A method and system for dispatching a network car, a storage medium and an electronic device

Abstract

The application discloses a network car-hailing order dispatching method and system, a storage medium and an electronic device. The method comprises the following steps: obtaining order information and transport capacity information; predicting newly generated order information and transport capacity information in a plurality of preset time domains in the future according to the obtained order information and transport capacity information; determining a control strategy in the future preset time domain according to the newly generated order information and transport capacity information, taking the maximization of the total order matching quantity in the plurality of preset time domains in the future as the target; and adopting batch matching to match the number of orders matched in any region in the future preset time domain and going to other regions, taking the optimization of the average pick-up distance as the target. Through the implementation of the application, the available transport capacity and the number of orders to be matched in each region in a future period of time are considered, as well as the influence of the future spatial distribution of the matching transport capacity, so that the solution can be obtained in a short time. Meanwhile, the travel experience is ensured, the total order matching quantity of the platform party is improved, and the application has certain interpretability and is convenient for practice landing.

Description

Technical Field

[0001] This invention relates to the field of ride-hailing technology, specifically to a ride-hailing dispatching method, system, storage medium, and electronic device. Background Technology

[0002] In recent years, the emergence and development of ride-hailing platforms have drastically changed people's travel habits. Before ride-hailing services, "difficulty in hailing a taxi" was a widespread problem. The main reasons were the mismatch between demand and capacity information and the inability to connect supply and demand information in real time. Ride-hailing platforms, relying on internet technology, integrate supply and demand information, matching users and drivers to provide online travel services. Ride-hailing services have had a huge impact on the transportation industry, significantly reducing passenger waiting times for taxis.

[0003] As ride-hailing services have developed, numerous problems have emerged that urgently need to be addressed, particularly the supply-demand matching issue. This problem affects the overall performance and efficiency of ride-hailing platforms, the volume of order matching, and revenue, while also impacting the user's travel experience. Therefore, improving the platform's order matching volume while ensuring a positive user experience is a crucial challenge. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a ride-hailing order dispatching method, system, storage medium, and electronic device to solve the technical problem of low order matching volume in existing ride-hailing platforms.

[0005] The technical solution proposed in this invention is as follows:

[0006] The first aspect of this invention provides a ride-hailing order dispatch method, comprising: acquiring order information and capacity information; predicting newly generated order information and capacity information in multiple preset time domains based on the acquired order information and capacity information; determining a control strategy within the future preset time domains based on the newly generated order information and capacity information, with the goal of maximizing the total order matching volume in the multiple preset time domains, the control strategy including the number of orders matched in any region that travel to other regions; and performing batch matching based on the number of orders matched in any region that travel to other regions within the future preset time domains, with the goal of optimizing the average pick-up distance.

[0007] Optionally, before predicting new order information and capacity information generated in multiple preset time domains based on the acquired order information and capacity information, the method further includes: dividing the entire operating area into multiple regions based on the order information and capacity information; and dividing the time into multiple time domains at equal intervals.

[0008] Optionally, based on newly generated order information and capacity information, with the objective of maximizing the total order matching volume across multiple preset time domains, the number of matched orders destined for other regions within any region in the future preset time domains is determined. This includes: constructing a mixed-integer linear programming model based on the objective of maximizing the total order matching volume across multiple preset time domains, capacity constraints, demand constraints, and variable constraints. The capacity constraints include the constraints between unmatched legacy capacity, capacity in newly generated capacity information, and arriving capacity. The demand constraints include the constraints between unmatched orders and orders in newly generated order information. The mixed-integer linear programming problem is then solved to obtain the number of matched orders destined for other regions within any region in the future preset time domains.

[0009] Optionally, based on the number of orders going to other regions matched in any region within a future preset time domain, batch matching is performed with the goal of optimizing the average pick-up distance. This includes: selecting a region, performing multiple batch matching operations based on the number of orders going to other regions matched in that region within a future preset time domain, and performing matching with the goal of optimizing the average pick-up distance in each batch matching operation; selecting other regions one by one, and performing multiple batch matching operations in each region until matching is completed for all regions.

[0010] Optionally, based on the number of matched orders destined for other regions within the future preset time domain, multiple batch matching is employed. In each batch matching, the goal is to optimize the average pick-up distance. This includes: before each batch matching, determining the target number to be completed based on the number of matched orders destined for other regions within the future preset time domain and the number of matched orders in the region; determining the available capacity in the region and the size of the number of orders to be matched from the region to other regions; and directly matching when the available capacity is greater than or equal to the number of orders to be matched.

[0011] Optionally, the ride-hailing dispatch method further includes: when the number of available capacity is less than the number of orders to be matched, determining whether the target number of orders to other regions matched in the current time domain for the region has been achieved; when the target number of orders to other regions matched in the current time domain for the region has been achieved, matching capacity and orders with the goal of optimizing the average pick-up distance.

[0012] Optionally, the ride-hailing dispatch method further includes: when the target number of orders to other regions matched in the current time domain of the region has not been achieved, determining the number of orders to each region in the next batch matching based on the proportion of orders to be completed in each region; using the number of orders to each region in the current time domain of the region as a constraint to minimize the average pick-up distance for matching capacity and orders.

[0013] A second aspect of this invention provides a ride-hailing dispatch system, comprising: a data layer for acquiring order information and capacity information; a prediction layer for predicting newly generated order information and capacity information in multiple preset time domains based on the acquired order information and capacity information; an allocation layer for determining a control strategy within the future preset time domains based on the newly generated order information and capacity information, with the goal of maximizing the total order matching volume in the multiple preset time domains, wherein the control strategy includes the number of orders matched in any region that travel to other regions; and an execution layer for performing batch matching based on the number of orders matched in any region within the future preset time domains that travel to other regions, with the goal of optimizing the average pick-up distance.

[0014] A third aspect of the present invention provides a computer-readable storage medium storing computer instructions for causing the computer to perform the ride-hailing dispatch method as described in the first aspect and any one of the first aspects of the present invention.

[0015] A fourth aspect of the present invention provides an electronic device, including: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the ride-hailing dispatch method as described in the first aspect and any one of the first aspects of the present invention.

[0016] The technical solution provided by this invention has the following effects:

[0017] The ride-hailing dispatch method, system, storage medium, and electronic device provided in this invention establish an online real-time dispatch method that considers both transportation capacity and orders. It takes into account the available transportation capacity and the number of orders to be matched in each region over a future period, as well as the impact of matching on the future spatial distribution of transportation capacity. This effectively decouples and breaks down large-scale, complex dispatching problems, enabling solutions in a short time. This dispatch method ensures a better travel experience, increases the platform's total order matching volume, and has a certain degree of interpretability, facilitating practical implementation in real-world scenarios.

[0018] The ride-hailing dispatch system provided in this invention utilizes a model-based predictive control framework, changing the current first-come-first-served and batch matching approach. This method only needs to construct a predictive model of newly added supply and newly generated demand in each region over the next T time domains using historical supply and demand data, and then input this model into the allocation layer model. The allocation layer considers future supply and demand information, matched order information, and the remaining supply and demand quantity, aiming to maximize the total order matching volume over the next T time domains. It determines the number of orders from matched orders in region i to travel to region j in the next time domain, and inputs this as the control strategy into the execution layer model. In each specific matching process in the execution layer, the capacity of each region is allocated to the demand traveling to each region according to the proportion of the number of orders to be matched. It does not require extensive real-time scenario judgment and data processing, thus placing low demands on processor speed and bandwidth. Attached Figure Description

[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a ride-hailing order dispatching method according to an embodiment of the present invention;

[0021] Figure 2 This is a flowchart of a ride-hailing order dispatching method according to another embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of the control strategy of the ride-hailing order dispatching method according to an embodiment of the present invention;

[0023] Figure 4 This is a flowchart of a ride-hailing order dispatching method according to another embodiment of the present invention;

[0024] Figure 5 This is a flowchart of a ride-hailing order dispatching method according to another embodiment of the present invention;

[0025] Figure 6 This is a flowchart of a ride-hailing order dispatching method according to another embodiment of the present invention;

[0026] Figure 7 This is a schematic diagram of area settings in a simulation example of the ride-hailing order dispatching method according to an embodiment of the present invention;

[0027] Figure 8 This is a schematic diagram showing the results of three matching methods in an application scenario of the ride-hailing dispatching method according to an embodiment of the present invention.

[0028] Figure 9 This is a schematic diagram of the results of three matching methods in another application scenario of the ride-hailing dispatching method according to an embodiment of the present invention;

[0029] Figure 10 This is a structural block diagram of a ride-hailing dispatch system according to an embodiment of the present invention;

[0030] Figure 11 This is a structural block diagram of a ride-hailing dispatch system according to another embodiment of the present invention;

[0031] Figure 12 This is a schematic diagram of the structure of a computer-readable storage medium provided according to an embodiment of the present invention;

[0032] Figure 13 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] As mentioned in the background section, improving platform order matching volume while ensuring a good user travel experience is a crucial issue. Existing order dispatch methods mainly fall into three categories: first-come, first-served (FFS), batch matching, and reinforcement learning-based methods. The FFS method assigns order requests to the nearest available capacity. However, when supply is tight, the platform, in pursuit of fairness, may match available capacity with passengers from further away and with the longest waiting times, resulting in a significant waste of effective capacity during pick-up and drop-off, without considering overall system efficiency. The batch matching method accumulates available capacity and demand over time to form capacity and demand pools, aiming to minimize total pick-up time through combination optimization. However, this method does not consider the impact of each supply-demand match on future capacity spatial distribution. Reinforcement learning-based supply-demand matching methods tend to select orders with higher future revenue value, but they require large amounts of data for training and the model's response time to real-time situations is slow, thus limiting their practical application.

[0035] Based on this, embodiments of the present invention provide a ride-hailing order dispatching method that, while ensuring the user's travel experience, aims to maximize the number of matching orders, considers future supply and demand information, and has a certain degree of interpretability. For example... Figure 1 As shown, the method includes the following steps:

[0036] Step S101: Obtain order information and transportation capacity information.

[0037] In one embodiment, order information includes the time and latitude / longitude information of the order creation. Capacity information includes the time and latitude / longitude information of capacity availability. Order information can also be represented as demand information, i.e., information about the need to book a ride; the latitude / longitude information in the order information can be the current location information of the user requesting a ride. Capacity information can also be represented as supply information; the latitude / longitude information in the capacity information can be the current location information of the vehicles.

[0038] In one embodiment, to facilitate the allocation of capacity and orders, the acquired information can be preprocessed. Specifically, the entire region can be divided into partitions based on the areas covered by the latitude and longitude information in the order and capacity information, such as the entire region being represented as... The partitioned region is denoted as r, where At the same time, time can also be divided into different time domains at equal intervals. Considering the significant uncertainties in demand and capacity, as well as the difficulty of forecasting, a time range of 5-15 minutes can be set.

[0039] Step S102: Predict newly generated order and capacity information for multiple preset time domains in the future based on the acquired order and capacity information. Specifically, in order to formulate a specific allocation strategy, information for multiple preset time domains in the future can be predicted based on the acquired information. For example, order and capacity information from previous days and order and capacity information up to the current time of the current day can be used as independent variables to predict newly generated order and newly added capacity information for each region in the next T time domains. Wherein, when the time window moves to the beginning of the k-th time domain, the multiple preset time domains in the future, or the next T time domains, can be represented as...

[0040] Step S103: Based on the newly generated order information and capacity information, with the goal of maximizing the total order matching volume in multiple preset time domains in the future, determine the control strategy within the preset time domains in the future. The control strategy includes the number of orders matched in any region that are destined for other regions.

[0041] Specifically, to maximize the total order matching volume across multiple preset time domains, a control strategy for these multiple preset time domains can be formulated. This strategy includes how to select orders based on their destination. However, considering the coupling between different time domains and the impact of the implemented strategy on the system state, inaccurate predictions of future supply and demand in various regions can occur. Therefore, although the total order matching volume across multiple preset time domains is formulated, the control strategy is only executed within the k-th preset time domain. When the time window moves forward to the beginning of the (k+1)-th time domain, the system state and the supply and demand in each region for the next T time domains are predicted again, and a new control strategy is formulated, rolling forward.

[0042] Step S104: Based on the number of matched orders to other regions in any region within a future preset time domain, batch matching is performed with the goal of optimizing the average pick-up distance. Specifically, when determining the number of matched orders to other regions in any region within a future preset time domain, the process can be partitioned. For example, first select a region, and based on the number of matched orders to other regions in that region within the future preset time domain, perform multiple batch matching operations, with the goal of optimizing the average pick-up distance in each batch matching operation. Then select other regions, and perform multiple batch matching operations in each region until matching is completed in all regions. During partitioned execution, the matching process can be performed simultaneously in these regions, or it can be performed sequentially region by region. In one embodiment, when executing the control strategy for the k-th time domain, it can be completed through multiple specific matching processes. Batch matching is used in each specific matching operation. The longer the specific matching time, the longer the passenger waiting time and the worse the user experience. Therefore, in practice, the duration of each specific matching operation can be around 2-20 seconds.

[0043] Batch matching refers to accumulating demand for a period of time (e.g., 5 seconds) instead of matching it immediately when demand arises, forming a capacity pool and a demand pool (e.g., 5 demands and 2 capacity). Then, the combination is optimized with the goal of minimizing the pick-up distance, that is, considering how to match 2 capacity to 5 demands.

[0044] The ride-hailing dispatch method provided in this invention establishes an online real-time dispatch method that considers both transportation capacity and orders. It takes into account the available transportation capacity and the number of orders to be matched in each region over a future period, as well as the impact of matching on the future spatial distribution of transportation capacity. This effectively decouples and breaks down large-scale, complex dispatching problems, allowing for quick solutions. This dispatch method ensures a better travel experience, increases the platform's total order matching volume, and has a degree of interpretability, making it easy to implement in real-world scenarios.

[0045] As an optional implementation of this invention, such as Figure 2 As shown, step S103, based on the newly generated order information and capacity information, aims to maximize the total order matching volume across multiple preset time domains by determining the number of orders matched in any region within the future preset time domains that are destined for other regions. This includes:

[0046] Step S201: Construct a mixed integer linear programming model based on the objective of maximizing the total order matching volume of multiple preset time domains in the future, capacity constraints, demand constraints, and variable constraints. Capacity constraints include the constraint relationship between unmatched legacy capacity, capacity in newly generated capacity information, and capacity that will arrive. Demand constraints include the constraint relationship between unmatched orders and orders in newly generated order information.

[0047] In one embodiment, such as Figure 3 As shown, the control strategy for the next T time domains can be formulated before the k-th control time domain by rolling the time window. This control strategy can be represented as the following mathematical programming problem: the decision variables are... This indicates that in the t-th control time domain, region i should have An order with region j as its destination is matched, or in other words, there should be an order in any region. Orders destined for the second region are matched. The objective is to maximize the total number of matched orders for each region over the next T control time domains. in and Representing respectively in the In each time domain, the number of orders to be matched and the number of transport capacity in region r are represented by the superscript d (demand), which represents demand (passengers, or the number of orders); and the superscript s (supply), which represents transport capacity (drivers, or the number of transport capacity). This represents the number of matched orders in time domain t and region r. The constraints are capacity constraints, demand constraints, and variable constraints.

[0048] In one embodiment, for capacity constraints, i.e., future time domain Central region The capacity consists of the unmatched capacity remaining in region r in time domain t-1, newly added capacity, and capacity currently in service that will arrive in region r in time domain t. The number of newly added capacity... The remaining capacity in the t-th time domain region r, as predicted, is denoted as... When formulating the control strategy for the k-th time domain, the number of unmatched transport capacities in the (k-1)-th time domain. Given a quantity, when capacity is waiting for a time domain that has not been matched, there is a probability that... Choose to leave. Capacity currently in service and about to arrive in region r includes capacity that was already booked before the start of time region k. And the capacity to accept orders according to the control strategy after time domain k. in Given quantities It can be represented as the capacity that, after time k, receives orders according to the control strategy and arrives in region r in time t. This indicates that in the t-th control time domain, region j should have An order destined for region r is matched. Starting from region j, to reach region r in time domain t, it needs to start in time domain t-a[jr], where a[jr] represents the number of time domains traversed to drive from region j to region r. Therefore, This is equivalent to summing the number of orders originating from all regions and arriving in region r from time domain t.

[0049] Specifically, capacity constraints can be represented by the following constraints:

[0050] (1)

[0051] (2)

[0052] (3)

[0053] (4)

[0054] (5)

[0055] in, Indicates that any t belongs to However, it does not include {k}. For time domain k, the number of transport capacities remaining from time domain k-1. It is known that in formula (3), t = k, therefore, in formula (3) It is known. However, for time domain k, none of the events have occurred, and t in formula (4) belongs to However, it does not include {k}, therefore, the remaining capacity in formula (4) It belongs to the unknown quantity. Formula (5) means: in the t-th time domain, the total transport capacity of region r is It equals the sum of the number of matched capacity and the number of unmatched, remaining capacity after the end of this time period.

[0056] In one embodiment, in the future time domain Central region The number of demands consists of unmatched orders from the previous time period and newly generated orders. When a demand remains unmatched for a certain period, there is a probability... Choose to leave. and Let represent the predicted value of newly generated orders and the number of unmatched orders in the time domain t region r, respectively. This can be expressed as the following constraint:

[0057] (1)

[0058] (2)

[0059] (3)

[0060] Among them, the number of requirements left over from the k-1 time domain is the same as in formulas (3) and (4) above. This is a known quantity. The number of remaining requirements after time domain k. It is an unknown quantity. Formula (3) means that in the time domain t, the total number of demands in region r is equal to the sum of the number of legacy demands and the number of matched demands.

[0061] In one embodiment, the variable constraints include legacy orders. Number of transport capacities and decision variables Furthermore, all three are non-negative numbers, specifically represented by the following three formulas:

[0062] (1)

[0063] (2)

[0064] (3)

[0065] However, the model obtained from the above mathematical programming is a nonlinear model, resulting in a slow solution speed. The model can be linearized by constructing integer variables of 0 and 1 using the minimum function. The minimum function refers to the minimum value where, for time t, the number of transport capacities to be matched in region r is equal to the sum of the available capacity and demand in that region. Let Z be the minimum value. tr Let be the order matching quantity for the t-th time-domain region r. The following constraints can be obtained. The min linear constraint is aided by an integer M, where M is a large number, such as 10000. Specifically, for Z... tr The following constraints can be applied:

[0066] (1)

[0067] (2)

[0068] (3)

[0069] (4)

[0070] (5)

[0071] (6)

[0072] in, and As an auxiliary variable, and The value can only be 0 or 1.

[0073] By linearizing the above mathematical programming problem, we can obtain the following mixed-integer linear programming (MILP) problem:

[0074]

[0075]

[0076] Subject to:

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094] Step S202: Solve the mixed-integer linear programming problem to obtain the number of orders matched in any region within the preset future time domain that travel to other regions. Specifically, by solving the above mixed-integer linear programming problem, the number of orders matched in region i and traveling to region j in the k-th time domain can be obtained.

[0095] As an optional implementation of this invention, such as Figure 4 As shown, based on the number of orders destined for other regions matched in this region within a future preset time domain, multiple batch matching is performed. In each batch matching, the goal is to optimize the average pick-up distance. The process includes the following steps:

[0096] Step S301: Before each batch matching, determine the matched orders and the number of orders to be completed based on the number of matched orders destined for other regions within the preset future time domain; specifically, when the time window rolls to the k-th time domain, before a specific matching in region i, it is known that: after the k-th time domain ends, region i should have A demand destined for region j is matched (i.e., the number of orders destined for other regions matched in any region within the future preset time domain, obtained through step S103, is known). In solving mixed-integer linear programming problems, in order to reduce computational complexity, no... Therefore, with integer constraints, It is a positive real number; it has been matched before this matching process. If there are orders destined for region j, then the number of orders to be completed, starting from region i and ending at region j, is... The three satisfy the following relationship:

[0097] Step S302: Determine the available capacity of the region and the number of orders from the region to other regions awaiting matching; specifically, before matching, it is also necessary to determine the available capacity of any region, i.e., the number of available vehicles in region i. And the size of the number of pending orders from any region to other regions, i.e., the actual number of pending orders from region i to region j. The size is used to determine the value.

[0098] Step S303: When the number of available shipping capacity is greater than the number of orders to be matched, match directly. Specifically, when the number of available shipping capacity is greater than or equal to the number of orders to be matched, i.e. At this point, regardless of whether the higher-level objective has been achieved, the number of objectives to be achieved is... Whether the order is completed or not, batch matching of available capacity and demand can be performed directly. At this point, all orders can be fulfilled, meaning that the current available capacity can fully meet the demand. When performing direct matching, to improve overall efficiency and the user's travel experience, matching can be performed with the goal of reducing the average pick-up distance.

[0099] Specifically, let all the unmatched demands in region i form a set. The capacity is aggregated as in Let p represent the p-th order in region i, with the destination being... This represents the q-th transport capacity in region i. (Set) medium elements This represents the proximity between demand (p) and supply (q). (Variable) when Match the p-th order in region i to the q-th capacity; otherwise...

[0100] Therefore, this batch matching process can be categorized into the following mathematical problem:

[0101]

[0102] Subject to

[0103]

[0104]

[0105] In one embodiment, such as Figure 5 As shown, this ride-hailing dispatching method also includes the following steps:

[0106] Step S401: When the number of available transport capacity is less than the number of orders to be matched, determine whether the number of orders to other regions matched in the current time domain for this region has been completed; specifically, when the number of available transport capacity is less than the number of orders to be matched, i.e. At this point, it is necessary to determine whether the upper-level goal has been achieved, i.e., the set number of orders. Is it complete?

[0107] Step S402: When the number of orders to other regions matched in the current time domain for this region is completed, optimize the average pick-up distance by matching capacity and orders. Specifically, when the set number of orders has been completed, it is necessary to match capacity and orders with the goal of optimizing the average pick-up distance. The specific matching process can be represented by the following mathematical problem:

[0108]

[0109] Subject to

[0110]

[0111]

[0112] In one embodiment, such as Figure 6 As shown, this ride-hailing dispatching method also includes the following steps:

[0113] Step S501: When the number of orders destined for other regions matched in the current time domain for this region is not yet completed, determine the number of orders that should be matched from this region to each region in the next batch matching based on the proportion of orders to be completed in each region; specifically, when the target number of orders to be completed is not yet completed, it is necessary to first determine the number of orders x that should be matched from the demand from region i to region j. ij x ij This is an integer variable. It represents the number of available vehicles in region i. The available vehicles need to travel to multiple regions, i.e., j = 1, 2, 3... The proportion of orders to each region is determined based on the number of orders that should travel to each region. Specifically, this can be achieved by considering the actual demand while allocating resources according to the proportion of orders to be completed in each region, using the following integer programming problem. For region i:

[0114]

[0115] Subject to

[0116]

[0117]

[0118]

[0119]

[0120] Step S502: Using the number of orders that should be matched from region i to each region in the current time domain as a constraint, capacity and orders are matched to minimize the average pick-up distance. Specifically, the number of matched orders x from region i to region j is determined. ij This constraint is incorporated into the following mathematical model. The objective function is to minimize the average pick-up distance of the matched orders.

[0121]

[0122] Subject to

[0123]

[0124]

[0125]

[0126] The ride-hailing order dispatch method provided in this embodiment can be applied to pick-up scenarios: Assuming there are three regions A, B, and C, and the time window is pushed to the k-th control time domain, step S103 calculates that in the next control time domain (here, 10 minutes), 10 orders to region A, 4 orders to region B, and 6 orders to region C should be matched in region A. Before the next specific matching process (batch matching every 10 seconds), 1, 1, and 3 orders to regions A, B, and C have already been matched respectively. In region A, there are 30, 20, and 20 orders to regions A, B, and C waiting to be matched respectively. It is known that there are 5 available vehicles in region A in the next supply and demand matching. Therefore, through step S104, combined with the number of orders to be matched, the number of vehicles matched to orders to each region is determined according to the ratio of the number of orders to be completed at each destination. In this specific embodiment, this is transformed into the 5 available vehicles in region A, according to the ratio of the number of orders to be completed at each destination: 9:3:3 (according to the formula). The result is (10-1):(4-1):(6-3). Five transport capacities are allocated to the demand in regions A, B, and C. The specific matching is constrained by the proportion of orders to be completed. In actual matching, the five available transport capacities are divided into 3, 1, and 1 orders respectively destined for regions A, B, and C. Once all specific matching processes in the k-th time domain are completed, the time window moves to the beginning of the next time domain, k+1, for re-prediction and re-formulation of control strategies, and so on, rolling forward.

[0127] In one embodiment, the technical effectiveness of the ride-hailing dispatch method provided by the present invention is verified through simulation experiments. To verify the performance of this ride-hailing dispatch method, traditional first-come-first-served and batch matching methods are selected as comparisons for simulation experiments. The experiments are designed to cover three regions: A, B, and C, which are plotted on the coordinate axis as follows: and A square with a side length of 3 centered at a certain point, such as Figure 7 As shown in the diagram. In this scenario, the distance between two points on the map is represented by Euclidean distance, travel time is simplified to be proportional to distance, the detour ratio is 1.3, and the vehicle speed is a constant 30 km / h. Both the driver and passengers have a certain tolerance for unmatched waiting times, which are assumed to follow a constant μ. s =1200,μ d =900, an exponential distribution. The timing of capacity rollout and demand generation in the three regions follows a Gaussian mixture distribution with different parameters.

[0128] Consider two different demand patterns: a small supply-demand gap and a large supply-demand gap. The specific simulation parameter settings are shown in Tables 1 and 2.

[0129] Table 1

[0130]

[0131]

[0132] Table 2

[0133]

[0134] Tables 1 and 2 provide metrics such as order matching volume and average pick-up distance for the three schemes under different demand patterns (First-Come, First-Served (FCFS), batch matching, and the MPC-dispatch method of this invention). Figure 8 and Figure 9 As shown. This ride-hailing dispatch method includes a filtering mechanism based on order destinations. This filtering significantly improves order matching rates, increasing the rate by 15.66% compared to first-come, first-served services when the supply-demand gap is small, and by 25.20% compared to first-come, first-pay services when the supply-demand gap is large. This is because the original order transfer matrix is ​​as follows:

[0135]

[0136] Each row represents the percentage of orders destined for a given region. As can be seen, only 50% of orders from regions A and B go to regions A and B, while the remaining 50% go to region C. Of the orders in region C, 60% are destined for region C. Using first-come, first-served and batch matching methods, the actual order transfer matrix is ​​proportional to the original order transfer matrix. This can easily lead to a situation where capacity originally in regions A and B is matched with demand going to region C, resulting in capacity accumulation in region C and a shortage of capacity in regions A and B. However, the order filtering strategy implemented in this embodiment affects each subsequent specific match, increasing the probability of matching orders from regions A and B to regions A and B, while decreasing the probability of matching orders going to region C. In other words, the ride-hailing dispatch method provided in this embodiment effectively incorporates future supply and demand information, effectively filters orders based on their destination, thereby increasing the order matching volume, and also has a faster calculation speed.

[0137] This invention also provides a ride-hailing dispatch system, such as... Figure 10 As shown, the system includes:

[0138] The data layer is used to obtain order information and transportation capacity information; for details, please refer to the relevant description of step S101 in the above method embodiment.

[0139] The prediction layer is used to predict new order information and capacity information generated in multiple preset time domains in the future based on the acquired order information and capacity information; for details, please refer to the relevant description of step S102 in the above method embodiment.

[0140] The allocation layer is used to determine the number of orders to be matched in any region within a future preset time domain and destined for a second region, based on newly generated order information and capacity information, with the goal of maximizing the total order matching volume across multiple preset time domains in the future; for details, please refer to the relevant description of step S103 in the above method embodiment.

[0141] The execution layer is used to perform batch matching based on the number of orders destined for the second region matched in any region within a preset future time domain, with the goal of optimizing the average pick-up distance. For details, please refer to the relevant description of step S104 in the above method embodiment.

[0142] The ride-hailing dispatch system provided in this invention establishes an online real-time dispatch method that considers both transportation capacity and orders. It takes into account the available transportation capacity and the number of orders to be matched in each region over a future period, as well as the impact of matching on the future spatial distribution of transportation capacity. This effectively decouples and breaks down large-scale, complex dispatching problems, enabling solutions in a short time. This dispatch method ensures a better travel experience, increases the platform's total order matching volume, and has a certain degree of interpretability, making it easy to implement in real-world scenarios.

[0143] In one embodiment, such as Figure 11 As shown, this ride-hailing dispatch method includes a data layer, a data processing layer, a prediction layer, and a model layer. The model layer includes an allocation layer and an execution layer. The data layer is used to acquire historical data on capacity and orders, including the order creation time and latitude / longitude information, and the capacity creation time and latitude / longitude information. The data processing layer is used to divide the entire region into partitions based on the acquired information. Divide time into different time domains at equal intervals. The prediction layer predicts the capacity and demand data for each region over the next T time domains based on historical information and known features of the day. The allocation layer above the model layer predicts the order matching volume over the next T time domains. The execution layer below the model layer executes the control strategy for the k-th control time domain based on the results of the allocation layer. Batch matching is used for each matching in the execution layer. If there is remaining capacity after matching, it is used for empty-run scheduling.

[0144] The ride-hailing dispatch system provided in this invention utilizes a model-based predictive control framework, changing the current first-come-first-served and batch matching approach. This method only needs to construct a predictive model of newly added supply and newly generated demand in each region over the next T time domains using historical supply and demand data, and then input this model into the allocation layer model. The allocation layer considers future supply and demand information, matched order information, and the remaining supply and demand quantity, aiming to maximize the total order matching volume over the next T time domains. It determines the number of orders from matched orders in region i to travel to region j in the next time domain, and inputs this as the control strategy into the execution layer model. In each specific matching process in the execution layer, the capacity of each region is allocated to the demand traveling to each region according to the proportion of the number of orders to be matched. It does not require extensive real-time scenario judgment and data processing, thus placing low demands on processor speed and bandwidth.

[0145] For a detailed description of the functions of the ride-hailing dispatch system provided in this embodiment, please refer to the description of the ride-hailing dispatch method in the above embodiments.

[0146] This invention also provides a storage medium, such as... Figure 12As shown, a computer program 601 is stored on it. When executed by a processor, this program implements the steps of the ride-hailing dispatch method described in the above embodiments. The storage medium also stores audio and video stream data, feature frame data, interactive request signaling, encrypted data, and a preset data size. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0147] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0148] This invention also provides an electronic device, such as... Figure 13 As shown, the electronic device may include a processor 51 and a memory 52, wherein the processor 51 and the memory 52 may be connected via a bus or other means. Figure 13 Taking the example of a connection between China and Israel via a bus.

[0149] Processor 51 can be a central processing unit (CPU). Processor 51 can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0150] The memory 52, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the corresponding program instructions / modules in the embodiments of the present invention. The processor 51 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions, and modules stored in the memory 52, thereby implementing the ride-hailing dispatch method in the above method embodiments.

[0151] The memory 52 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor 51, etc. Furthermore, the memory 52 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 52 may optionally include memory remotely located relative to the processor 51, and these remote memories may be connected to the processor 51 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0152] The one or more modules are stored in the memory 52, and when executed by the processor 51, they perform the following: Figure 1 -9 shows the ride-hailing dispatch method in the embodiment.

[0153] For specific details regarding the aforementioned electronic devices, please refer to the relevant documentation. Figures 1 to 9 The relevant descriptions and effects in the illustrated embodiments are for understanding purposes only and will not be repeated here.

[0154] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for dispatching ride-hailing orders, characterized in that, include: Obtain order information and shipping capacity information; Based on the acquired order and capacity information, predict new order and capacity information generated in multiple preset time domains in the future; Based on newly generated order information and capacity information, a mixed-integer linear programming model is constructed and solved with the objective of maximizing the total order matching volume across multiple preset time domains in the future. This model determines the control strategy within the preset time domains, where the control strategy includes the number of matched orders destined for other regions in any region. The mixed-integer linear programming model is constructed based on capacity constraints and demand constraints. The capacity constraints include unmatched legacy capacity, newly generated capacity, and arriving capacity. The demand constraints include unmatched orders and newly generated orders. According to the control strategy, multiple batch matching is adopted to optimize the average pick-up distance. Before each batch matching, the target number to be completed is determined based on the number of orders in the control strategy and the matched orders. Different matching methods are selected based on the available capacity in the region, the number of orders to be matched, and the completion status of the control strategy.

2. The ride-hailing dispatching method according to claim 1, characterized in that, According to the control strategy, multiple batch matching is employed to optimize the average pick-up distance, including: Select a specific region, and based on the number of orders from that region to other regions matched within a future preset timeframe, perform multiple batch matching operations, with the goal of optimizing the average pick-up distance in each batch matching operation; select other regions, and perform multiple batch matching operations in each region until matching is completed for all regions; Before each batch matching, the target number to be completed is determined based on the number of orders to other regions matched in this region within a future preset time domain and the number of orders already matched in this region; the available capacity in this region and the size of the number of orders to be matched in other regions in this region are also determined. When the number of available shipping capacity is greater than or equal to the number of orders to be matched, match directly. When the number of available transport capacity is less than the number of orders to be matched, determine whether the number of orders to other regions matched in the current time domain for this region has been completed. When the number of orders to other regions matched in the current time domain for this region is completed, capacity and orders are matched with the goal of optimizing the average pick-up distance; If the number of orders from this region to other regions matched in the current time domain is not completed, the number of orders from this region to each region that should be matched in the next batch matching is determined based on the proportion of orders to be completed in each region. The number of orders from this region to each region that should be matched in the current time domain is used as a constraint to minimize the average pick-up distance for matching capacity and orders.

3. A ride-hailing dispatch system, characterized in that, include: The data layer is used to obtain order information and transportation capacity information; The prediction layer is used to predict new order and capacity information generated in multiple preset time domains in the future based on the acquired order and capacity information. The allocation layer is used to construct and solve a mixed-integer linear programming model based on newly generated order information and capacity information, with the goal of maximizing the total order matching volume across multiple preset time domains in the future. This model determines the control strategy within the preset time domains, where the control strategy includes the number of matched orders destined for other regions in any region. The mixed-integer linear programming model is constructed based on capacity constraints and demand constraints. Capacity constraints include unmatched legacy capacity, newly generated capacity, and arriving capacity. Demand constraints include unmatched orders and newly generated orders. The execution layer is used to perform multiple batch matching operations based on the control strategy, with the goal of optimizing the average pick-up distance. Before each batch matching operation, the target number to be completed is determined based on the number of orders in the control strategy and the already matched orders. Different matching methods are selected based on the available capacity in the region, the number of orders to be matched, and the completion status of the control strategy.

4. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the ride-hailing dispatch method as described in any one of claims 1-2.

5. An electronic device, characterized in that, include: The system includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the ride-hailing dispatch method as described in any one of claims 1-2.

Images