A new demand response type public transportation system vehicle scheduling fast response method

Abstract

The application provides a novel demand response type public transportation system vehicle scheduling quick response method, which is a pre-allocation algorithm before vehicle assignment for the vehicle scheduling process of the novel demand response type public transportation system, accurately predicts the transport capacity in the corresponding time-space interval through a solving strategy of a time-space-vehicle-flow four-dimensional graph and a heuristic algorithm based on a moving time window, and pre-allocates vehicle resources through a quick response algorithm.The application can better instantaneously respond to continuous pre-booking travel requests in a short time, and the utilization of vehicle resources is relatively balanced, and there is sufficient empty vehicle scheduling time between two trips of the same vehicle.The order quick response result contains the response condition of the booking and the information of path planning, and the pre-allocation of vehicle resources integrates the orders with the same O-D and departure time, so that the workload of formal vehicle assignment and scheduling can be reduced, and the efficiency of the scheduling algorithm is improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology. Background Technology

[0002] New demand-responsive public transportation systems are an emerging type of urban mobility solution, including "micro-circulation" dynamic buses, intelligent shuttle buses, and smart train transportation systems. They are a public transportation mode without fixed routes or stops, which combines intelligent connected vehicles, big data, cloud computing and other technologies, and can provide point-to-point, high-capacity travel services that can be booked in advance or ordered on the spot.

[0003] The new demand-responsive public transportation system can meet the diverse travel needs of passengers. Passengers can book their trips in advance or place orders instantly, enjoying customized "point-to-point" travel using dedicated vehicles. It also adopts an intensive shared mobility model to improve capacity and efficiency. Its vehicle dispatching system can promptly process passenger requests, flexibly adjust vehicle routes, and assign different vehicle types to adapt to various travel scenarios. Compared to traditional buses, ride-hailing services, and other public transportation modes, this system combines the features of intensive travel, precise booking, and real-time dispatching.

[0004] Existing demand-response public transportation system scheduling algorithms, such as Figure 1 As shown, most of the studies focus on finding the optimal route when travel demand is known. However, there are no mature theoretical and practical results for accurately judging transport capacity in the future and responding to the dynamic demand of passengers with uncertainties. Summary of the Invention

[0005] The new demand-responsive public transport system adopts a reservation-based operation model, allowing passengers to book their trips for any future operating period at any time via a client application. Because passenger travel requests are unpredictable, and the scheduling process must comprehensively consider operating costs, it is impossible to allocate vehicle resources and plan routes in real time for each passenger request. Therefore, this invention uses rapid response as real-time feedback to passenger reservation requests to ensure service response quality, with vehicle assignment completed just before the actual travel time.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A novel rapid response method for vehicle dispatching in demand-responsive public transportation systems is characterized by a pre-allocation algorithm, designed for the vehicle dispatching process of new demand-responsive public transportation systems. This algorithm, acting as a pre-allocation algorithm before vehicle assignment, accurately predicts the transport capacity within the corresponding spatiotemporal interval through a solution strategy of a four-dimensional "time-space-vehicle-flow" graph and a heuristic algorithm based on a moving time window. By pre-allocating vehicle resources through a rapid response algorithm, the results can better adapt to the needs of new demand-responsive public transportation system scenarios.

[0008] A novel demand-responsive public transportation system vehicle dispatching rapid response method employs a solution strategy based on a "time-space-vehicle-flow" four-dimensional graph to calculate transport capacity within each spatiotemporal interval. The spatiotemporal four-dimensional graph includes four elements to consider when calculating transport capacity: time, space, vehicles, and load factor. In this graph, the horizontal axis represents the time series, evenly divided by the minimum time interval (t) at which passengers can book; the vertical axis represents the station sequence, divided proportionally by the distance between two adjacent stations (A, B, C, ..., where A, B, C are station codes). The station at the bottom and the station at the top of the vertical axis are the same station (A), forming a loop to achieve cyclical vehicle dispatch; diagonal lines represent vehicle resources, with the slope representing vehicle speed. Dashed diagonal lines indicate unused vehicle resources, while solid diagonal lines indicate occupied vehicle resources in that spatiotemporal period. The numbers next to the lines indicate the sequence number of the occupied vehicle; the color of the diagonal lines indicates the vehicle load factor, with darker colors indicating a higher load factor, and a load factor of 1 being dark red.

[0009] During the rapid response phase, the horizontal and vertical axes and dashed diagonal lines of the "time-space-vehicle-flow" four-dimensional graph need to be initialized according to the actual route conditions. When a passenger's travel request containing information such as origin and destination stations, travel time, and number of passengers is received, the usage of vehicle resources in the corresponding time and station interval in the four-dimensional graph can be queried. Based on the vehicles and the occupancy rate, it is decided whether to accept the passenger's reservation request. After successfully accepting the passenger's reservation request, vehicle resources are pre-allocated for them. If an empty vehicle is allocated, a solid diagonal line is added in the corresponding time and space interval to indicate vehicle resource occupancy, and the color of the diagonal line is increased according to the number of passengers. If the vehicle is shared with other reservations, the color of the corresponding diagonal line is updated to indicate the increase in vehicle occupancy rate.

[0010] A heuristic algorithm based on a moving time window is used to solve the problem:

[0011] STEP1: Initialize the vehicle resource pool, and deploy a total of n vehicles to the resource pool, numbered 1, 2, ... n respectively. Correspondingly, release vehicle resources in the "time-space-vehicle-flow" four-dimensional graph.

[0012] STEP2: Receive real-time booking requests from passengers; trip information includes origin and departure stations. Terminal Departure time of the trip Total number of participants ,in Not exceeding the vehicle's maximum passenger capacity ;

[0013] STEP 3: Based on the reservation information , Perform path planning and predict the future. Always set off, from arrive The trip duration is used to calculate the end time of the trip. At the same time, the corresponding spacetime segments are marked in the four-dimensional diagram and represented by red dashed lines;

[0014] STEP4: Select trips from the historical itinerary pool with departure times less than [time missing]. Plus maximum redundancy time or the end time is greater than minus The trips that have been successfully booked and pre-assigned vehicle numbers are sorted by vehicle number from largest to smallest and departure time from earliest to latest to form a vehicle candidate set. At the same time, all the oblique solid lines in the four-dimensional map whose starting and ending abscissas are within the above range are marked according to the size of the vehicle number and the order of the beginning and ending abscissas to form a vehicle resource set.

[0015] STEP 5: Perform different actions depending on the results of the successfully booked itinerary set:

[0016] a. If there is a historical itinerary with the same departure time, starting station, and ending station as the new itinerary, and the remaining number of seats can meet the number of passengers booked for the new itinerary, then the vehicle resources of that train sequence can be pre-allocated to the new itinerary, and the color of the corresponding solid line in the four-dimensional graph can be updated.

[0017] b. If there is a historical itinerary with the same departure time, starting station, and ending station as the new itinerary, but the remaining number of seats is insufficient to meet the number of passengers booking the new itinerary, the vehicle resources for this train sequence cannot be pre-allocated. Continue to determine whether the following conditions are met.

[0018] c. Sort by the sequence number of the successfully booked trips, and check in turn whether the successfully booked orders for the current trip are empty. The vehicle resources of the first empty trip will be pre-allocated to the new trip. Change the dotted line of the specified time and space interval to a diagonal solid line, and update the color of the diagonal line according to the number of people in the order.

[0019] d. If there are no empty vehicles, sort the vehicles in the candidate set by their sequence number, check in turn whether there are remaining seats for each vehicle sequence in the current time and space, calculate the time utility, and then pre-allocate the optimal vehicle sequence resources for the new trip based on the utility. Time utility is represented by the size of the time interval between available seats, that is, the departure time of the subsequent trip minus the end time of the preceding trip, which is the difference between the horizontal coordinate interval of the beginning and end of the solid line and the beginning and end of the colored dashed line of the same vehicle sequence marked in the four-dimensional graph; assign the vehicle sequence with the lowest time utility, change the red dashed line to a solid line, and update the line color synchronously according to the number of orders.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] This invention can effectively and instantly respond to consecutive booking requests within a short period of time, and the utilization of vehicle resources is relatively balanced, with sufficient empty vehicle scheduling time between two consecutive trips for the same vehicle. The order rapid response result includes information on the response status of the booking and route planning. At the same time, the pre-allocation of vehicle resources integrates orders with the same origin-destination (OD) and departure time, which can reduce the workload of subsequent formal vehicle assignment and scheduling, and improve the efficiency of the scheduling algorithm. Attached Figure Description

[0022] Figure 1 This is a novel approach to solving the vehicle scheduling problem in demand-responsive public transportation systems, which is currently the only technology available.

[0023] Figure 2 This is a schematic diagram of the "time-space-vehicle-flow" four-dimensional diagram of the present invention.

[0024] Figure 3 This is a schematic diagram of the algorithm flow based on the moving time window of the present invention.

[0025] Figure 4 The diagram in 'af' is a schematic diagram of the algorithm flow of this invention corresponding to the four-dimensional "time-space-vehicle-flow" diagram.

[0026] Figure 5 This is a schematic diagram of the line stations in an embodiment.

[0027] Figure 6 This is a four-dimensional diagram of the scheduling response results of the embodiment, which is a "time-space-vehicle-flow" diagram.

[0028] Figure 7 This is the reservation mini-program interface of the embodiment.

[0029] Figure 8 These are images of intelligent connected vehicles used in the embodiments. Detailed Implementation

[0030] Existing demand-responsive public transport systems employ a reservation-based operating model, allowing passengers to book trips for any future operating time slot via a client application. Due to the uncertainty of passenger travel requests, and the need to comprehensively consider operating costs during the scheduling process, it's impossible to allocate vehicle resources and plan routes in real-time for each passenger request. Therefore, a rapid response mechanism is used to provide real-time feedback on passenger reservation requests to ensure service quality. Vehicle assignment is then completed just before the actual travel time. Figure 1 As shown.

[0031] The new demand-responsive public transport system's vehicle dispatching is divided into two phases: rapid response and vehicle assignment. The rapid response phase involves assessing future vehicle capacity and responding to passenger reservation requests in real time; the vehicle assignment phase involves centralized dispatching of passenger reservation requests and formally assigning vehicles. This invention is applied in the rapid response phase.

[0032] A novel rapid response method for vehicle dispatching in demand-responsive public transportation systems employs a solution strategy based on a four-dimensional "time-space-vehicle-flow" graph to calculate transport capacity within each spatiotemporal interval. This spatiotemporal four-dimensional graph incorporates four elements—time, space, vehicles, and occupancy rate—that need to be considered in capacity calculation. Figure 2 As shown: In the four-dimensional graph, the horizontal axis represents the time series, which is evenly divided according to the minimum time interval (t) at which passengers can make reservations; the vertical axis represents the station sequence, which is divided proportionally according to the distance between two adjacent stations (A, B, C, ..., where A, B, C are station codes). The station at the bottom and the station at the top of the vertical axis are the same station (A), forming a loop to realize the cyclical assignment of vehicles; the diagonal lines represent vehicle resources, and their slope is the vehicle speed. Dashed diagonal lines indicate that the vehicle resources are not used, and solid diagonal lines indicate that the vehicle resources are occupied in that time and space. The numbers next to the lines indicate the sequence number of the vehicle occupying the resources; the color of the diagonal lines indicates the vehicle load factor. The darker the color, the higher the load factor. When the load factor is 1, it is dark red.

[0033] During the rapid response phase, the horizontal and vertical axes and dashed diagonal lines of the "time-space-vehicle-flow" four-dimensional graph need to be initialized according to the actual route conditions. When a passenger's travel request containing information such as origin and destination stations, travel time, and number of passengers is received, the usage of vehicle resources in the corresponding time and station interval in the four-dimensional graph can be queried. Based on the vehicles and the occupancy rate, it is decided whether to accept the passenger's reservation request. After successfully accepting the passenger's reservation request, vehicle resources are pre-allocated for them. If an empty vehicle is allocated, a solid diagonal line is added in the corresponding time and space interval to indicate vehicle resource occupancy, and the color of the diagonal line is increased according to the number of passengers. If the vehicle is shared with other reservations, the color of the corresponding diagonal line is updated to indicate the increase in vehicle occupancy rate.

[0034] A heuristic algorithm based on a moving time window was used to solve the problem (e.g., Figure 3 (As shown).

[0035] The algorithm idea is as follows:

[0036] STEP 1: Initialize the vehicle resource pool, and deploy a total of n vehicles to the pool, numbered 1, 2, ..., n. Correspondingly, in the "Time-Space-Vehicle-Flow" four-dimensional graph, vehicle resources are released. All diagonal lines are represented by dashed lines (e.g., ...). Figure 4 (as shown in a)

[0037] STEP2: Receive real-time booking requests from passengers; trip information includes origin and departure stations. Terminal Departure time of the trip Total number of participants ,in Not exceeding the vehicle's maximum passenger capacity ;

[0038] STEP 3: Based on the reservation information , Perform path planning and predict the future. Always set off, from arrive The trip duration is used to calculate the end time of the trip. At the same time, the corresponding spatiotemporal segments are marked in the four-dimensional diagram, represented by red dashed lines (e.g., Figure 4 (as shown in b)

[0039] STEP4: Select trips from the historical itinerary pool with departure times less than [time missing]. Plus maximum redundancy time or the end time is greater than minus The trips that have been successfully booked and pre-assigned vehicle numbers are sorted by vehicle number from largest to smallest and departure time from earliest to latest to form a vehicle candidate set. Simultaneously, the trips are marked on a four-dimensional graph (e.g., based on vehicle number size and the order of the origin's x-axis). Figure 4 The vehicle resource set is composed of all the oblique solid lines whose starting and ending abscissas fall within the above range (as shown in c).

[0040] STEP 5: Perform different actions depending on the results of the successfully booked itinerary set:

[0041] a. If a historical itinerary with the same departure time, starting station, and ending station as the new itinerary exists, and the remaining seats are sufficient to accommodate the number of passengers booking for the new itinerary, then the vehicle resources of that itinerary can be pre-allocated to the new itinerary, and the color of the corresponding solid line in the four-dimensional graph can be updated (e.g., ...). Figure 4 (as shown in d)

[0042] b. If there is a historical itinerary with the same departure time, starting station, and ending station as the new itinerary, but the remaining number of seats is insufficient to meet the number of passengers booking the new itinerary, the vehicle resources for this train sequence cannot be pre-allocated. Continue to determine whether the following conditions are met.

[0043] c. Sort by the sequence number of successfully booked trips, and sequentially check if the successfully booked order for that trip is empty in the current time and space. The vehicle resource of the first empty trip will be pre-allocated to a new trip. Change the dotted line of the specified time and space interval to a solid diagonal line, and update the color of the diagonal line according to the number of orders (e.g., ...). Figure 4 (as shown in e).

[0044] d. If there are no empty vehicles, sort the vehicles in the candidate set by their sequence number, and sequentially determine whether there are remaining seats for each vehicle sequence in the current time-space context. Calculate the time utility, and then pre-allocate the optimal vehicle sequence resources for the new trip based on the utility. Time utility is represented by the time interval between available seats, i.e., the departure time of the subsequent trip minus the end time of the preceding trip, which is the difference in the horizontal coordinate interval between the start and end points of the solid line and the colored dashed line for the same vehicle sequence marked in the four-dimensional graph (e.g., ...). Figure 4 (As shown in f). Assign the vehicle with the least time utility, and change the red dashed line to a diagonal solid line. The color of the diagonal line will be updated synchronously according to the number of orders.

[0045] Example 1

[0046] Using the actual road conditions of Huyi Highway in Jiading District, Shanghai as a case study application scenario, this new demand-responsive public transportation system case line is 8.2km long, with 12 stations in each direction, totaling 24 stations in both directions (e.g., Figure 5 (As shown in the image) The operation consists of 5 vehicles, each accommodating up to 4 passengers. Through central control, intelligent connected vehicles, and a green wave control system, the operating speed is maintained at 60 km / h. This case study route implements a reservation-based, streamlined, and direct-access operation mode, allowing passengers to book their trips in 5-minute increments. It possesses all the key characteristics of a novel demand-responsive public transportation system and can be used to verify the usability and effectiveness of rapid response algorithms.

[0047] Under the above vehicle and route conditions, a certain number of passenger travel orders were simulated and sent out in sequence. The specific information and order of the orders are shown in Table 1. For each order, the minimum selectable departure time granularity is 5 minutes, and the maximum number of passengers that can be booked is 4. To better verify the effectiveness of the proposed scheduling algorithm, the departure times of the simulated orders were concentrated between 10:00 and 16:00. For example, the booking request with order number 1 indicates that the passenger's departure time is 10:00, the starting station is Zhaoxian Road A Station, the destination station is Yagang Road A Station, and the number of passengers booked is 2.

[0048] Table 1. Order Information Sequence Table for the Case Study

[0049]

[0050] For each trip, route planning is performed according to the order of booking to obtain the corresponding trip end time. Then, the fast response algorithm is used to calculate each order in turn. The results of the fast response and the pre-assigned vehicle sequence number are shown in Table 2. The * in parentheses indicates another order number that shares a vehicle.

[0051] Table 2 Fast Response Processing Results

[0052]

[0053] The vehicle resource usage of the above scheduling response results in the "time-space-vehicle-flow" four-dimensional graph is as follows: Figure 6 As shown.

[0054] The preliminary vehicle assignment scheme obtained through the fast response algorithm is shown in Table 3. Since the fast response algorithm responds instantly according to the order reservation order, the vehicle assignment scheme may not be able to make the best use of vehicle resources. A more optimized vehicle scheduling scheme will be determined in the specific vehicle assignment stage before the passengers actually depart.

[0055] Table 3 Vehicle Assignment Scheme

[0056]

Claims

1. A novel rapid response method for vehicle dispatching in a demand-responsive public transportation system, characterized in that, For the vehicle scheduling process of new demand-responsive public transportation systems, as a pre-allocation algorithm before vehicle assignment, it uses a solution strategy of "time-space-vehicle-flow" four-dimensional graph and a heuristic algorithm based on moving time windows to accurately predict the transport capacity in the corresponding time and space interval. A solution strategy based on a "time-space-vehicle-flow" four-dimensional graph is adopted to calculate the transport capacity in each spatiotemporal interval. The spatiotemporal four-dimensional graph includes four factors that need to be considered when calculating transport capacity: time, space, vehicles, and load factor. In the four-dimensional graph, the horizontal axis represents the time series, which is evenly divided according to the minimum time interval t that passengers can book; the vertical axis represents the station sequence, which is divided proportionally according to the distance between two adjacent stations. The station at the bottom and the station at the top of the vertical axis are the same station A, forming a loop to realize the cyclical allocation of vehicles; the diagonal lines represent vehicle resources, and their slope is the vehicle speed. Dashed diagonal lines indicate that the vehicle resources are not used, and solid diagonal lines indicate that the vehicle resources in that spatiotemporal space are occupied. The numbers next to the lines indicate the sequence number of the vehicles occupying the resources; the color of the diagonal lines indicates the vehicle load factor, and the darker the color, the higher the load factor. A load factor of 1 is dark red. A heuristic algorithm based on a moving time window is used to solve the problem: STEP1: Initialize the vehicle resource pool, and deploy a total of n vehicles to the resource pool, numbered 1, 2, ... n respectively. Correspondingly, release the vehicle resources in the "time-space-vehicle-flow" four-dimensional graph. STEP2: Receive real-time booking requests from passengers; trip information includes origin and departure stations. Terminal Departure time of the trip Total number of participants ,in Not exceeding the vehicle's maximum passenger capacity ; STEP 3: Based on the reservation information , Perform path planning and predict the future. Always set off, from arrive The trip duration is used to calculate the end time of the trip. At the same time, the corresponding spacetime segments are marked in the four-dimensional diagram and represented by red dashed lines; STEP4: Select trips from the historical itinerary pool with departure times less than [time missing]. Plus maximum redundancy time or the end time is greater than minus The trips that have been successfully booked and pre-assigned vehicle numbers are sorted by vehicle number from largest to smallest and departure time from earliest to latest to form a vehicle candidate set. At the same time, all the oblique solid lines in the four-dimensional map whose starting and ending abscissas are within the above range are marked according to the size of the vehicle number and the order of the beginning and ending abscissas to form a vehicle resource set. STEP5: Perform different operations depending on the results of the successfully booked itinerary set.

2. The novel demand-responsive public transportation system vehicle dispatching rapid response method as described in claim 1, characterized in that, During the rapid response phase, the horizontal and vertical axes and dashed diagonal lines of the "time-space-vehicle-flow" four-dimensional graph are initialized according to the actual route conditions to represent vehicle resources. When a passenger's travel request containing information such as origin and destination stations, travel time, and number of passengers is received, the usage of vehicle resources in the corresponding time and station interval in the four-dimensional graph is queried. Based on the vehicles and the occupancy rate, it is decided whether to accept the passenger's reservation request. After successfully accepting the passenger's reservation request, vehicle resources are pre-allocated for them. If an empty vehicle is allocated, a solid diagonal line is added in the corresponding time and space interval to indicate vehicle resource occupancy, and the diagonal line color is increased according to the number of passengers. If the vehicle is shared with other reservations, the color of the corresponding diagonal line is updated to indicate the increase in vehicle occupancy rate.

3. The novel demand-responsive public transportation system vehicle dispatching rapid response method as described in claim 2, characterized in that, In STEP5, a. If there is a historical itinerary with the same departure time, starting station, and ending station as the new itinerary, and the remaining number of seats can meet the number of passengers booked for the new itinerary, then the vehicle resources of that train sequence are pre-allocated to the new itinerary, and the color of the corresponding solid line in the four-dimensional graph is updated. b. If there is a historical itinerary with the same departure time, starting station, and ending station as the new itinerary, but the remaining number of seats is insufficient to meet the number of passengers booking the new itinerary, the vehicle resources for this train sequence cannot be pre-allocated. Continue to determine whether the following conditions are met. c. Sort by the sequence number of the successfully booked trips, and check in turn whether the successfully booked orders for the current trip are empty. The vehicle resources of the first empty trip will be pre-allocated to the new trip. Change the dotted line of the specified time and space interval to a diagonal solid line, and update the color of the diagonal line according to the number of people in the order. d. If there are no empty vehicles, sort the vehicles in the candidate vehicle set by their sequence number, check whether there are any remaining seats for each vehicle sequence in the current time and space, calculate the time utility, and then pre-allocate the optimal vehicle sequence resources for the new trip based on the utility. The time utility is represented by the size of the time interval between empty seats, that is, the departure time of the subsequent trip minus the end time of the preceding trip, which is the difference between the horizontal coordinate interval of the beginning and end of the solid line and the beginning and end of the colored dashed line of the same vehicle sequence marked in the four-dimensional graph. Assign the vehicle sequence with the lowest time utility, change the red dashed line to a solid line, and update the color of the line synchronously according to the number of orders.

Images