Intelligent Dispatch Method and System for Hub Taxi Based on Dynamic Partitioning

Abstract

This invention belongs to the field of intelligent transportation and low-carbon technology, specifically disclosing a hub taxi intelligent dispatching method and system based on dynamic zoning. The method includes: classifying passengers into three levels: short-distance, medium-distance, and long-distance; performing status detection and coding management on taxis entering the hub area to construct a vehicle attribute set; dividing the lanes in the hub waiting area into three types: dedicated short-distance lanes, dedicated long-distance lanes, and flexible adjustment lanes, and periodically adjusting the number of each type of lane based on real-time collected passenger type distribution data and taxi queuing data; and achieving optimal matching between passengers and taxis through a multi-factor weighted matching algorithm based on passenger level, identification number, and vehicle attribute set, and planning the optimal driving route for the order. This invention solves the problem that existing technologies lack a systematic solution for demand perception, dynamic adjustment, precise matching, and effect feedback, and cannot simultaneously meet the needs of drivers, passengers, and the environment.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation and low-carbon technology, specifically relating to a hub taxi intelligent dispatching method and system based on dynamic zoning. Background Technology

[0002] With the acceleration of urbanization, passenger traffic at large transportation hubs continues to grow, with daily taxi demand reaching 12,000-18,000 trips, and drivers experiencing average waiting times of 60-120 minutes. The current dispatching model faces three core problems:

[0003] 1. Road rage among drivers: If drivers receive short-distance orders after waiting in long queues, the input-output ratio becomes unbalanced, and the incidence of road rage is as high as 38%. This manifests as poor service attitude, sudden acceleration and deceleration, etc., which directly threaten passenger safety.

[0004] 2. Inefficient resource allocation: The traditional "first-come, first-served" model lacks differentiated management, resulting in long-distance passengers waiting for more than 40 minutes, short-distance order rejection rates reaching 25%, and lane resource utilization rates of less than 60%.

[0005] 3. Lack of low-carbon operation: Vehicle idling time accounts for more than 35% of the total operating time, the empty running rate is 25%, and the average daily additional carbon emissions per vehicle reach 12kg, which is contrary to the "dual carbon" target.

[0006] Existing technologies only solve local problems and lack a systematic solution of "demand perception, dynamic adjustment, precise matching, and effect feedback," which cannot simultaneously meet the needs of drivers, passengers, and the environment. Therefore, it is urgent to build a multi-objective collaborative intelligent scheduling system. Summary of the Invention

[0007] The purpose of this invention is to address the problem that existing technologies lack a systematic solution of "demand perception-dynamic adjustment-precise matching-effect feedback" and cannot simultaneously meet the needs of drivers, passengers, and the environment. This invention proposes a hub taxi intelligent dispatching method and system based on dynamic partitioning.

[0008] The technical solution of the present invention is as follows: Firstly, a hub taxi intelligent dispatching method based on dynamic partitioning, comprising the following steps:

[0009] Passengers entering the hub are divided into three categories—short-haul, medium-haul, and long-haul—based on their destination distance, and a unique identification number is generated for each passenger.

[0010] Perform status detection and coding management on taxis entering the hub area, and construct a set of vehicle attributes;

[0011] The waiting area lanes in the hub are dynamically divided into three types: short-distance dedicated lanes, long-distance dedicated lanes, and flexible adjustment lanes. Based on real-time collected passenger type distribution data and taxi queuing data, the number of each type of lane is periodically adjusted.

[0012] Based on passenger class, identification number and vehicle attribute set, the optimal matching of passengers and taxis is achieved through a multi-factor weighted matching algorithm, and dispatch instructions are generated to guide the vehicle into the lane and generate an order.

[0013] Based on travel time and carbon emissions, the system plans the optimal travel route for orders and completes intelligent dispatching of taxis at hubs.

[0014] As a preferred option, the vehicle attribute set includes:

[0015] Basic coding: company code, vehicle ID, and service preferences; among which, service preferences include short-distance priority, medium-distance priority, long-distance priority, and flexible service;

[0016] Dynamic attributes: current location, cumulative waiting time, remaining battery or fuel, and historical service records.

[0017] Preferably, the lane number adjustment cycle is 15 minutes; the lane number adjustment is based on factors including queue length of each lane, passenger type distribution, taxi waiting time, and real-time traffic flow; the lane number adjustment decision is generated through a reinforcement learning model, and the reward function of the reinforcement learning model... for:

[0018]

[0019] in, , and This represents the weighting coefficient.

[0020] As a preferred option, the multi-factor weighted matching algorithm employs a two-layer screening mechanism, specifically:

[0021] First-level filtering: Compatibility filtering based on passenger level and taxi service preference identifiers;

[0022] Second-level screening: For the candidate taxi set that passed the first-level screening, calculate its matching score with passengers. The calculation formula is:

[0023]

[0024] in, , , and These are the weighting coefficients. For a function of the waiting time, For the service matching degree function, This is a proximity function. For energy state function, Indicates the first The passenger's class, Indicates taxi service preferences, Indicates the first The current location of the taxi Indicates the first The current location of each passenger. This indicates the remaining battery power or fuel level.

[0025] Preferably, the method further includes:

[0026] For taxis that complete short-distance orders, compensation points are automatically calculated and issued based on their actual waiting time and travel distance. The formula for calculating compensation points is as follows:

[0027]

[0028] in, Indicates the compensation points. Indicates taxi waiting time. , Indicates the distance to the passenger's destination. This represents the time period coefficient.

[0029] As a preferred option, the method for planning the optimal travel route for orders is as follows:

[0030] Generate multiple candidate routes, estimate the travel time and carbon emissions for each route, and calculate the overall cost for each route:

[0031]

[0032]

[0033] in, Indicates total cost. Indicates travel time. Indicates carbon emissions. Indicates idling time. Indicates path length. Indicates the number of traffic lights. Indicates the number of congested road sections. Indicates the length of the congested road section. Indicates the speed in congested sections of the road;

[0034] Choose the route with the lowest overall cost as the optimal driving route.

[0035] The beneficial effects of this invention are:

[0036] 1. This invention addresses the supply-demand mismatch at its source by adjusting lanes every 15 minutes and using a two-way passenger-taxi screening system, thereby reducing the likelihood of road rage among drivers.

[0037] 2. This invention calculates the correlation between waiting time and distance through compensation points and uses blockchain to ensure transparency and traceability, thereby increasing the willingness to accept short-distance orders.

[0038] 3. This invention achieves a win-win situation of "high efficiency and low carbon" by weighting driving time with carbon emissions, which is in line with the "dual carbon" goal.

[0039] 4. Through systematic design, this invention achieves for the first time the synergistic optimization of the interests of the "driver-passenger-environment" in hub taxi dispatch.

[0040] Secondly, a hub taxi intelligent dispatching system based on dynamic partitioning includes:

[0041] The perception layer is used to collect passenger data, taxi data, lane status data, and points storage data in real time.

[0042] The data layer is used to receive passenger data, taxi data, lane status data, and points storage data, and then build a database to store passenger data, taxi data, lane status data, and points storage data.

[0043] The core module layer is used to perform passenger classification, dynamic lane management, intelligent matching and scheduling, route generation and short-distance compensation based on passenger data, taxi data, lane status data and points storage data, and generate decision instructions.

[0044] The execution layer is used to perform lane type switching, order push, and points redemption operations based on decision instructions;

[0045] The feedback layer includes a passenger evaluation module, a driver behavior analysis module, and a carbon emission monitoring module. These modules are used to collect passenger evaluation data, driver behavior analysis data, and carbon emission monitoring data, respectively, and transmit them to the data layer to provide a basis for optimization of the core module layer, thereby achieving closed-loop control.

[0046] As a preferred option, the driver behavior analysis module analyzes drivers' historical service records and real-time driving behavior data to establish a risk assessment model for driver behavior risk warning; the formula for the risk assessment model is as follows:

[0047]

[0048] in, Indicates the risk value, when Intervention is initiated when the value exceeds a preset threshold. Indicates taxi waiting time. Indicates the first The proportion of short-distance orders for each taxi Indicates the first The complaint rate of taxis.

[0049] Thirdly, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method as described in the first aspect.

[0050] Fourthly, a non-transitory computer-readable storage medium is provided that stores computer instructions for causing a computer to perform the method as described in the first aspect. Attached Figure Description

[0051] Figure 1 The diagram shows a flowchart of a hub taxi intelligent scheduling method based on dynamic partitioning provided in Embodiment 1 of the present invention.

[0052] Figure 2 The figure shown is an architecture diagram of a hub taxi intelligent dispatching system based on dynamic partitioning provided in Embodiment 2 of the present invention.

[0053] Figure 3 The diagram shown is a flowchart of passenger classification processing provided in Embodiment 2 of the present invention.

[0054] Figure 4 The diagram shown is a flowchart of the dynamic lane adjustment algorithm provided in Embodiment 2 of the present invention.

[0055] Figure 5 The diagram shown is a schematic diagram of the passenger-taxi matching logic provided in Embodiment 2 of the present invention.

[0056] Figure 6 The diagram shown is a flowchart of carbon emission optimization path planning provided in Embodiment 2 of the present invention. Detailed Implementation

[0057] Exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the drawings are merely exemplary and are intended to illustrate the principles and spirit of the invention, and are not intended to limit the scope of the invention.

[0058] Example 1:

[0059] like Figure 1 As shown, a hub taxi intelligent dispatching method based on dynamic partitioning includes the following steps:

[0060] S1. Based on destination distance Passengers entering the hub are divided into three categories: short-distance, medium-distance, and long-distance, and a unique identification number is generated for each passenger.

[0061] Short-distance passenger P_short: ≤ 10km, numbered in the format P [YYYYMMDD] S [XXX];

[0062] Mid-journey passenger P_mid: 10km< ≤ 30km, numbered in the format P [YYYYMMDD] M [XXX];

[0063] Long-distance passenger P_long: >30km, numbered in the format P [YYYYMMDD] L [XXX];

[0064] S2. Perform status detection and coding management on taxis entering the hub area, and construct a set of vehicle attributes;

[0065] S3. The lanes in the hub waiting area are dynamically divided into three types: short-distance dedicated lanes, long-distance dedicated lanes, and flexible adjustment lanes. Based on real-time collected passenger type distribution data and taxi queuing data, the number of each type of lane is periodically adjusted.

[0066] Short-distance dedicated lane L short Numbers LS01-LS12;

[0067] Long-distance dedicated lane L long Numbers LL01-LL12;

[0068] Flexible lane adjustment L flex Numbers: LF01-LF08.

[0069] S4. Based on passenger level, identification number and vehicle attribute set, the optimal matching of passengers and taxis is achieved through a multi-factor weighted matching algorithm, and a dispatch instruction is generated to guide the vehicle into the lane and generate an order;

[0070] S5. Based on travel time and carbon emissions, plan the optimal travel route for orders and complete intelligent dispatching of taxis at hubs.

[0071] In this embodiment, the vehicle attribute set includes:

[0072] Basic coding: company code, vehicle ID, and service preferences; among which, service preferences include short-distance priority, medium-distance priority, long-distance priority, and flexible service;

[0073] Dynamic attributes: current location, cumulative waiting time, remaining battery or fuel, and historical service records.

[0074] In this embodiment, the lane number adjustment cycle is 15 minutes; the lane number adjustment is based on factors including queue length of each lane, passenger type distribution, taxi waiting time, and real-time traffic flow; the lane number adjustment decision is generated through a reinforcement learning model, and the reward function of the reinforcement learning model... for:

[0075]

[0076] in, , and This represents the weighting coefficients (β=0.5, γ=0.3, δ=0.2).

[0077] The formula for adjusting the number of lanes of each type is:

[0078]

[0079] in, This indicates the adjusted number of dedicated short-distance lanes. Indicates the proportion of short-distance passengers. This represents the influence coefficient of waiting time. This indicates the deviation rate between the current average waiting time and the baseline value.

[0080] In this embodiment, the multi-factor weighted matching algorithm adopts a two-layer screening mechanism, specifically:

[0081] First-level filtering: Compatibility filtering based on passenger level and taxi service preference identifiers;

[0082] The second layer of screening: For the candidate taxi set that passed the first layer of screening, a matching score between the taxi and the passenger is calculated. Long-distance passengers are preferentially matched with taxis with the longest waiting time, and short-distance passengers are matched with taxis with shorter waiting times or those with a preference for short-distance travel. The matching score... The calculation formula is:

[0083]

[0084] in, , , and These are the weighting coefficients. For a function of the waiting time, For the service matching degree function, This is a proximity function. For energy state function, Indicates the first The passenger's class, Indicates taxi service preferences, Indicates the first The current location of the taxi Indicates the first The current location of each passenger. This indicates the remaining battery power or fuel level.

[0085] In this embodiment, the method further includes:

[0086] For taxis that complete short-distance orders, compensation points are automatically calculated and issued based on their actual waiting time and travel distance. The formula for calculating compensation points is as follows:

[0087]

[0088] in, Indicates the compensation points. Indicates taxi waiting time. , Indicates the distance to the passenger's destination. The coefficient represents the time period (peak 1.2, off-peak 1.0, low-peak 0.8). Compensation points can be redeemed for: charging / fueling discounts (e.g., 1 point = 0.5 yuan), priority queuing rights (e.g., 100 points = 1 time), or parking fee reductions (e.g., 50 points = 1 hour), and the points records are tamper-proof through blockchain technology.

[0089] In this embodiment, the method for planning the optimal driving route for an order is as follows:

[0090] Generate multiple candidate routes, estimate the travel time and carbon emissions for each route, and calculate the overall cost for each route:

[0091]

[0092]

[0093] in, Indicates total cost. Indicates travel time. Indicates carbon emissions. Indicates idling time. Indicates path length. Indicates the number of traffic lights. Indicates the number of congested road sections. Indicates the length of the congested road section. Indicates the speed in congested sections of the road;

[0094] Choose the route with the lowest overall cost as the optimal driving route.

[0095] Example 2:

[0096] Based on Example 1, such as Figure 2 As shown, this embodiment of the invention provides a hub taxi intelligent dispatching system based on dynamic partitioning, which can be used to implement the hub taxi intelligent dispatching method based on dynamic partitioning as described in the foregoing embodiments. The system includes:

[0097] The perception layer is used to collect passenger data, taxi data, lane status data, and points storage data in real time. The core components of the perception layer include passenger terminals (self-service registration screen / APP / voice terminal), vehicle perception devices (GPS / license plate recognition / fuel level sensor), and environmental perception devices (lane camera / millimeter-wave radar / CO2 detector). It is supported by hardware such as a 10.1-inch touch screen, 4G / 5G module, high-definition camera (2 megapixels), and millimeter-wave radar (detection distance of 50m).

[0098] The data layer receives passenger data, taxi data, lane status data, and points storage data, and then builds a database to store these data. The core components of the data layer include a passenger database, a taxi database, a traffic database (real-time road conditions / lane status), and a blockchain storage repository (compensation points). It is supported by cloud servers (8 cores and 16GB) and blockchain nodes (consortium blockchain), supports the storage of structured data across the entire chain, supports more than 1,000 data queries per second, and ensures that points records are tamper-proof.

[0099] The core module layer is used to perform passenger classification, dynamic lane management, intelligent matching and scheduling, route generation, and short-distance compensation based on passenger data, taxi data, lane status data, and points storage data, and to generate decision instructions. The core module layer includes a passenger classification module, a dynamic lane management module, an intelligent matching and scheduling module, a short-distance compensation module, and a carbon emission optimization module. It adopts the hardware support of an edge computing gateway (2TOPS computing power) and has a response time of ≤100ms.

[0100] The execution layer is used to perform lane type switching, order push, and points redemption operations according to decision instructions. The core components of the execution layer include lane execution equipment (LED guidance screen / physical isolation controller / lane indicator light), dispatch terminal (driver APP / management platform), and compensation redemption terminal (charging pile / parking lot interface). The hardware includes LED screen (2560×1440 resolution), motor controller (0.5s response time) and Android smart terminal.

[0101] The feedback layer includes a passenger evaluation module, a driver behavior analysis module, and a carbon emission monitoring module. These modules collect passenger evaluation data, driver behavior analysis data, and carbon emission monitoring data, respectively, and transmit them to the data layer to provide a basis for optimization of the core module layer, thus achieving closed-loop control. The feedback layer hardware includes a QR code generator, an in-vehicle microphone (for noise reduction), and an OBD interface.

[0102] In this embodiment, as Figure 3 As shown, the passenger classification processing algorithm is as follows:

[0103] Passengers enter the hub and select their input method: self-service terminal / APP / human assistance;

[0104] Enter destination information (address / POI name);

[0105] GIS map system calculates destination distance D i (Using the Gaode API, accuracy ≤100m);

[0106] Judge D i Range: D i ≤ 10km, classified as P short Generate a serial number: P+YYYYMMDD+S+3-digit serial number (e.g., P20240827S045); 10km <D i ≤ 30km, classified as P mid Generate the following number: P+YYYYMMDD+M+3-digit serial number (e.g., P20240827M012); D i >30km, classified as P long Generate the following number: P+YYYYMMDD+L+3-digit serial number (e.g., P20240827L008).

[0107] Verification: Is the address ambiguous? If so, push a list of popular locations for the passenger to select again; if not, push a recommended waiting area (P) on the LED signage screen. short →LS lane, P long →LL lane) + Estimated waiting time (based on real-time queue);

[0108] Passengers proceed to the designated waiting area and enter the matching queue;

[0109] The grading process has ended; now we await intelligent matching.

[0110] The threshold for address ambiguity determination is "POI positioning accuracy > 500m", and the estimated waiting time is calculated based on "number of vehicles queuing in the current lane × average service time (8 minutes / order for short distances, 15 minutes / order for long distances)".

[0111] In this embodiment, as Figure 4 The specific flow of the dynamic lane adjustment algorithm shown is as follows:

[0112] Start adjusting the cycle (triggered every 15 minutes);

[0113] Data collection: Queue length L for each lane queue Passenger type percentage (R) short / R mid / R long Average taxi waiting time T avg Real-time traffic flow Q;

[0114] Constructing the reinforcement learning state space S={L queue , R short , T avg , Q};

[0115] Calculate the demand ratio R short =N short / N total R mid =N mid / N total R long =N long / N total (N) total (Total number of passengers currently waiting)

[0116] Calculate the baseline number of lanes: Short-distance lane baseline number of lanes L shortbase =Total number of lanes × R short The number of intermediate lanes, L midbase =Total number of lanes × R mid Long-distance lane base number L longbase =Total number of lanes × R long ;

[0117] Waiting time correction ΔT=(T avgshort -T base ) / T base (T) base =15 minutes, representing the baseline waiting time); α=0.2 (when ΔT>0, a short-distance lane needs to be added) / 0.1 (when ΔT≤0, fine-tuning is required); L shortnew =L shortbase ×(1+α×ΔT);

[0118] Constraint checks: ①L shortnew ≥2 (ensuring basic capacity for short-distance passengers); ②L flexnew =Total number of lanes - L shortnew -L midbase -L longbase ≥1 (with room for flexible adjustment); if not met, reduce L proportionally. midbase / L longbase ;

[0119] Calculate the reinforcement learning reward function R = -[β×T] avg + γ×C emission + δ×switching frequency], where β=0.5 (waiting time weight), γ=0.3 (carbon emission weight), and δ=0.2 (switching cost weight);

[0120] Determine if R is better than the historical best value (last 3 periods). If so, determine the new lane configuration: L shortnew / L midnew (L) midbase (Revised) / L longnew (L) longbase (Revised) / L flexnew If not, fine-tune the α value (±0.05) and re-perform the constraint check;

[0121] Execute the adjustment: ① Send a type switching command to the lane controller (e.g., “LF02→LS lane”); ② Update the LED guidance screen display (e.g., “Current LS lane: LS01-LS08”); ③ Trigger the movement of the physical barrier (electrically controlled, time ≤ 2 minutes);

[0122] Record the adjustment log to the blockchain (including adjustment time, previous and current configurations, and R value) to end this cycle.

[0123] Example of a certain station, with a total of 16 lanes (all lanes), during the morning rush hour (8:00-10:00) R short =65%, L shortbase =10.4, ΔT=0.4 (T avgshort =21 minutes).

[0124] L shortnew = 10.4×(1+0.2×0.4) =12, final configuration: LS01-LS12 (12 lines), LM01-LM02 (2 lines), LL01 (1 line), LF01 (1 line).

[0125] In this embodiment, as Figure 5 As shown, the passenger-taxi matching logic is as follows:

[0126] The input is the passenger queue P. queue (Passenger Class P) type Passenger Identification Number P id (and pick-up point), taxi queue T) queue (Taxi number T) id Taxi service preferences W j Taxi waiting time Taxi's current location (LOC) j , indicates remaining battery power / fuel level / E j );

[0127] First-level filtering: Type matching (excluding incompatible combinations)

[0128] Press P type Traffic splitting:

[0129] P short (Short distance), filter W j Taxis ∈ {short-distance preferred, flexible service} form a candidate set C. j1 (excluding W) j =Mid-distance priority / Long-distance priority);

[0130] P mid (Midway) Filter W j Taxis ∈ {priority for mid-journey, flexible service} form a candidate set C. j2 (excluding W) j =Short distance preferred / Long distance priority);

[0131] P long (Long distance), filter W j Taxis ∈ {long-distance priority, flexible service} form a candidate set C. j3 (excluding W) j =Short route preferred / Medium route preferred);

[0132] Second-level screening: Multi-factor scoring

[0133] Calculate the waiting time weight W t =1-exp(- / 30)( The unit is minutes; W t Range 0-1, the longer the wait W t The larger exp is, the more exponential the function (exp represents an exponential function).

[0134] Calculate type matching degree Service Match : Exact match (P short &W j =W j ) = 1.0, partial match (P short &W j =Flexible services)=0.7;

[0135] Calculate the distance weight Proximity = 1 - (LOC) j Distance from boarding point / 2000 (distance in meters; ≤500m=1.0, >2000m=0.2).

[0136] Calculate the energy weight E j For gasoline vehicles, the fuel percentage is calculated as 100; for electric vehicles, the battery percentage is calculated as 100 (≥80% = 1.0, ≤20% = 0.2).

[0137] Calculate the total score. ij =0.4×W t + 0.3×Service Match + 0.2×Proximity + 0.1×E j ;

[0138] Press P type Determine the sorting rules

[0139] P short Press Score ij Descending order (W preferred) t Smaller taxis (to avoid drivers who have to wait a long time taking short trips).

[0140] P long Press Score ij Descending order (W preferred) t Large taxis ensure that drivers who have been waiting for a long time can pick up long-distance passengers.

[0141] P mid Press Score ij Descending order (no additional bias, balanced supply and demand);

[0142] Select Top1 Taxi T best Generate matching pairs (P) i T best );

[0143] Pushing P-related messages to driver apps id (e.g., P20240827S045), boarding point (e.g., "LS03 lane at the East Square of a certain station"), D i Orders with a distance of "8.5km" or similar;

[0144] Pushing T-shirts to passengers id The matching notification includes (e.g., T03A1256S), license plate, and estimated pick-up time (e.g., "5 minutes").

[0145] Matching complete, driver proceeds to pick up passenger, who waits.

[0146] In terms of matching efficiency, this invention achieves a single batch matching time of ≤3 seconds, supports the simultaneous processing of matching needs for 500+ passengers and 300+ taxis, and achieves a matching accuracy of ≥95%.

[0147] In this embodiment, as Figure 6 As shown, the carbon emission optimization pathway planning process is as follows:

[0148] Input includes matching orders (origin S = passenger pick-up point, destination E = passenger destination) and taxi model parameter V (fuel / electric); real-time traffic API (map) is called to obtain traffic data: congestion index (0-10) for each road segment, number of traffic lights, and construction information; 3-5 candidate routes R are generated. candidate ;

[0149] Initialize path index k=1;

[0150] Extract the current path R k The basic data, including length D k (km), Number of traffic lights N light Number of congested road sections N congest ;

[0151] Estimated travel time T travel T travel = D k / (V base × (1 - 0.03 × congestion index)), where the vehicle speed V base =60km / h on city roads, 80km / h on highways;

[0152] Estimate idling time T idle (Unit: minutes):

[0153] T idle = N light ×30s (idle at a single traffic light) + N congest ×(D congest / V congest )×60s (idling speed in congested areas)

[0154] Among them, D congest V represents the length of the congested road segment. congest Indicates the speed in congested sections of the road;

[0155] Calculate carbon emissions C emission Gasoline-powered vehicle C emission = 0.2×D k + 0.05×T idle (kg);

[0156] Electric vehicle C emission = 0.08×D k + 0.02×T idle (kg, converted to carbon emissions from electricity);

[0157] Calculate the overall cost (Cost) k = 0.6×T travel (Time weight) + 0.4×Cemission (Carbon emission weights);

[0158] Check if k equals the total number of candidate paths. If not, increment the count k by 1 and recalculate the travel time. If yes, filter by Cost. k The path with the shortest length is the optimal path R. opt ;

[0159] Perform a second check to determine the optimal path R. opt Are there any sudden traffic jams (congestion index ≥ 8) or temporary construction? If so, exclude R. opt The optimal route is re-selected from the remaining candidate routes and validated a second time; if not, a notification is sent to the driver's app. opt Includes real-time steering guidance, estimated energy consumption (e.g., "estimated fuel consumption 0.8L"), and estimated arrival time (T). travel + Pick-up time);

[0160] Push the estimated arrival time to the passenger's app;

[0161] Route planning complete, driver presses R opt While driving, the system updates road conditions in real time and dynamically adjusts the route.

[0162] In this embodiment, the driver behavior analysis module analyzes the driver's historical service records and real-time driving behavior data to establish a risk assessment model for driver behavior risk warning; the formula for the risk assessment model is as follows:

[0163]

[0164] in, Indicates the risk value, when Intervention is initiated when the value exceeds a preset threshold. Indicates taxi waiting time. Indicates the first The proportion of short-distance orders for each taxi Indicates the first The complaint rate of each taxi. This invention constructs a risk model based on driving behavior and historical data to identify high-risk drivers in advance and intervene proactively rather than handling the situation afterward.

[0165] According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0166] In an exemplary embodiment, the electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the hub taxi intelligent dispatching method based on dynamic partitioning as described in Embodiment 1 above.

[0167] In an exemplary embodiment, the readable storage medium may be a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the hub taxi intelligent dispatching method based on dynamic partitioning as described in Embodiment 1 above.

[0168] In an exemplary embodiment, the computer program product includes a computer program that, when executed by a processor, implements the hub taxi intelligent dispatching method based on dynamic partitioning as described in Embodiment 1 above.

[0169] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0170] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0171] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0172] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0173] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0174] Example 3:

[0175] Based on Example 1, this embodiment of the invention uses a certain station as an example to illustrate the technical effects of the invention through effect calculation. The facility configuration is shown in Table 1.

[0176] Table 1. Facility Configuration of a Certain Station

[0177]

[0178] 1. Road rage incidence calculation

[0179] Data sources: driver service ratings (1-5 points), passenger complaint records (monthly statistics), and in-vehicle audio sentiment analysis (keyword identification: "annoyed", "too slow", etc.);

[0180] Calculation formula: Road rage incidence rate = (Number of orders with a score ≤ 2 + Number of valid complaints × 5) / Total number of orders × 100% (complaint weight × 5, as a single complaint affects multiple passengers);

[0181] Result: 38% before implementation → 14.3% after implementation, a decrease of 62.3%.

[0182] 2. Carbon emission calculation

[0183] Target of the calculation: 12,000 taxis per day;

[0184] Carbon emission formula for a single vehicle: C = 0.2 × D actual + 0.05×(T idle + T wait (Before implementation); C' = 0.2×D opt + 0.05×(T idleopt + T waitopt (After implementation); where C' represents the carbon emissions per vehicle after adopting the method proposed in this invention, and D opt R represents the optimal path opt The length of T idleopt R represents the optimal path opt Idle time, T waitopt R represents the optimal path opt The corresponding waiting time.

[0185] Result: Daily total carbon emissions were 14.4 tons before implementation, reduced to 11.7 tons after implementation, a decrease of 18.7%.

[0186] 3. Passenger Experience Measurement

[0187] Indicators: average waiting time, service satisfaction (out of 5), complaint rate;

[0188] Results: Waiting time decreased from 42 minutes to 24 minutes (a reduction of 42.9%), satisfaction score increased from 3.2 to 4.7 (an increase of 46.9%), and complaint rate decreased from 8.5% to 2.1% (a decrease of 75.3%).

[0189] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A dynamic partition-based hub taxi intelligent dispatching method, characterized in that, The method comprises the following steps: According to the distance of the destination, the passengers entering the hub are divided into three levels including short, medium and long, and a unique identification number is generated for each passenger; The state of the taxi entering the hub area is detected and coded, and a vehicle attribute set is constructed; The lanes in the hub waiting area are dynamically divided into three types: short-distance dedicated lanes, long-distance dedicated lanes, and flexible adjustment lanes, and the number of each type of lane is periodically adjusted based on real-time collected passenger type distribution data and taxi queue data; the adjustment period of the number of lanes is 15 minutes; the adjustment of the number of lanes is based on the queue length of each lane, the passenger type distribution, the taxi waiting time, and the real-time traffic flow; the lane number adjustment decision is generated by a reinforcement learning model, and the reward function of the reinforcement learning model is: F: wherein , and represent weight coefficients; Based on the passenger level, identification number and vehicle attribute set, the optimal matching of passengers and taxis is realized through a multi-factor weighted matching algorithm, and a dispatching instruction is generated to guide the vehicle into the lane and generate an order; The multi-factor weighted matching algorithm adopts a double-layer screening mechanism, specifically: First layer screening: compatibility filtering according to passenger level and service preference identification of taxi; Second layer screening: for the candidate taxi set passing the first layer screening, calculate its matching score with the passenger, the matching score The calculation formula is: wherein, , , and are weight coefficients, is a function related to waiting time, is a service matching function, is a distance proximity function, is an energy status function, denotes a rank of the th passenger, denotes a taxi service preference, denotes a current location of the th taxi, denotes a current location of the th passenger, denotes a remaining amount of electricity or oil; Based on the driving time and carbon emissions, the optimal driving path of the order is planned, and the intelligent dispatching of the hub taxi is completed.

2. The dynamic zone-based hub taxi dispatching method of claim 1, wherein, The vehicle attribute set includes: Basic coding: company code, vehicle ID and service preference; Among them, the service preference includes short trip priority, medium trip priority, long trip priority and flexible service; Dynamic attributes: current location, cumulative waiting time, remaining power or oil, and historical service records.

3. The dynamic zone-based hub taxi cab intelligent dispatching method according to claim 1, wherein, The method further comprises: For the taxi completing the short trip order, based on its actual waiting time and driving distance, the compensation points are automatically calculated and issued, and the calculation formula of the compensation points is: wherein, denotes a compensation integral, denotes a taxi waiting time, , denotes a passenger destination distance, denotes a time period coefficient.

4. The dynamic zone-based hub taxi cab intelligent dispatching method according to claim 1, wherein, The method for planning the optimal driving path of the order is specifically: Generate multiple candidate paths, estimate the driving time and carbon emissions of each path, and calculate the comprehensive cost of each path: wherein, denotes the overall cost, denotes the travel time, denotes the carbon emission, denotes the idling time, denotes the path length, denotes the number of traffic lights, denotes the number of congested segments, denotes the length of congested segments, denotes the speed of congested segments; Select the path with the minimum comprehensive cost as the optimal driving path.

5. A hub taxi intelligent dispatching system based on the hub taxi intelligent dispatching method based on dynamic zoning according to any one of claims 1-4, characterized in that, It includes: The perception layer is used to collect passenger data, taxi data, lane state data and integral storage data in real time; The data layer is used to receive passenger data, taxi data, lane state data and integral storage data, and then construct a database for storing passenger data, taxi data, lane state data and integral storage data; The core module layer is used to classify passengers, dynamically manage lanes, intelligently match and dispatch, generate paths and short trip compensation based on passenger data, taxi data, lane state data and integral storage data, and generate decision instructions; The execution layer is used to perform lane type switching, order pushing and integral redemption operations according to the decision instructions; The feedback layer includes a passenger evaluation module, a driver behavior analysis module and a carbon emission monitoring module, which are used to collect passenger evaluation data, driver behavior analysis data and carbon emission monitoring data respectively, and transmit them to the data layer to provide basis for optimization of the core module layer, realizing closed-loop control.

6. The hub taxi brokerage system of claim 5, wherein, The driver behavior analysis module establishes a risk assessment model for driver behavior risk warning by analyzing the historical service records and real-time driving behavior data of the driver; The expression formula of the risk assessment model is: in, Indicates the risk value, when Intervention is initiated when the value exceeds a preset threshold. Indicates taxi waiting time. Indicates the first The proportion of short-distance orders for taxis. Indicates the first The complaint rate of taxis.

7. An electronic device, comprising: It includes: At least one processor; And a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to make the at least one processor execute the method of any one of claims 1-4.

8. A non-transitory computer-readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to make the computer execute the method of any one of claims 1-4.

Images