In some embodiments, after obtaining the overall suitability of all candidate drop-off points Compare the overall suitability of all candidate drop-off points serving the target passenger group. The candidate drop-off point with the highest overall suitability is selected, and the final drop-off point is determined as the target drop-off point serving the target passenger group. This point represents the optimal drop-off location at the current moment, striking the best balance between pedestrian accessibility and the safety of the drop-off environment.

[0047] By determining the target drop-off point, local delays or safety hazards caused by improper drop-off point selection are avoided. It also provides a more reasonable and feasible destination node for subsequent global route planning, ensuring the efficiency and reliability of the last mile in the rapid delivery process.

[0048] Step S104: Based on the traffic flow data and the regional feature information, the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point is calculated by the path prediction algorithm, and an equilibrium constraint mathematical programming model is constructed using each path prediction cost. The scheduling path scheme is obtained by solving the model with the goal of minimizing the driving cost, and is used to optimize the scheduling of the target vehicle. The target vehicle is determined based on the vehicle status data.

[0049] Please refer to Figure 4 In some embodiments, the step of calculating the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point based on the traffic flow data and the regional feature information using a path prediction algorithm includes steps S401 to S406: Step S401: Determine the target vehicle based on the vehicle status data of available shuttle buses in the office building; In some embodiments, determining the target vehicle based on the vehicle status data of available office building shuttle buses includes: acquiring vehicle status data of all available office building shuttle buses, wherein the vehicle status data includes remaining driving range and rated passenger capacity; filtering candidate vehicle sets whose rated passenger capacity is greater than the total number of people in the target passenger group based on the rated passenger capacity of each available shuttle bus; and determining the vehicles in the candidate vehicle set whose remaining driving range is greater than a preset threshold as the target vehicles. Specifically, firstly, real-time vehicle status data of all available office building shuttle buses is acquired through the data interface of an Internet of Things (IoT) platform or a fleet management system, wherein the vehicle status data includes at least the remaining driving range reported by the battery management system (BMS) and the fixed parameter rated passenger capacity read from the vehicle file. Subsequently, the total number of people in the target passenger group is compared with the rated passenger capacity of each vehicle, and all vehicles whose rated passenger capacity is greater than the total number of people are included in the candidate vehicle set. Finally, the vehicles in the candidate vehicle set whose remaining driving range is greater than a preset threshold are finally determined as the target vehicles.

[0050] It should be noted that the preset threshold is not a fixed value, but is dynamically calculated based on the estimated round-trip distance from the office building to the target drop-off point, plus a safety margin (e.g., an additional 20% mileage) to ensure that the vehicle can successfully complete the trip and return safely, avoiding the risk of insufficient range during the journey.

[0051] Step S402: Based on traffic control rules, determine several candidate driving routes for the target vehicle from the office building to the target drop-off point; In some embodiments, after the target vehicle is determined, the map service API is invoked to request several candidate driving routes on the road network (usually 3-5 alternative routes with the best time or distance) starting from the office building location and ending at the target drop-off point. Each route is represented by an ordered sequence of road nodes.

[0052] It should be noted that when generating routes, the embedded traffic control rule database (such as no-entry, restricted-entry, and height-restriction information) automatically filters out routes that do not comply with the rules, ensuring that every candidate driving route obtained is legal and passable.

[0053] Step S403: Input the historical traffic flow data, the traffic flow data and the reported traffic incident information into the preset fusion model to predict the baseline travel time and the probability of congestion for each candidate travel route within the expected arrival time period. In some embodiments, for each candidate travel route, the system inputs historical traffic flow data from the same period, currently collected real-time traffic flow data (such as average vehicle speed), and reported traffic incident information (such as construction or accidents) into a pre-trained fusion model for prediction. This allows for a comprehensive analysis of temporal patterns and the impact of unforeseen events, and outputs the baseline travel time required for a vehicle to travel the route within the desired arrival time period. And the probability of significant congestion occurring on the path during that time period. .

[0054] It should be noted that the fusion model adopts a structure that combines a long short-term memory network (LSTM) with an attention mechanism. The training data consists of a large amount of historical traffic data and its corresponding actual travel time and congestion labels. During training, the model learns the complex relationship between the temporal evolution of traffic flow and the impact of events. For a given path and expected arrival time, it outputs the baseline travel time and the probability of congestion in parallel.

[0055] Step S404: Based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, the baseline travel time and the estimated energy consumption cost are corrected to obtain the target travel time and the target energy consumption cost. In some embodiments, weather information is first obtained, and the corresponding speed influence factor is looked up from a preset lookup table according to the weather type (such as rain, snow, fog) and intensity. (For example, during moderate rain) Meanwhile, based on the predicted congestion probability The time delay coefficient is converted into an actual time delay coefficient using the function f(P). And utilize this speed influence factor and time delay coefficient For the reference passage time The adjustments were made to obtain a target travel time that is closer to reality: .

[0056] In some embodiments, the correction of energy consumption cost is similar to the correction of the baseline travel time described above. Specifically, firstly, based on the real-time energy consumption model of the target vehicle, the distance L of the candidate path and the predicted average vehicle speed within the expected arrival time period provided by the path prediction algorithm are combined. The baseline energy cost was calculated. Then, based on the weather type (e.g., high temperature, low temperature) and intensity, an auxiliary energy consumption impact coefficient is obtained from a pre-set lookup table. Next, since congestion leads to frequent starts and stops and low-speed driving, significantly increasing energy consumption, the congestion probability output in the path prediction step is used. The energy consumption coefficient is converted into a congestion energy consumption coefficient by using an empirical function g(P) fitted based on historical driving data. Finally, the baseline energy cost will be... Weather-related energy consumption coefficient and congestion energy consumption coefficient —Conduct comprehensive calculations to obtain the final target energy consumption cost. The relevant formula is: .

[0057] It should be noted that the function f(P) was obtained through fitting historical data analysis, and the formula is as follows: , where a is the fitting parameter. The process of determining the empirical function g(P) is similar, so it will not be described in detail here.

[0058] It should be noted that the real-time energy consumption model was established through bench testing and road spectrum data calibration before the vehicle was put into operation. Its core is to reflect the energy consumption per unit distance of a specific vehicle type under a specific load. (e.g., kilowatt-hours per kilometer or liters per 100 kilometers) and vehicle speed The functional relationship is as follows. Therefore, the formula for calculating the baseline energy cost is: .

[0059] It should be noted that the energy consumption impact coefficient The value can be obtained by analyzing the correlation between historical energy consumption data and meteorological data, and this application does not impose any restrictions. For example, when the ambient temperature is higher than 28°C, the air conditioning system is turned on. A possible value is 1.08 to 1.15, which means that the baseline energy consumption will increase by 8% to 15%.

[0060] Step S405: Based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path, calculate the path risk coefficient of each path. In some embodiments, road attribute data (including the number of lanes, slope, curve radius, and presence of median strips) is obtained from a map service interface; real-time traffic condition information (such as warnings of slippery road surfaces and low visibility) is obtained from a traffic management platform; and historical accident data (including accident frequency and severity) for the road segment is obtained from an accident database. This multi-source heterogeneous data is then standardized and quantified to form a set of risk factor vectors. Subsequently, these risk factor vectors are input into a preset risk assessment model to calculate the path risk coefficient. The calculation formula is: ,in, Let m be the m-th risk characteristic factor (such as the sharp bend factor, accident rate factor). This is a preset risk factor.

[0061] It should be noted that the risk assessment model uses a weighted linear combination or a small neural network, and its weights or structure are determined during the training phase by fitting a large amount of labeled historical accident data with road feature data. This application does not impose any restrictions on this.

[0062] Step S406: Based on the target travel time, the target energy consumption cost, and the path risk coefficient, the path prediction cost of each candidate travel path is integrated to obtain the path prediction cost.

[0063] In some embodiments, the target travel time is determined using a preset risk weighting coefficient. Target energy consumption cost and path risk coefficient Weighting is applied to calculate the path prediction cost for each candidate driving path. The relevant formula is ,in, , , Risk weights pre-assigned.

[0064] In some embodiments, the step of constructing an equilibrium-constrained mathematical programming model using the predicted costs of each path and solving it with the objective of minimizing travel costs to obtain a scheduling route scheme includes: constructing an equilibrium-constrained mathematical programming model with the objective of minimizing travel costs based on each candidate travel path and its corresponding predicted cost, wherein the equilibrium-constrained mathematical programming model includes coverage constraints, resource constraints, and decision variable constraints; solving the equilibrium-constrained mathematical programming model, and determining the set of candidate travel paths corresponding to decision variables with a value of 1 in the solution results as the scheduling route scheme. Specifically, firstly, after determining each candidate travel path and its corresponding predicted cost (Cost), for the set L of all candidate paths, a decision variable is defined. when If the first option is selected, then path k is chosen in the final scheduling scheme; otherwise, it is not selected. The core of the model is to minimize the travel cost, and the objective function is constructed as follows: ,in, The path prediction cost for path k is given. Simultaneously, the model constructs three types of constraints: the first is a coverage constraint: for each target passenger group i, its corresponding subset of candidate paths... At least one of them is selected, that is The second constraint is resource constraint: the total number of selected routes (i.e., the number of shuttle buses that need to be dispatched) must not exceed the total number N of shuttle buses currently available in the office building. The third type is decision variable constraint: explicitly defining all decision variables as integers between 0 and 1, i.e. This model formalizes the scheduling problem as an explicit 0-1 integer linear programming problem. Then, a mathematical optimization solver (such as CPLEX, Gurobi, or an open-source solver) is invoked to solve the equilibrium-constrained mathematical programming model, seeking the set of decision variables that satisfies all constraints and minimizes the objective function value. After the solution is completed, the system extracts all... The decision variables, and the corresponding candidate driving paths, are selected into the final scheduling path scheme.

[0065] By using path prediction costs as the basic data, a balanced constrained mathematical programming model is constructed with the goal of minimizing the system's operating costs. This method can systematically weigh the needs of all passenger groups, the costs of all available paths, and the status of vehicle resources, generating a scheduling route plan from a globally optimal perspective. This allows the target vehicles to be scientifically directed to serve each passenger group in the most efficient order and route, ultimately maximizing the improvement of bus scheduling efficiency at the system level.

[0066] This invention provides a precise and original data foundation for subsequent intelligent demand aggregation and route planning by acquiring users' personalized travel needs. Through intelligent clustering of spatiotemporal demands, it can automatically and efficiently integrate scattered, individual travel requests into several target passenger groups with a common destination area and similar time windows. This achieves large-scale demand aggregation, transforming the problem of "serving multiple people" into "serving several groups," fundamentally creating conditions for designing efficient and intensive shuttle bus routes. By introducing and analyzing real-time monitoring images of the destination area, it can dynamically identify the actual regional characteristics of each candidate drop-off point, thereby intelligently selecting a route for each passenger group that comprehensively considers walking convenience and the suitability of the real-time drop-off environment. The optimal drop-off point not only avoids local delays or safety hazards caused by improper drop-off point selection, but also provides a more reasonable and feasible destination node for subsequent global route planning, ensuring the efficiency and reliability of the last mile in the rapid delivery process. By estimating the future cost of each alternative route through a route prediction algorithm, and using this as the basis, an equilibrium-constrained mathematical programming model with the goal of minimizing system operating costs is constructed for solution. This method can systematically weigh the needs of all passenger groups, the costs of all available routes, and the status of vehicle resources, generating a scheduling route plan from a globally optimal perspective. This scientifically directs target vehicles to serve each passenger group in the most efficient order and route, ultimately maximizing the improvement of bus scheduling efficiency at the system level.

[0067] like Figure 4 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided; An embodiment of the present invention provides a bus optimization scheduling system based on equilibrium constraint mathematical programming, comprising: The receiving module 100 is used to receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time. Processing module 200 is used to group users into several target passenger groups based on the destination area to which the destination location belongs and the expected arrival time period using a clustering algorithm. The determination module 300 is used to collect traffic flow data from the office building to each destination area within the expected arrival time period for each target passenger group, as well as real-time monitoring images of several candidate drop-off points in each destination area. The real-time monitoring images are analyzed using an image analysis algorithm to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, the target drop-off point corresponding to the target passenger group is determined. The scheduling module 400 is used to calculate the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point based on the traffic flow data and the regional feature information, and to construct an equilibrium constraint mathematical programming model using the path prediction costs, and solve the scheduling path scheme with the goal of minimizing the driving cost, for optimizing the scheduling of the target vehicle, wherein the target vehicle is determined according to vehicle status data.

[0068] In some embodiments, the determining module 300 includes: The first calculation unit is used to calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; The first correction unit is used to correct the estimated walking time based on real-time weather information to obtain the target walking time. The second calculation unit is used to calculate the regional risk coefficient of each of the candidate drop-off points using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. The third calculation unit is used to calculate the overall suitability of each of the candidate drop-off points by combining the target walking time and the regional risk coefficient. The first determining unit is used to select the candidate drop-off point with the highest overall suitability as the target drop-off point corresponding to the target passenger group.

[0069] In some embodiments, the scheduling module 400 includes: The second determining unit is used to determine the target vehicle based on the vehicle status data of available shuttle buses in the office building; The third determining unit is used to determine several candidate driving routes for the target vehicle from the office building to the target drop-off point in combination with traffic control rules; The prediction unit is used to input historical traffic flow data, the traffic flow data and the reported traffic incident information into a preset fusion model to predict the baseline travel time and the probability of congestion for each candidate travel route within the expected arrival time period. The second correction unit is used to correct the baseline travel time and the estimated energy consumption cost based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, so as to obtain the target travel time and the target energy consumption cost. The fourth calculation unit is used to calculate the path risk coefficient of each path based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path. The fifth calculation unit is used to integrate the target travel time, the target energy consumption cost, and the path risk coefficient to obtain the path prediction cost of each candidate travel path.

[0070] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the bus optimization scheduling method based on equilibrium constraint mathematical programming provided by any of the above-described method embodiments of the present invention.

[0071] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0072] Based on the above embodiments of the bus optimization scheduling method based on equilibrium constraint mathematical programming, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the bus optimization scheduling method based on equilibrium constraint mathematical programming of any embodiment of the present invention.

[0073] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0074] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0075] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0076] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the bus optimization scheduling method based on equilibrium constraint mathematical programming as described in any of the above-described method embodiments of the present invention.

[0077] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0078] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for optimal bus scheduling based on equilibrium-constrained mathematical programming, characterized in that, include: Receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time. Based on the destination area to which the destination location belongs and the expected arrival time period, users are grouped using a clustering algorithm to form several target passenger groups; For each target passenger group, traffic flow data from the office building to each destination area during the expected arrival time period is collected, as well as real-time monitoring images of several candidate drop-off points in each destination area. The real-time monitoring images are analyzed using image analysis algorithms to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, the target drop-off point corresponding to the target passenger group is determined. Based on the traffic flow data and the regional feature information, the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point is calculated by the path prediction algorithm. The equilibrium constraint mathematical programming model is constructed using the path prediction costs. The scheduling path scheme is obtained by solving the scheme with the goal of minimizing the driving cost. The target vehicle is determined based on the vehicle status data.

2. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 1, characterized in that, Based on the destination location's location within its designated area and the expected arrival time, users are grouped using a clustering algorithm to form several target passenger groups, including: The destination area to which each of the stated destination locations belongs is determined using a geographic information system; Users with the same destination region are clustered using a first clustering algorithm to divide the users into several initial passenger groups; The second clustering algorithm is used to cluster users in each initial passenger group according to their expected arrival time periods, so as to aggregate users whose expected arrival time periods overlap or have an interval of less than a preset threshold into a subgroup, thereby obtaining several target passenger groups.

3. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 1, characterized in that, The step of analyzing the real-time monitoring images using image analysis algorithms to identify the regional feature information of each candidate drop-off point includes: The real-time monitoring images are preprocessed to obtain preprocessed images; The preprocessed image is subjected to target detection by an image analysis algorithm to obtain several target detection boxes. Semantic segmentation is then performed on each target detection box to extract regional feature information of each candidate drop-off point. The regional feature information includes at least one of the following: pedestrian density, traffic sign status, distribution of stationary obstacles, and road surface condition.

4. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 3, characterized in that, The step of determining the target drop-off point corresponding to the target passenger group based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information includes: Calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; The estimated walking time is corrected based on real-time weather information to obtain the target walking time; The regional risk coefficient of each candidate drop-off point is calculated using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. The overall suitability of each candidate drop-off point is calculated by combining the target walking time and the regional risk coefficient. The candidate drop-off point with the highest overall suitability is selected as the target drop-off point for the target passenger group.

5. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 1, characterized in that, The step of calculating the path prediction cost for each candidate driving path of the target vehicle from the office building to the target drop-off point using a path prediction algorithm based on the traffic flow data and the regional feature information includes: The target vehicle is determined based on the vehicle status data of available shuttle buses in the office building; Based on traffic control rules, several candidate driving routes for the target vehicle from the office building to the target drop-off point are determined; Historical traffic flow data, the traffic flow data, and reported traffic incident information are input into a preset fusion model to predict the baseline travel time and congestion probability of each candidate travel route within the expected arrival time period. Based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, the baseline travel time and the estimated energy consumption cost are corrected to obtain the target travel time and the target energy consumption cost. Based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path, the path risk coefficient of each path is calculated. The path prediction cost for each candidate travel path is obtained by integrating the target travel time, the target energy consumption cost, and the path risk coefficient.

6. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 5, characterized in that, The process of determining the target vehicle based on the vehicle status data of available shuttle buses in the office building includes: Obtain vehicle status data for all available shuttle buses in the office building, wherein the vehicle status data includes remaining driving range and rated passenger capacity; Based on the rated passenger capacity of each available shuttle bus, a set of candidate vehicles whose rated passenger capacity is greater than the total number of people in the target passenger group is selected. Vehicles in the candidate vehicle set whose remaining driving range is greater than a preset threshold are identified as the target vehicles.

7. The bus optimization scheduling method based on equilibrium constraint mathematical programming according to claim 1, characterized in that, The step of constructing an equilibrium-constrained mathematical programming model using the predicted costs of each path, and solving it with the objective of minimizing travel costs to obtain a scheduling route scheme, includes: Based on each candidate driving path and the corresponding path prediction cost, an equilibrium constrained mathematical programming model is constructed with the goal of minimizing driving costs. The equilibrium constrained mathematical programming model includes coverage constraints, resource constraints, and decision variable constraints. The equilibrium constraint mathematical programming model is solved, and the set of candidate driving paths corresponding to the decision variables with a value of 1 in the solution results is determined as the scheduling path scheme.

8. A shuttle bus optimization scheduling system based on equilibrium constraint mathematical programming, characterized in that, include; The receiving module is used to receive travel request data submitted by several users in the office building, wherein the travel request data includes the destination location and the expected arrival time period; The processing module is used to group users into several target passenger groups based on the destination area to which the destination location belongs and the expected arrival time period, using a clustering algorithm. The determination module is used to collect traffic flow data from the office building to each destination area within the expected arrival time period for each target passenger group, as well as real-time monitoring images of several candidate drop-off points in each destination area. The real-time monitoring images are analyzed using an image analysis algorithm to identify the regional feature information of each candidate drop-off point. Based on the destination location of each passenger in each target passenger group, each candidate drop-off point, and the corresponding regional feature information, the target drop-off point corresponding to the target passenger group is determined. The scheduling module is used to calculate the path prediction cost of each candidate driving path of the target vehicle from the office building to the target drop-off point based on the traffic flow data and the regional feature information, and to construct an equilibrium constraint mathematical programming model using the path prediction costs, and solve the scheduling path scheme with the goal of minimizing the driving cost, for optimizing the scheduling of the target vehicle, wherein the target vehicle is determined according to vehicle status data.

9. The bus optimization scheduling system based on equilibrium constraint mathematical programming according to claim 8, characterized in that, The determining module includes: The first calculation unit is used to calculate the estimated walking time for each passenger in each of the target passenger groups from each of the candidate drop-off points to the corresponding destination location; The first correction unit is used to correct the estimated walking time based on real-time weather information to obtain the target walking time. The second calculation unit is used to calculate the regional risk coefficient of each of the candidate drop-off points using a preset safety scoring model, wherein the safety scoring model is constructed based on the pedestrian density, the status of traffic signs, the distribution of stationary obstacles, and the road surface conditions. The third calculation unit is used to calculate the overall suitability of each of the candidate drop-off points by combining the target walking time and the regional risk coefficient. The first determining unit is used to select the candidate drop-off point with the highest overall suitability as the target drop-off point corresponding to the target passenger group.

10. The bus optimization scheduling system based on equilibrium constraint mathematical programming according to claim 8, characterized in that, The scheduling module includes: The second determining unit is used to determine the target vehicle based on the vehicle status data of available shuttle buses in the office building; The third determining unit is used to determine several candidate driving routes for the target vehicle from the office building to the target drop-off point in combination with traffic control rules; The prediction unit is used to input historical traffic flow data, the traffic flow data and the reported traffic incident information into a preset fusion model to predict the baseline travel time and the probability of congestion for each candidate travel route within the expected arrival time period. The second correction unit is used to correct the baseline travel time and the estimated energy consumption cost based on weather information, the probability of congestion, and the real-time energy consumption model of the target vehicle, so as to obtain the target travel time and the target energy consumption cost. The fourth calculation unit is used to calculate the path risk coefficient of each path based on the acquired road attribute data, real-time traffic information and historical accident data of each candidate driving path. The fifth calculation unit is used to integrate the target travel time, the target energy consumption cost, and the path risk coefficient to obtain the path prediction cost of each candidate travel path.

Images