A demand response type urban and rural bus and town and village bus two-way joint scheduling method

Abstract

The application discloses a demand response type urban and rural public transport and town and village public transport two-way joint scheduling method, which comprises the following steps: dynamically determining the number and positions of transfer points of urban and rural DRT and town and village DRT based on orders; using a rolling time domain framework to divide the all-day uplink and downlink driving plans of the urban and rural DRT into a plurality of periods, each period containing one uplink and one downlink urban and rural DRT train, and a town and village DRT first cooperates with the uplink urban and rural DRT, and then cooperates with the downlink urban and rural DRT to complete all order tasks in the period; using a four-stage cooperative optimization algorithm based on site spatial clustering guidance to sequentially perform initial path solving and optimization, initial path repair, optimized order distribution and path updating, obtain the uplink and downlink scheduling final scheme in the current period and enter the next period, and the like. The application improves the rural area public transport operation efficiency and passenger travel experience, and takes into account the passenger demand and the interests of the public transport company.

Description

Technical Field

[0001] This invention relates to the field of public transportation scheduling technology, specifically a demand-responsive two-way joint scheduling method for urban and rural public transportation and town and village public transportation. Background Technology

[0002] Demand-Responsive Transit (DRT) systems can address travel issues in low-demand-density areas and have been widely adopted in many locations. Applying DRT in rural areas can provide convenient and efficient transportation services for rural residents.

[0003] Research combining multi-point transfers and bidirectional integration of multiple DRT lines in joint scheduling is still relatively rare, mainly facing the following challenges: ① The number of alternative transfer points is limited, making it difficult to effectively cover rural areas with large station spacing, long lines, and dispersed demand in time and space; ② How to adapt to the characteristics of bidirectional joint scheduling involving multiple directions, multiple stations, multiple train types, and multiple train services, including dividing the daily train schedule, determining the service order of town and village DRT and urban and rural DRT in both directions, and how to efficiently connect when switching between inbound and outbound tasks; ③ How to determine the task allocation based on actual orders and dynamically adjust the location and number of inbound and outbound transfer points.

[0004] In existing technologies, dynamic transfers typically employ fixed transfer schemes with few alternative transfer points (<3), generally only one. In suburban and rural areas, bus routes have large station spacing and long spans, and rural road networks are underdeveloped. Fixed transfer modes easily lead to route duplication between transfer lines, resulting in unnecessary detours for passengers, affecting system operational efficiency and passenger travel experience. Regarding station response rules, existing solutions use a combination of express and local buses, flexible station skipping, and A / B crossing of stations to reduce service stops, as well as time-based pricing strategies to guide flexible demand groups to travel during off-peak hours. However, due to limited public transportation options and resources in rural areas and relatively long departure times, while reducing service stops improves operational efficiency, it is difficult to balance passenger travel needs and bus company profits. Furthermore, fare discounts are not suitable during off-peak hours, while doubling DRT service fees during peak hours may exceed the economic affordability of some rural residents. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] Therefore, the purpose of this invention is to provide a demand-responsive two-way joint scheduling method for urban and rural public transport and town and village public transport, so as to solve the problems of limited alternative transfer points, difficult scheduling connection, and inflexible order allocation in the prior art, improve the operating efficiency of public transport in rural areas and the travel experience of passengers, and take into account both passenger needs and the interests of public transport companies.

[0007] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution:

[0008] A demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport includes the following steps:

[0009] S1. Determine the number and location of transfer points for urban and rural DRT and town and village DRT based on actual order data;

[0010] S2. Using a rolling time-domain framework, the daily up and down train schedule of the urban and rural DRT is divided into several cycles. Each cycle contains one up and one down urban and rural DRT train. One town and village DRT train first coordinates with the up urban and rural DRT, and then with the down urban and rural DRT to complete all order tasks within the cycle.

[0011] S3. A four-stage collaborative optimization algorithm based on site spatial clustering is adopted to sequentially solve and optimize the initial path, repair the initial path, optimize order allocation and update the path, obtain the final uplink and downlink scheduling scheme in the current cycle and enter the next cycle.

[0012] S4. Adopt a station response rule that coordinates multi-passenger response with peak fare. When a demand response station meets the needs of at least two UR or DD passengers getting on and off, or when a passenger is willing to pay the peak fare, the station will be included in the scheduling scope.

[0013] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, in step S1, the number and location of transfer points for urban-rural DRT and town-village DRT are dynamically determined based on actual orders as follows:

[0014] Demand response stations will be set up in densely populated areas on both sides of the benchmark urban and rural bus routes;

[0015] Based on spatial distribution characteristics, demand response sites are clustered to transform point demand into area demand and determine the fixed sites corresponding to each region.

[0016] The location and number of transfer points are dynamically adjusted based on the order allocation results.

[0017] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, step S2, the periodic division of the rolling time-domain frame further includes:

[0018] Determine the number of town and village buses. When the number of town and village DRTs is greater than 1, different town and village DRTs will provide services in a sequential order within each cycle.

[0019] Define the period range, including the previous period, the current period, and the next period, and ensure that the previous period, the current period, and the next period are not continuous.

[0020] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, step S3, the initial path solving and optimization stage of the four-stage collaborative optimization algorithm, includes:

[0021] Determine the DRT site set and service area set;

[0022] Determine the initial location of the town / village DRT and its surrounding area;

[0023] Orders are allocated based on regional priority, with the highest priority regional orders assigned to town and village DRTs and the rest assigned to urban and rural DRTs;

[0024] Verify the feasibility of the initial path.

[0025] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, step S3, the initial path repair stage of the four-stage collaborative optimization algorithm includes:

[0026] Expand the DRT service area in towns and villages and update the routes;

[0027] When the DRT service capacity in towns and villages reaches saturation, the areas to be added for service will be listed as pending areas;

[0028] Calculate the order rejection cost for each demand response site and discard the site with the lowest order rejection cost.

[0029] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, step S3, the four-stage collaborative optimization algorithm's optimized order allocation and route update stage includes:

[0030] An adaptive large neighborhood search algorithm is used to optimize order allocation;

[0031] Calculate the rejection cost for sites within the pending area, insert sites according to priority, and update the paths.

[0032] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint dispatching method described in this invention, step S4, the station response rules for multi-passenger response and peak fare coordination, further include:

[0033] If a passenger pays a peak-hour fare and the subsequent demand for more passengers increases, the fare difference will be refunded.

[0034] Passengers' acceptance of peak-hour fares is inversely proportional to the coefficient.

[0035] As a preferred embodiment of the demand-responsive urban and rural public transport and town and village public transport two-way joint dispatching method described in this invention, it also includes a differentiated fare strategy, in which the passenger fare consists of a base fare and a service fare. The service fare for UR passengers and DD passengers is determined by the service distance, while the service fare for UW passengers and DF passengers is zero.

[0036] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, the following constraints are also included:

[0037] Passenger constraints include: reservation volume not exceeding the maximum value in each period; UR or DD passengers' drop-off station being the demand response station; and passengers arriving at the boarding station earlier than the vehicle's arrival time.

[0038] Urban and rural DRT constraints include: one-way travel time not exceeding a multiple of the fixed route time, arrival time at a fixed station not exceeding a multiple of the fixed route time, time interval between fixed stations not exceeding a multiple of the benchmark departure interval, and remaining passenger capacity constraints.

[0039] Town and village DRT constraints include secondary transfer constraints, downlink reverse constraints, location reset constraints, and transfer time window constraints.

[0040] As a preferred embodiment of the demand-responsive urban-rural public transport and town-village public transport two-way joint scheduling method described in this invention, the rules for station spatial clustering include:

[0041] The areas must not be too large and must be independent of each other and not overlap.

[0042] The area is located on one side of the urban and rural DRT line, and is named sequentially by odd or even numbers with no repetition of names;

[0043] Area priority is determined by driving direction and route location, with areas on the same side of the same direction having higher priority than areas on opposite sides of the opposite direction.

[0044] Compared with existing technologies, the beneficial effects of this invention are as follows: Based on actual order conditions, this invention dynamically determines the number and location of transfer stations for urban-rural DRT and town-village DRT, improving transfer flexibility, ensuring optimal transfer plans, reducing passenger detour distances, and improving system operational efficiency. The station response rule of "multi-passenger response and peak fare coordination" improves operational efficiency while also taking into account the low payment capacity of some rural residents. Through a four-stage collaborative optimization algorithm guided by station spatial clustering, order allocation, transfer plans, and optimal routes can be optimized in a coordinated manner, thereby improving overall transportation efficiency. At the same time, by connecting upstream and downstream urban-rural DRT trains in a rolling cycle, and considering the empty running distance and location reset of town-village DRT during upstream and downstream task switching, efficient task connection within the cycle is achieved, further reducing the empty running rate and optimizing resource allocation. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0046] Figure 1 This is a schematic diagram of the dynamic transfer two-way joint scheduling organization provided by the present invention;

[0047] Figure 2 This is a flowchart of a demand-responsive two-way joint scheduling method for urban and rural public transport and town and village public transport according to the present invention;

[0048] Figure 3 The classification and naming results of one side of the urban and rural DRT reference line provided by this invention;

[0049] Figure 4 A schematic diagram of urban and rural DRT train services provided by the present invention;

[0050] Figure 5 This invention provides schematic diagrams of the enumeration and insertion methods.

[0051] Figure 6 This invention provides a schematic diagram of local and global paths.

[0052] Figure 7 The diagram below illustrates the internal destruction operator provided by the present invention, wherein (a) destroys the town / village DRT path and repairs the urban / rural DRT path; (b) destroys the urban / rural DRT path and repairs the town / village DRT path; and (c) simultaneously destroys and repairs both the town / village DRT and urban / rural DRT paths. Detailed Implementation

[0053] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0054] Urban and rural public transport adopts a segmented demand-response organizational model. In rural areas with low demand density, a semi-flexible, variable route system is used, while in urban areas with high demand density, fixed routes are maintained. Town and village public transport uses a more flexible demand-response connection system, connecting with urban and rural public transport at transfer points. Dispatch organization is as follows: Figure 1 As shown, the direction from rural to urban areas is defined as the upward direction. The solid line represents the baseline route of the urban-rural DRT, with demand response stations set at densely populated areas on both sides. AH represents the demand response stations for the current cycle, where A, B, C, and H represent upward demand, and D, E, F, and G represent downward demand. After the order allocation task is executed, AF is served by the town-village DRT, while G and H are served by the urban-rural DRT. The initial position of the town-village DRT is Z, and the optimal path after decision is Z→A→1→B→C→2→3→D→E→F (the initial position of the town-village DRT in the next cycle). Fixed stations 1, 2, and 3 are transfer points. Specifically, 1 and 2 are transfer points for passengers boarding at stations A, B, and C, respectively, and 3 is the transfer point for downward passengers. Simultaneously, 2 is both the endpoint of the upward task and the starting point of the downward task for the town-village DRT, and there may be empty running distances between 2 and 3 for both upward and downward connections. The solid arrows in the diagram represent the paths of the urban-rural DRT deviating from the baseline route's pick-up and drop-off stations H and G.

[0055] To ensure a good travel experience for passengers in the main passenger flow directions and to concentrate capacity to maximize service within the limited time of the current cycle, a differentiated time window strategy is adopted at transfer points: transfer points in the main passenger flow directions are set as hard time windows, while those in the secondary directions are set as soft time windows to retain scheduling flexibility. It is assumed that most passenger flow origin-destination (OD) points span multiple sections.

[0056] The rolling time-domain framework is used to divide the daily up-and-down train schedule of the urban-rural DRT into several cycles. Each cycle includes one up-and-down urban-rural DRT train trip, with one town-village DRT train coordinating with the up-and-down urban-rural DRT train first, and then with the down-and-down urban-rural DRT train. The three trains work together to complete all orders within one cycle.

[0057] The problem addressed in this invention is the optimal allocation of limited resources under a rolling cycle, specifically manifested in order allocation strategy, transfer plan determination, and optimal route planning. These three elements are significantly coupled: different order allocation strategies directly affect the transfer plan and optimal route results, and the quality of the transfer plan and optimal route results, in turn, affects the next order allocation decision. This characteristic makes the problem quite complex. Based on this, as... Figure 2As shown, a demand-responsive two-way joint scheduling method for urban and rural public transport and town and village public transport is designed, and the solution is obtained based on a four-stage collaborative optimization algorithm guided by station spatial clustering: the first, second, third and fourth stages are respectively initial path solving and optimization, initial path repair, optimized order allocation and path update, obtaining the final up and down scheduling scheme in the current cycle and entering the next cycle.

[0058] The model preparation section mainly includes:

[0059] (1) Determine the location of the demand response site

[0060] Demand response stations will be set up in densely populated areas on both sides of the benchmark urban and rural bus routes;

[0061] (2) Determine the demand response zone

[0062] Rural bus routes in rural areas and urban routes in high-demand-density urban areas will maintain fixed routes.

[0063] (3) Determine the service distance of the demand response site

[0064] Assume that the service distance of a demand response station is its distance from the nearest fixed station.

[0065] (4) Cluster demand response sites and determine the fixed sites corresponding to each region:

[0066] Compared to urban areas, rural areas have lower population density and more concentrated settlements, with a relatively smaller number of villages and more pronounced spatial distribution characteristics among them. Therefore, clustering demand response stations based on these spatial distribution characteristics transforms point demands into fewer area demands during order allocation, reducing the dimensionality of the solution space. To facilitate algorithm implementation, the following clustering and naming rules are set: ① Regions should not be too large; ② Regions are independent and cannot overlap; ③ Regions must be on one side of the urban / rural DRT line and cannot cross lines; ④ Regions on one side of the station should be named sequentially using odd or even numbers; ⑤ Region names cannot be duplicated.

[0067] Figure 3 The diagram illustrates the classification and naming results on one side of the urban and rural DRT benchmark lines. q Let q represent the set of fixed stations corresponding to region q. During scheduling, regional priority is determined by the direction of travel and the location of each region on both sides of the urban and rural public transport base line. Specifically, regions traveling in the same direction and on the same side of the line have higher priority than regions traveling in the opposite direction and regions on opposite sides. The set of all regions is Q; the set of all stops is P = P G ∪S, P G S and S represent the set of fixed sites and the set of demand response sites, respectively.

[0068] (5) Determine the number of town and village buses

[0069] The number of town and village DRT services (δ) participating in the service is determined by actual factors such as the departure interval of urban and rural DRT, the length of rural demand response sections, DRT demand, and the number of vehicles available locally. For example... Figure 4 As shown, if δ = 1, then one town / village DRT needs to participate in all tasks; if δ > 1, then different town / village DRTs will serve in a sequential manner within each cycle. To avoid ambiguity, the terms "previous cycle," "current cycle," and "next cycle" mentioned below specifically refer to the service cycle of one town / village DRT, which are τ-δ, τ, and τ+δ, respectively; when δ > 1, the above three cycles are not continuous.

[0070] (6) Determine the period range

[0071] That is, the periodic set T contained in the model.

[0072] Two-way scheduling process:

[0073] 1) Passenger Classification

[0074] A two-dimensional classification framework based on "travel direction-service mode" is used to distinguish passenger groups in the model. This framework divides passenger groups into four categories by analyzing the interaction between travel direction and service mode:

[0075] ①Upward-Reserved (UR): Passengers who make reservations before the departure of the urban and rural DRT;

[0076] ②Upward-Waiting (UW) refers to passengers traveling upwards and waiting at fixed stops;

[0077] ③ Downward-Demand (DD) type, which refers to passengers traveling downhill and arriving at a demand-response station;

[0078] ④ Downward-Fixed (DF) type, which refers to passengers traveling downhill and arriving at a fixed station.

[0079] 2) Uplink scheduling process

[0080] When the upbound rural-urban DRT departs, the system collects upbound reservation (UR) passenger orders from the previous cycle and the current cycle, and performs order allocation, transfer planning, and optimal route decision-making tasks. Specifically, if there is no demand-response station on the town / village DRT route, the train remains stationary. The detailed route-solving process is described in the solution algorithm.

[0081] Phase 1: Initial Path Solving and Optimization

[0082] Step 1: Determine the DRT site set S U and service area collection Q U

[0083] S U For UR passengers' drop-off points; Q U For S U The collection of regions where each station is located.

[0084] Step 2: Determine the initial location of the town / village DRT. and its surrounding area

[0085] For demand response sites where the previous cycle of town and village DRT service ended, for The area where it is located.

[0086] Step 3: Initial Order Allocation and Route Solving

[0087] Judgment set Q U Does it contain If a region is specified in the corresponding regional priority list, then orders from the highest priority region will be assigned to the town / village DRT, and the remaining orders will be assigned to the urban / rural DRT; otherwise, all orders will be assigned to the urban / rural DRT.

[0088] Step 4 Initial Path Feasibility Verification

[0089] Verify if the initial path meets the constraints. If yes, proceed to stage three; otherwise, proceed to stage two.

[0090] Phase Two: Initial Path Repair

[0091] The initial path was repaired by adjusting the service area and filtering orders. The specific steps are as follows:

[0092] Step 1: Expand the DRT service area in towns and villages

[0093] Referring to Step 3 of Phase 1, if there are new service areas in the town / village DRT, update the path and proceed to Step 2; otherwise, proceed to Step 4.

[0094] Step 2 Feasibility Verification

[0095] Determine if the new path satisfies the constraints. If it does, proceed to stage three; otherwise, proceed to Step 3.

[0096] Step 3: Dynamically adjust based on capacity grading

[0097] As shown in Table 1, when the town and village DRT does not meet the constraints, it indicates that its service capacity has become saturated and it does not have the capacity to serve a new area. The new service area is listed as a pending area. If the urban and rural DRT still does not meet the constraints at this time, it should abandon the order to meet its own constraints.

[0098] Table 1 - Hierarchical Dynamic Adjustment

[0099]

[0100] Step 4: Filter orders and reject them.

[0101] Calculate the rejection cost for each demand response station s in the current urban and rural DRT route. give up The smallest site. Including ticket price losses Penalty Cost Savings in driving costs

[0102]

[0103] In the formula: X csk It is a decision variable, determining whether train number k serves passenger c at station s; f c Here, C represents the fare for passenger c; C represents the set of all passenger orders; V and B represent the sets of town / village DRT services and urban / rural DRT services, respectively; c1 is the unit penalty cost; n s This indicates the number of passengers boarding (alighting) at station s (both directions); d ij c1 represents the distance between stations i and j; c2 and c3 represent the per-kilometer operating cost of town / village DRT and urban / rural DRT, respectively. It is a decision variable, used to determine whether station s is served by town / village bus k. v Service; in formula (4), train number k continuously serves stations i, s, and j.

[0104] Phase 3: Optimize order allocation, update paths, and assign orders to pending regions.

[0105] An adaptive large neighborhood search algorithm is used to optimize order allocation. Referring to Step 4 of Phase 2, the rejection cost of each station within the undetermined area is calculated. Stations are prioritized according to their rejection costs. Based on the priority order, the insertion method is used to determine whether each station should reject the order. If it is accepted, the path is updated.

[0106] Downlink scheduling process

[0107] After completing the upbound task, the town / village DRT will carry out the tasks of order allocation, transfer planning, and optimal route decision-making for downbound demand-response (DD) passengers. The specific steps are as follows:

[0108] Step 1: Determine the DRT site set S D and service area collection Q D

[0109] S D For DD passengers, this is the set of drop-off points; Q D For S D The collection of regions where each station is located.

[0110] Step 2: Determine the initial location of the town / village DRT.

[0111] This refers to the transfer point (fixed station) where the town / village DRT upbound task ends.

[0112] Step 3: Determine the optimal transfer point range and initial transfer point

[0113] Based on the site correspondence results obtained in 1.2.1, set S is determined. D Corresponding fixed site set The fixed stations with the smallest and largest serial numbers are... and The initial transfer point is

[0114] Step 4: Assign orders and solve for the optimal path.

[0115] In the same uplink scheduling process, "Phase 1 Step 3 to Phase 3" are obtained. This represents the scheduling results at transfer points. Specifically, when solving for the objective function value, the town / village DRT needs to be considered. and empty driving distance between

[0116] Phase 4: Determine the optimal transfer point for the outbound journey and proceed to the next cycle.

[0117] Proceed to Step 4, when The loop terminates at time, and the result is obtained. The scheduling results when each fixed station is the first downlink transfer point are used to select the solution with the optimal objective function value as the downlink scheduling scheme and proceed to the next cycle.

[0118] Peak operation scheduling strategy

[0119] During peak hours, the impact on system efficiency and the travel experience of both up-to-wait (UW) and down-to-fixed-station (DF) passengers should be minimized. Simultaneously, the travel needs of UR and DD passengers boarding and alighting at demand response stations must be considered. To this end, drawing inspiration from HOV lanes, a station response rule coordinating multi-passenger response and peak fare is proposed. Specifically, the system only includes a demand response station in its scheduling scope if it meets the multi-passenger response condition of at least two UR (DD) passengers boarding or alighting, or if a passenger is willing to pay an additional β times the service fare, i.e., the peak fare. After a passenger pays the peak fare, if the multi-passenger response condition is subsequently met, the fare difference is refunded, and the normal service fare is charged.

[0120] Assuming all UR and DD passengers are willing to accept normal service fares, and their acceptance of peak-hour fares... Primarily influenced by the coefficient β, the lower the β, the stronger the willingness to pay. Assuming an inverse relationship exists between the two, i.e.

[0121]

[0122] Differentiated pricing strategy

[0123] Differentiated fares are applied to different types of passengers, and a passenger fare f is defined. c From the base ticket price Service ticket price constitute. The fare for fixed-route buses. The additional cost for providing DRT service to UR and DD passengers is determined by the service distance s f Decision. UW passengers, DF passengers UR passengers, DD passengers for:

[0124]

[0125] Where: H c γ is the decision variable used to determine whether passengers actually paid extra for peak-hour fares; γ is the service fee coefficient, calculated as 1 kilometer if the distance is less than 1 kilometer.

[0126] The system's ticket revenue is

[0127]

[0128] In the formula: X c It is a decision variable used to determine whether passenger C receives service.

[0129] Optimization model:

[0130] Model assumptions

[0131] To simplify the model, the following assumptions are made: ① When the downlink urban-rural DRT reaches the section boundary point in the current cycle, the uplink urban-rural DRT in the next cycle begins; ② The average speed of the urban-rural DRT in rural areas is faster than that in urban areas; ③ The speed of the town-village DRT is equal to that of the urban-rural DRT in rural areas; ④ The buses travel at a constant speed between stops; ⑤ Each order corresponds to one passenger; ⑥ Except for passengers who are refused or actively give up their bus trips, the travel information of other passengers will not change midway; ⑦ The distance between stops is Euclidean distance; ⑧ Road conditions are good, and there are no external factors causing vehicle operation interruptions.

[0132] objective function

[0133] From the perspective of the bus company, with the objective function of maximizing system profit, the system profit consists of fixed costs Z1, operating costs Z2, order refusal penalty costs Z3, and fare revenue Z4. Based on the actual situation of rural public transport resource allocation, this paper includes vehicle idle costs within the category of fixed costs; that is, fixed costs exist regardless of whether the vehicle is in use.

[0134] Z = minZ4 - Z1 - Z2 - Z3 (9)

[0135]

[0136] Z3=c1n J (12)

[0137] Z4=∑f c ,X cJ =0 (13)

[0138] In the formula: c4 and c5 represent the fixed cost of each town / village DRT and urban / rural DRT vehicle operating one trip, respectively; Representing a periodic set The number of periods τ contained in X; ijk It is a decision variable, used to determine whether train number k serves stations i and j consecutively; n J Indicates the total number of orders rejected; X cJ It is a decision variable used to determine whether passenger C's order is rejected.

[0139] Constraints

[0140] (1) Passenger constraints

[0141] To meet and ensure the travel needs and fairness of UW and DF passengers, maximum reservation volume constraints and reservation station constraints are set for UR and DD passengers, and arrival time window constraints are added.

[0142]

[0143] Equation (14) represents the reservation amount Y in each period τ. τ Less than the maximum allowed reservation value Y max Equation (15) represents passenger type l c For UR or DD passenger drop-off stations Only demand-response stations are allowed; Equation (16) represents the time it takes for a passenger to arrive at the boarding station. Earlier than the arrival time of DRT vehicles X ck It is a decision variable used to determine whether passenger c is served by train k.

[0144] (2) Urban and rural DRT constraints

[0145] To avoid excessively long one-way travel times, ensure frequent and punctual departures in both directions, and minimize the impact on the total travel time and waiting time at stations for UW and DF passengers, travel time constraints, on-the-way time constraints, and fixed station interval constraints are set for urban and rural DRT systems, and passenger capacity constraints are added.

[0146]

[0147] Equation (17) represents train number k. b One-way travel time T k Not exceeding λ times the fixed line travel time T0; Equation (18) represents the time T required to reach each fixed station. ik Not exceeding α times the fixed line operating time. Indicates the fixed train number k. b The arrival time at each stop i is Indicates train number k b The departure time; Equation (19) indicates that the time interval between the arrival of urban and rural DRT vehicles at the same fixed station does not exceed ψ times the benchmark departure interval T0. Equation (20) indicates the departure time of train number k. b Remaining passenger capacity leaving current station j Remaining passenger capacity when leaving the previous station i Subtract the net number of passengers boarding at the current station j Indicates train number k b Set of stopover stations; Equation (21) represents the remaining passenger capacity of the first station. For the maximum passenger capacity R b Subtract the total number of reservations for train number k

[0148] (3) Town and Village DRT Constraints

[0149] To reduce or avoid unnecessary onboard time for UR and DD passengers and to prevent resource waste, secondary transfer constraints, downlink reverse constraints, and location reset constraints are set for town and village DRT systems, and transfer time window constraints are added, as shown in equations (22)-(27).

[0150] ①Two-way transfer constraints

[0151] Two-way transfers include cross-line transfers and upward reverse transfers. Cross-line transfers refer to situations where the town / village DRT crosses the urban / rural DRT base line, meaning when traveling from one side of the urban / rural DRT base line to the other, a stop is required at the nearest fixed station for UR and DD passengers to transfer. Upward reverse transfers refer to situations where, when traveling upwards, continuous service within the upward-bound area on the same side of the base line does not require a transfer, but if serving the downward-bound area on the same side, a stop is required at the nearest fixed station for passengers to disembark before proceeding.

[0152]

[0153] In the formula: It is a decision variable for determining the number of DRT trains traveling to towns and villages. Continuous service area q i and q j Whether the secondary transfer constraint is triggered.

[0154] ② Downward Reverse Constraint P j

[0155] During the downlink, in region q i Corresponding fixed site set In the middle, if there is a transfer point exist After the last fixed station in the downlink direction, the service to that area cannot be reversed.

[0156]

[0157] In the formula: It is a decision variable for determining the outbound DRT train services to towns and villages. Is it possible to serve the area q? i .

[0158] ③ Position Reset Constraints

[0159] Due to dynamic transfers and regional priorities, town and village DRT systems may cluster at both ends of the line, leading to resource waste. For example, if a town or village DRT completes its uplink connection at a section boundary but is not assigned downlink tasks, it will have to completely reverse its direction to perform uplink tasks in the next cycle, impacting system efficiency and passenger experience. Therefore, during periods of high passenger flow, after completing its current cycle's tasks, the town or village DRT should proceed to the midpoint p of the demand response section based on the spatial distribution of orders for the next cycle.Z The nearest and most needed demand response site, i.e.

[0160]

[0161] After the location reset, urban-rural DRT and town-village DRT are more evenly distributed within the rural demand response zones, and urban-rural DRT can start DRT service as soon as the train departs; town-village DRT focuses on serving the zone boundary and midpoint p. Z For sections of the DRT line, urban and rural DRT systems can maintain a fixed route as much as possible in these sections, thereby reducing the detour distance for passengers. During periods of low to medium passenger flow, DRT does not require rerouting due to the lower demand.

[0162] ④ Transfer time window constraints

[0163] When the main passenger flow is upward, the town / village DRT must arrive at each transfer point earlier than the urban / rural DRT, and maintain a flexible time window of εmin at the initial transfer point on the downward route; when the main passenger flow is downward, the opposite strategy is implemented. Simultaneously, the town / village DRT must arrive at the departure time of the urban / rural DRT on the upward route in the next cycle. Complete all tasks for the current cycle.

[0164]

[0165] In the formula: Indicates the transfer point for the upward journey; Indicates the first transfer point on the outbound route; H U (H D () is the set of transfer points for the up (down) route; It is a decision variable, determining the upstream (downstream) town / village DRT train services. Does it serve urban and rural DRT train services in both directions?

[0166] Solution Algorithm

[0167] Based on the characteristics of the model, a four-stage collaborative optimization algorithm guided by station spatial clustering is designed to solve the problem. This algorithm integrates enumeration and insertion methods, genetic algorithms, and adaptive large neighborhood search algorithms. The enumeration and insertion methods and genetic algorithms collaboratively optimize the transfer schemes and optimal paths within a single route, while the adaptive large neighborhood search algorithm focuses on optimizing the order allocation between town / village DRT and urban / rural DRT routes.

[0168] 2.1 Insertion Method and Enumeration Method

[0169] 2.1.1 Regional Allocation and Adjustment

[0170] Step 1: Determine basic information

[0171] gather For the newly added service area q j Corresponding fixed stations; collection For q j All sites within the system that are yet to be served.

[0172] Step 2: Construct the enumeration path

[0173] The enumerated path consists of three parts: starting point + intermediate point + ending point. For town / village DRT paths, proceed to Step 3; for urban / rural DRT paths, proceed to Step 4.

[0174] Step 3: Update town and village DRT routes

[0175] Origin: ① Initial Path: For uphill travel, the origin is the station where the previous cycle of the town / village DRT service ended; for downhill travel, the origin is the transfer point where the uphill task ended. ② New Cross-Line Area: The last station on the current town / village DRT path. ③ New Same-Side Area: The last demand response station on the current path, i.e., the second to last station. Destination: Any station in the network, i.e., a transfer point. Intermediate point: P j .

[0176] The shortest path is found by enumeration and then inserted at the end of the original path.

[0177] Step 4: Update urban and rural DRT routes

[0178] like Figure 5 As shown in (a), the starting point (end point): The first (last) fixed stops g1 and g2 along the direction of travel are for both directions, with the start and end points reversed for both directions. Midpoints: stops between g1 and g2. A set of.

[0179] The shortest path is found using an enumeration method, replacing the original paths g1, g2, and the stations in between. Specifically, if the number of intermediate points exceeds 8, a genetic algorithm is used to avoid a rapid expansion of the solution space.

[0180] 2.1.2 Point Allocation and Adjustment

[0181] like Figure 5 As shown in (b), when inserting a demand response station in a certain path, the nearest neighbor insertion method is used, and the steps are as follows:

[0182] Step 1: Find the stop closest to the station to be removed in the path to be inserted.

[0183] Step 2: Insert the removed stations to the front and rear of the stop station respectively, to obtain two new paths to be determined.

[0184] Step 3: Select the shortest path from the two pending new paths as the result after insertion.

[0185] Step 4: Determine whether the inserted path meets the constraints. If it does, update the path; otherwise, restore the path.

[0186] 2.2 Genetic Algorithm

[0187] While insertion and enumeration methods can achieve optimal solutions when generating local paths, repeated replacement and insertion operations in the initial path can lead to inconsistent connections between local paths, affecting the overall path's coherence and global optimality. Therefore, this paper introduces a genetic algorithm to further optimize the path. Fixed urban route sections do not require optimization.

[0188] 2.2.1 Local and Global Paths

[0189] Optimize local long arc paths, such as Figure 6 As shown. A genetic algorithm is used to optimize local and global paths. The global path is the entire path of the urban-rural DRT within the rural demand response zone. The steps for extracting the local path are as follows:

[0190] Step 1: Determine the long arc path

[0191] In segment 1, find the path g4 to g5 with the most stations between two fixed stations.

[0192] Step 2: Expand the long arc path into a local path

[0193] Extend g4 and g5 forward and backward by one fixed station to g3 and g6 respectively to obtain the local path to be optimized, g3~g6;

[0194] 2.2.2 Genetic Algorithm Design

[0195] A sequential coding rule is adopted, encoding the fixed stops along the urban and rural DRT fixed lines sequentially. The first and last fixed stops in the downlink direction (first and last stops) are coded as 1 and n, respectively, and demand response stops are coded sequentially starting from n+1. The objective function value is used as a measure of individual fitness. During the selection operation, an elite preservation strategy is combined with a roulette wheel selection strategy to perform a survival-of-the-fittest operation. The first and last stops of the path to be optimized remain unchanged, and the crossover and mutation range is between the second and second-to-last chromosomes of the population. A random two-point crossover and mutation rule is adopted, and the number of chromosomes involved in crossover and mutation is... N is the number of stations in the path to be optimized. The maximum number of iterations is 1000. If no better solution is found in 200 consecutive iterations, the iteration stops.

[0196] 2.3 Adaptive Large Neighborhood Search Algorithm

[0197] To avoid the limitation of simply adjusting order allocation between regions, and to reduce the dependence of different site spatial clustering methods on the system path planning results, an adaptive large neighborhood search algorithm with 4 destruction operators and 1 repair operator is designed to further optimize order allocation, so as to optimize order allocation by transferring sites without reducing the number of service sites.

[0198] 2.3.1 Destruction Operator

[0199] Demand response stations are removed from town / village or urban / rural DRT routes using disruption operators. Four disruption operators are designed: nearest neighbor area, least contiguous station, nearest neighbor station, and random disruption. Each disruption operator contains three internal disruption operators, such as... Figure 7 As shown, the methods are: (a) destroying the town / village DRT path and repairing the urban / rural DRT path; (b) destroying the urban / rural DRT path and repairing the town / village DRT path; (c) simultaneously destroying and repairing the town / village DRT and urban / rural DRT paths.

[0200] The destruction operator selects the transfer site from the DRT path. Taking internal operator 1 as an example, the destruction operators are explained as follows:

[0201] ①Nearest Neighbor Area: Within the area served by the Town / Village DRT, destroy the area that is closest to the Urban / Rural DRT service area and transfer the order tasks from that area to the Urban / Rural DRT. Proximity is determined based on regional priority.

[0202] ② Lowest Convenience Station: Among the stations served by town and village DRT systems, the tasks of the stations with the lowest convenience will be transferred to urban and rural DRT systems. The travel distance saved by not serving this demand response station indicates the degree of convenience.

[0203] ③Nearest station: Among the demand response stations of town and village DRT services, select the station that is spatially closest to the urban and rural DRT path, and transfer the task of that station to the urban and rural DRT.

[0204] ④ Random disruption: Randomly select a demand response site for town / village DRT services and transfer it to the task of urban / rural DRT.

[0205] 2.3.2 Repair Operator

[0206] Repair operator will Figure 7 The transfer stations are re-inserted into the town / village or urban / rural DRT path, as indicated by the dashed arrows in the diagram, thus achieving the transfer of order allocation status, using the nearest neighbor insertion principle. When the disruption operator is 1, it is a regional order allocation transfer, and the repair method and steps are the same as in 2.1.1; when the disruption operator is 2-4, it is a point order allocation transfer, and the repair method and steps are the same as in 2.1.2. No repair is required when a transfer station is removed from the path.

[0207] Specifically, if the internal disruption operator is 1, it is necessary to determine whether there are consecutive fixed stops in the town / village DRT route. If so, the task at the next fixed stop is deleted. If the internal disruption operator is 2 or 3, it is necessary to determine whether the relevant constraints of the town / village DRT are met. If they are met, it cannot be inserted at the very beginning of the route as the starting point. If it is inserted at the very end of the route, a transfer point needs to be added to construct a complete route.

[0208] 2.3.3 Adaptive Selection Strategy for Destruction Operators

[0209] The destructive operator is selected using the concept of roulette wheel selection, and the probability of each operator being selected is...

[0210]

[0211] In the formula: w ij The weight of the destruction operator i to which the internal operator j belongs, i.e., the destruction operator ij.

[0212] At the beginning of the iteration, each operator has the same initial weight. During the iteration, the operator weights are dynamically adjusted based on changes in performance scores. The weight update formula is:

[0213]

[0214] In the formula: ζ is the operator weight update coefficient; θ ij Let η be the cumulative score of operator ij within a given update period; ij θ represents the cumulative number of times the operator ij is called within a given update cycle. ij and η ij It is reset to 0 after each weight update.

[0215] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport, characterized in that, The steps are as follows: S1. Determine the number and location of transfer points for urban and rural DRT and town and village DRT based on actual order data; S2. Using a rolling time-domain framework, the daily up and down train schedule of the urban and rural DRT is divided into several cycles. Each cycle contains one up and one down urban and rural DRT train. One town and village DRT train first coordinates with the up urban and rural DRT, and then with the down urban and rural DRT to complete all order tasks within the cycle. S3. A four-stage collaborative optimization algorithm based on site spatial clustering is adopted to sequentially solve and optimize the initial path, repair the initial path, optimize order allocation and update the path, obtain the final uplink and downlink scheduling scheme in the current cycle and enter the next cycle. S4. Adopt the station response rule that coordinates multi-passenger response and peak fare. When a demand response station meets the needs of at least two up-reservation type or down-demand response type passengers getting on and off, or when a passenger is willing to pay the peak fare, the station will be included in the scheduling scope. In step S3, the initial path solving and optimization stage of the four-stage collaborative optimization algorithm includes: Determine the DRT site set and service area set; Determine the initial location of the town / village DRT and its surrounding area; Orders are allocated based on regional priority, with the highest priority regional orders assigned to town and village DRTs and the rest assigned to urban and rural DRTs; Verify the feasibility of the initial path; The initial path repair phase of the four-stage collaborative optimization algorithm includes: Expand the DRT service area in towns and villages and update the routes; When the DRT service capacity in towns and villages reaches saturation, the areas to be added for service will be listed as pending areas; Calculate the cost of order rejection at each demand response site and discard the site with the lowest order rejection cost; The four-stage collaborative optimization algorithm's order allocation and path update stage includes: An adaptive large neighborhood search algorithm is used to optimize order allocation; Calculate the rejection cost for sites within the pending area, insert sites according to priority and update the paths; In step S4, the station response rules for coordinating multi-passenger response and peak fare also include: If a passenger pays a peak-hour fare and the subsequent demand for more passengers increases, the fare difference will be refunded. Passengers' acceptance of peak-hour fares is inversely proportional to the coefficient.

2. The demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport according to claim 1, characterized in that, In step S1, the number and location of transfer points for urban-rural DRT and town-village DRT are dynamically determined based on the actual orders, as follows: Demand response stations will be set up in densely populated areas on both sides of the benchmark urban and rural bus routes; Based on spatial distribution characteristics, demand response sites are clustered to transform point demand into area demand and determine the fixed sites corresponding to each region. The location and number of transfer points are dynamically adjusted based on the order allocation results.

3. The demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport according to claim 1, characterized in that, In step S2, the periodic division of the rolling time-domain frame also includes: Determine the number of town and village buses. When the number of town and village DRTs is greater than 1, different town and village DRTs will provide services in a sequential order within each cycle. Define the period range, including the previous period, the current period, and the next period, and ensure that the previous period, the current period, and the next period are not continuous.

4. The demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport according to claim 1, characterized in that, It also includes a differentiated fare strategy, where passenger fares consist of a base fare and a service fare. The service fare for uphill (reservation-based) and downhill (demand-response) passengers is determined by the service distance, while the service fare for uphill (waiting) passengers and downhill (fixed-station) passengers is zero.

5. The demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport according to claim 1, characterized in that, It also includes the following constraints: Passenger constraints include: reservations in each period not exceeding the maximum value; passengers on the up-to-reservation or down-to-demand response routes must alight at the demand response station; and passengers must arrive at the boarding station before the vehicle arrives. Urban and rural DRT constraints: including one-way travel time not exceeding the fixed route time. The arrival time at the fixed station shall not exceed the fixed route time. The fixed station time interval shall not exceed the benchmark departure interval. Multiple, remaining passenger capacity constraints; Town and village DRT constraints include secondary transfer constraints, downlink reverse constraints, location reset constraints, and transfer time window constraints.

6. The demand-responsive two-way joint scheduling method for urban and rural public transport and town-village public transport according to claim 1, characterized in that, The rules for site spatial clustering include: The areas must not be too large and must be independent of each other and not overlap. The area is located on one side of the urban and rural DRT line, and is named sequentially by odd or even numbers with no repetition of names; Area priority is determined by driving direction and route location, with areas on the same side of the same direction having higher priority than areas on opposite sides of the opposite direction.

Images