Spatial crowdsourcing online user recruitment method and device, medium and product

Abstract

The invention discloses an online user recruitment method and device for spatial crowdsourcing, a medium and a product, and relates to the technical field of spatial crowdsourcing, and the method comprises the steps: carrying out the revenue modeling of an online user recruitment problem for spatial crowdsourcing in an intelligent vehicle freight scene, and obtaining an intelligent vehicle freight model; acquiring a historical data set; the historical data set is used for recording order data of the intelligent vehicle in a historical time period; and performing order selection decision according to the intelligent vehicle freight model and the historical data set based on a prophet inequality mechanism of historical data sampling. According to the spatial crowdsourcing-based online user recruitment method and device, the medium and the product provided by the invention, the lower bound of revenue can be ensured, and the global fairness is also ensured.

Description

Technical Field

[0001] The present application relates to the field of space crowdsourcing technology, and in particular to a method, device, medium and product for recruiting online users for space crowdsourcing. Background Art

[0002] With the development of cutting-edge technologies such as smart cities and industrial Internet, spatial crowdsourcing has become an intelligent paradigm for completing spatial information-related tasks or projects. Spatial crowdsourcing uses mobile devices carried by a large number of users as basic perception units, collaborates through the network, submits or collects information about specific locations, such as pictures, comments, and ratings of places, or performs tasks at specific locations, such as intra-city delivery, and ultimately completes large-scale and complex urban perception tasks.

[0003] User recruitment is a core research problem in spatial crowdsourcing systems. It is necessary to select a group of appropriate perception units to complete spatial crowdsourcing tasks under certain constraints, such as cost constraints and distance constraints. Most of the current research on user recruitment problems models it as a constrained optimization problem. For example, each user will generate a certain value, and the spatial crowdsourcing platform must also pay a certain amount of compensation to the user. Then the user recruitment problem will be modeled as an optimization problem that maximizes the platform's revenue value under certain compensation cost constraints. However, this modeling method is often only applicable to offline scenarios, that is, all candidate user sets are known, and then the optimal solution is searched. This offline method will encounter challenges when facing online problems due to the lack of theoretical support. Because in online scenarios, the platform does not know the value of users who will come in the future. If it only spends its budget based on the current user set and ignores the higher-value users who may come in the future, it will only find a local optimal solution and cannot guarantee profitability. Some current user recruitment strategies for online scenarios often add an observation phase before the selection phase. For example, in the classic secretary problem, it is a well-known probability theory problem and decision theory problem: suppose an employer needs to select a secretary from n applicants. Applicants come for interviews one by one in order, and the employer can only decide whether to hire the applicant after the interview. Once a candidate is rejected, he / she cannot be selected again. The goal is to maximize the probability of selecting the best applicant (that is, the one with the strongest ability among all applicants). The secretary problem is a classic online scenario problem, that is, the challenge is that we don’t know the ability of future applicants, and we can only make judgments based on the applicants who have been interviewed. A classic solution to the problem is the "37% rule": reject the first n / e (e is the base of the natural logarithm) applicants (approximately the first 37%), this is the observation stage, and then select the first applicant who is better than all the previous ones. This strategy can maximize the probability of selecting the best applicant, which is about 37%. However, this strategy is unfair to candidates in the observation stage and cannot meet fairness. Summary of the invention

[0004] The purpose of this application is to provide an online user recruitment method, device, medium and product for spatial crowdsourcing, which can ensure the lower bound of profitability while ensuring global fairness.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides an online user recruitment method for space crowdsourcing, comprising:

[0007] In the scenario of smart truck cargo transportation, the profitability modeling of online user recruitment problem of spatial crowdsourcing is carried out to obtain the smart truck cargo transportation model;

[0008] Acquire a historical data set; the historical data set is used to record the order data of the smart car in the historical time period;

[0009] Based on the prophet inequality mechanism of historical data sampling, order selection decision is made according to the smart vehicle freight model and the historical data set.

[0010] Optionally, in the scenario of smart vehicle cargo transportation, the profitability modeling of online user recruitment problem of spatial crowdsourcing is performed to obtain a smart vehicle cargo transportation model, which specifically includes:

[0011] In the scenario of smart car cargo transportation, the online user recruitment problem of spatial crowdsourcing is defined as the online order selection problem of smart cars;

[0012] The smart car object, online delivery order and smart car revenue are modeled respectively to obtain the smart car cargo transportation model.

[0013] Optionally, the smart car object includes: the current location of the smart car, the current capacity of the smart car, the current speed of the smart car, the current route of the smart car and the set of orders currently accepted and not completed by the smart car; the online delivery order includes: the time of the order, the pickup point of the order and the delivery point of the order; the smart car revenue includes: the difference between the revenue of the order and the path distance cost.

[0014] Optionally, based on a prophet inequality mechanism of historical data sampling, an order selection decision is made according to the smart vehicle freight model and the historical data set, specifically including:

[0015] Determine the sample set capacity according to the historical data set; the sample set capacity is equal to the average number of orders within a set period of time in the historical data set;

[0016] Constructing a threshold sample set according to the sample set capacity and the historical data set; the threshold sample set includes the revenue increments corresponding to a set number of order data randomly sampled from the historical data set;

[0017] Using the shortest path planning algorithm, path planning is performed under the assumption that the target order is accepted to obtain candidate paths;

[0018] Using the intelligent vehicle freight model, the corresponding revenue increment is calculated according to the candidate path;

[0019] If the revenue increment corresponding to the candidate path is greater than the maximum revenue increment in the threshold sample set, the target order is accepted and the subsequent path is updated;

[0020] If the revenue increment corresponding to the candidate path is less than or equal to the maximum revenue increment value in the threshold sample set, the target order is rejected and the original path is maintained.

[0021] Optionally, a shortest path planning algorithm is used to perform path planning under the assumption that the target order is accepted, and a candidate path is obtained, specifically including:

[0022] Obtaining a current route set of the smart car, and adding the pickup point and delivery point of the target order to the current route set;

[0023] All possible paths in the current route set are traversed, and the path with the shortest distance that meets the set constraints is selected as a candidate path.

[0024] Optionally, the set constraints include: the pickup point of the order is before the delivery point of the order, the number of orders in the current route set is less than or equal to the current capacity of the smart car, and the revenue of the target order is greater than the additional cost added by the new route.

[0025] Optionally, the set number is equal to the sample set capacity minus 1.

[0026] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the online user recruitment method for spatial crowdsourcing.

[0027] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the online user recruitment method for spatial crowdsourcing.

[0028] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the online user recruitment method for spatial crowdsourcing.

[0029] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0030] The present application provides a method, device, medium and product for recruiting online users for spatial crowdsourcing, the method comprising: in a scenario of intelligent vehicle cargo transportation, modeling the profitability of online user recruitment for spatial crowdsourcing to obtain an intelligent vehicle cargo transportation model; obtaining a historical data set; the historical data set is used to record the order data of intelligent vehicles in historical time periods; based on a prophet inequality mechanism for historical data sampling, order selection decisions are made according to the intelligent vehicle cargo transportation model and the historical data set. In the scenario of intelligent vehicle cargo transportation, the present application models the specific spatiotemporal crowd intelligence perception task of online user recruitment for spatial crowdsourcing into an online instance selection problem, specifically an online order selection problem, and designs a prophet inequality mechanism based on historical data sampling, with the intelligent vehicle as the center. For each incoming online order, it can provide guidance with a guaranteed lower limit of expected benefits, helping the intelligent vehicle to determine whether to accept the real-time order, while ensuring a fair solution to the problem and ensuring the lower limit of the expected benefits of the subject. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 A flowchart of the online user recruitment method for spatial crowdsourcing provided for this application;

[0033] Figure 2 Technical details of the prophet inequality mechanism based on historical data sampling provided for this application;

[0034] Figure 3 A flowchart of online instance selection for the online user recruitment method for spatial crowdsourcing provided in this application. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0036] Current research on user recruitment in real-world scenarios has the following limitations:

[0037] 1. Online instance decision-making scenario of spatial crowdsourcing. In the spatial crowdsourcing of smart cities, researchers are faced with a real online scenario: some instances (orders, users, etc.) arrive in real time, and the spatial crowdsourcing platform needs to make a decision on whether to accept the instance at the same time as the instance arrives. This is an online instance decision-making scenario of spatial crowdsourcing. Among them, the intelligent cargo transportation task is a classic scenario. The spatial crowdsourcing platform will issue a series of cargo transportation orders. These cargo transportation orders are online attributes, that is, when the order is issued, the smart car must immediately make an irrevocable decision, that is, whether to accept the cargo transportation task. If the order is rejected, it cannot be recalled in the future because the order is likely to have been accepted by other smart cars. The cargo transportation order includes taking the goods from a certain location and then delivering them to the required location. The smart car will choose whether to accept the current task based on its current route planning and current space capacity. The goal is to reasonably select the cargo transportation task to maximize its own benefits.

[0038] 2. Guaranteeing the benefits of perception tasks. Spatial crowdsourcing tasks in smart cities are generally subject to cost constraints. Under these constraints, they need to make continuous and irrevocable decisions in the face of the continuous flow of online orders, while also ensuring a certain expected benefit. This is also a challenge for online delivery tasks.

[0039] 3. Fairness in user recruitment. Due to the online nature of freight orders and the non-recallability of decisions, a certain observation phase is often required to assist in making recruitment decisions. For example, in the online secretary mechanism, first in the observation phase, observe the first 1 / e part of the online orders, find the most cost-effective profit value as the selection threshold S', and then in the selection phase, select the first order with a profit value higher than S', so that the global optimal value, that is, the best order, can be selected with a probability of 1 / e. However, this selection mode will cause the first batch of orders that go online to be used as the observation group, and they will not be selected, which is unfair to them. The fairness issue is also a challenge for the online order selection problem.

[0040] Based on the above points, this application proposes an online user recruitment method, equipment, medium and product for spatial crowdsourcing, which reasonably solves the online order selection problem of spatial crowdsourcing in smart cities based on the prophet inequality mechanism of historical data sampling, which can guarantee the lower bound of profitability while also ensuring global fairness.

[0041] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0042] In an exemplary embodiment, the present application provides an online user recruitment method for space crowdsourcing, which is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. Figure 1 As shown, the method includes the following steps 1 to 3.

[0043] Step 1: In the scenario of smart vehicle cargo transportation, the profitability modeling of online user recruitment problem of spatial crowdsourcing is carried out to obtain the smart vehicle cargo transportation model.

[0044] Spatial crowdsourcing is a form of crowdsourcing that uses a large number of volunteers to collect and analyze geospatial data through mobile smart devices. This model is usually applied to tasks that require geographic location information, such as logistics and transportation, environmental monitoring, event reporting, or urban planning. In spatial crowdsourcing, task publishers publish geo-related tasks on the platform, and then volunteers or participants accept and complete these tasks based on their own locations. They usually use mobile devices such as smartphones to collect and upload data.

[0045] The scenario of smart car cargo transportation is used as an example of the online instance selection problem. In the scenario of smart car cargo transportation, the spatial crowdsourcing system can be used to optimize logistics, transportation and delivery services. Analogous to the secretary problem, the "employer" here usually refers to ordinary people or professionals who participate in the spatial crowdsourcing task. They can be drivers, delivery personnel or other volunteers who use their mobile devices and smart vehicles to participate in specific cargo transportation tasks. "Users" or "candidates" are defined as orders in this scenario. Each order will bring certain benefits to the employer, and the employer will also pay certain costs (time cost + energy cost). In this scenario, the smart car faces continuous incoming orders. According to its current status, it needs to reasonably formulate strategies to choose which orders to accept to maximize its own benefits.

[0046] Specifically, in the scenario of smart car transportation, the online user recruitment problem of spatial crowdsourcing is defined as the online order selection problem of smart cars. The smart car objects, online delivery orders and smart car revenues are modeled separately to obtain the smart car transportation model.

[0047] First, the smart car object is modeled, including: 1) the current location wl of the smart car, 2) the current capacity wc of the smart car, 3) the current speed wv of the smart car, 4) the current route wr of the smart car, and 5) the set of orders wo currently accepted and not completed by the smart car.

[0048] Secondly, the online delivery order is modeled, including: 1) the order time ot, 2) the order pickup point (starting point) o.ol, 3) the order delivery point (end point) o.dl. Through the above modeling, the optimization objective and related constraints can be obtained.

[0049] Finally, the revenue of the smart car is modeled. Usually, the cost of a delivery task is related to the distance. Every time a smart car accepts an order, the path will change, which will increase the distance and the cost accordingly. The total path distance of the task is expressed as: in represents the position of the i-th point in the intelligent vehicle path, represents the position of the i+1th point in the smart car path, and m' represents the total number of smart car path points. The revenue of the smart car is modeled as the difference between the order revenue and the path distance cost, that is, Where R j R represents the revenue brought by the sensing task part responsible for the jth order, D represents the path distance cost, and n' represents the total number of orders. j ∝distance(o j .dl,o j .ol) indicates R j It is directly proportional to the delivery distance of the order.

[0050] Step 2: Obtain a historical data set. The historical data set is used to record the order data of smart cars in a historical period of time.

[0051] Step 3: Based on the prophet inequality mechanism of historical data sampling, make order selection decisions according to the smart vehicle freight model and the historical data set.

[0052] The prophet inequality mechanism is a decision-making tool used in operations research and economics, especially in situations involving sequential decision making and future uncertainty. The mechanism is derived from the prophet inequality problem, which involves optimal decision making for a series of unknown values, where the decision maker wants to choose a value that is as close as possible to the expected maximum value (prophet value) among a set of random variables. The technical details of the prophet inequality mechanism are as follows Figure 2 As shown, firstly, a corresponding number of sample sets are collected from the historical data set according to the estimated number of online instances. Whenever an online instance arrives, its utility is compared with the maximum utility in the sample set. If it is greater than the maximum value, the instance is selected, otherwise, an element in the sample set is replaced immediately.

[0053] Based on the prophet inequality mechanism of historical data sampling, order selection decision is made according to the smart vehicle freight model and the historical data set, which specifically includes the following steps (1) to (6).

[0054] (1) Determine the sample set capacity according to the historical data set. The sample set capacity is equal to the average number of orders within a set time in the historical data set. Preferably, the set time is the time taken by the smart car to travel from the current path point to the next path point within the past week.

[0055] Assume that the current time is t 0 , current vehicle speed wv, current path point wr[0], if no new orders are accepted, the current route will be maintained, and the next path point is wr[1]. The time from the current path point wr[0] to the next path point wr[1] is Comprehensive statistical historical data in the past week 0 to 1 The average number of orders during a period of time is denoted as

[0056] (2) Constructing a threshold sample set according to the sample set capacity and the historical data set. The threshold sample set includes revenue increments corresponding to a set number of order data randomly sampled from the historical data set. Preferably, the set number is equal to the sample set capacity minus 1.

[0057] The sample set is constructed based on the historical data set. The historical data records the order status in the past period of time, including the order time ot, the order pickup point o.ol, and the order delivery point o.dl. Based on the historical order data, the current time t is estimated at the same time. 0 to 1 The number of orders that may be encountered Then randomly sample from the historical order set Then assume that we accept these orders. orders, perform shortest path planning, and calculate the incremental revenue P(c)+P(ok)-(D(wr(c+ok))-D(wr(c))), denoted as p k , for The number of the order, where P(c) represents the current accumulated revenue of the smart car, P(ok) represents the reward paid for order ok, D is the distance cost calculation function, wr(c) represents the current route set of the smart car, and wr(c+ok) represents adding order ok to the current route set of the smart car. k} is defined as the threshold sample set S.

[0058] (3) Using the shortest path planning algorithm, path planning is performed under the assumption that the target order is accepted to obtain candidate paths.

[0059] Specifically, obtain the current route set of the intelligent vehicle, and add the pick-up point and delivery point of the target order to the current route set; traverse all possible paths in the current route set, and select the path with the shortest distance that meets the set constraint conditions as the candidate path.

[0060] Among them, the set constraint conditions include: the pick-up point of the order is before the delivery point of the order, the number of orders in the current route set is less than or equal to the current capacity of the intelligent vehicle, and the revenue of the target order is greater than the additional cost increased by the new path.

[0061] The shortest path planning algorithm specifically includes: each intelligent vehicle will store a current route set w.r(c), which contains the path points to be passed in the future, that is, the pick-up points and delivery points of the orders that have not yet been reached. When making an order selection decision, add the target order o k .ol and o k .dl to the current route set w.r(c+o.k), where k is the serial number of the target order. Traverse all possible paths in the set w.r and select the shortest distance that meets the following constraints: 1) The pick-up point of the order must be before the delivery point of the order: for any element w.r[a]=o k .dl∈w.r in the set, which is the delivery point of a certain order, if there is a corresponding pick-up point w.r[b]=o k .ol∈w.r in the set, then b<a, where a and b represent the serial numbers of the element in the set w.r; 2) The number of orders in the set cannot exceed the capacity of the intelligent vehicle, that is, |w.r|≤w.c; 3) The revenue of the target order must be greater than the additional cost increased by the new path, that is, P(c)+P(o.k)>D(w.r(c+o.k))-D(w.r(c)), where P(c) represents the cumulative revenue of the current intelligent vehicle, P(o.k) represents the remuneration paid by the order o.k, and D is the distance cost calculation function.

[0062] (4) Adopt the intelligent vehicle cargo transportation model, and calculate the corresponding revenue increment according to the candidate path.

[0063] (5) If the revenue increment corresponding to the candidate path is greater than the maximum value of the revenue increment in the threshold sample set, accept the target order and update the subsequent path.

[0064] (6) If the revenue increment corresponding to the candidate path is less than or equal to the maximum value of the revenue increment in the threshold sample set, reject the target order and keep the original path.

[0065] Specifically, after constructing the threshold sample set, whenever the intelligent vehicle receives an online order o kFirst, we plan a new path and calculate the revenue increment p under the assumption that the order is accepted, and compare it with the maximum value p in the threshold sample set. max Compare. If p>p max , then accept the order and update the subsequent path; if p≤p max , then reject the order, keep the original path, and wait for new orders. This mechanism can guarantee the lower limit of the expected return of the subject, that is, Among them, OPT represents the optimal solution and E represents the expectation.

[0066] like Figure 3 As shown in the figure, this application models a specific spatiotemporal crowd intelligence perception task into an online instance selection problem. Taking the online order decision scenario of smart cars as an example, a prophet inequality mechanism based on historical data sampling is designed. First, a sample set is constructed from the historical data set, and then its utility benefit is calculated in reality and compared with the maximum value in the sample set. The sample set is dynamically updated until the decision is made, which ensures the lower limit of the expected benefit of the subject while ensuring fairness in solving the problem.

[0067] Specifically, assume an online selection scenario simulation, for example, a smart car encounters n orders and evaluates the value E(X) that is expected to be selected at the τth time when the distribution is unknown. τ ) is as close to the expected maximum as possible. In each round, n-1 samples are collected and the sample set of the τth round is defined as X τ =[X 1 ,X 2 ,…,X n-1 ]. According to the prophet inequality theory, if an online instance, that is, an order in a spatial crowdsourcing scenario, is selected, then it is hoped that the incremental revenue brought by the instance is greater than or equal to the expected maximum value E of the sample. max [X 1 ,X 2 ,…,X n-1 ]. When faced with online decision-making, the online instance will be compared with n-1 samples, and the largest one will be selected. Since the samples are randomly sampled and randomly replaced, in the smart car delivery scenario, these samples are the incremental revenue of the sampled historical orders to the current smart car, P(c)+P(ok)-(D(wr(c+ok))-D(wr(c))). This means that a maximum instance is selected from n instances, and the probability of an online instance being selected is If not selected (with probability ), then randomly replace one of the sample sets until the τth round of selection ends. In summary, the utility expectation of the instance selected in the τth round can be calculated as follows: For any ω<τ, The probability of an online instance being selected in τ rounds is: So the expected value of the selected instance and the sample expected maximum value satisfy:

[0068] The first equal sign is formed by the first n terms of the geometric sequence and The second approximation sign is derived from the limit equation It is deduced that the lower limit of the expected return of the subject that can be achieved through this application is about 0.632 times the expected maximum return.

[0069] In an exemplary embodiment, the present application further provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0070] In an exemplary embodiment, the present application further provides a computer-readable storage medium storing a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0071] In an exemplary embodiment, the present application further provides a computer program product, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0072] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant legal provisions.

[0073] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0074] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0075] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0076] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A spatial crowdsourcing online user recruitment method, characterized in that: include: In the scenario of smart truck cargo transportation, the profitability modeling of online user recruitment problem of spatial crowdsourcing is carried out to obtain the smart truck cargo transportation model; Acquire a historical data set; the historical data set is used to record the order data of the smart car in the historical time period; Based on the prophet inequality mechanism of historical data sampling, order selection decision is made according to the smart vehicle freight model and the historical data set.

2. The online user recruitment method for space crowdsourcing according to claim 1, characterized in that: In the scenario of smart truck cargo transportation, the profitability modeling of online user recruitment problem of spatial crowdsourcing is carried out to obtain the smart truck cargo transportation model, which specifically includes: In the scenario of smart car cargo transportation, the online user recruitment problem of spatial crowdsourcing is defined as the online order selection problem of smart cars; The smart car object, online delivery order and smart car revenue are modeled respectively to obtain the smart car cargo transportation model.

3. The online user recruitment method for space crowdsourcing according to claim 2, characterized in that: The smart car object includes: the current location of the smart car, the current capacity of the smart car, the current speed of the smart car, the current route of the smart car and the set of orders currently accepted and not completed by the smart car; the online delivery order includes: the time of the order, the pickup point of the order and the delivery point of the order; the smart car revenue includes: the difference between the revenue of the order and the path distance cost.

4. The online user recruitment method for space crowdsourcing according to claim 3, characterized in that: Based on the prophet inequality mechanism of historical data sampling, order selection decision is made according to the smart vehicle freight model and the historical data set, specifically including: Determine the sample set capacity according to the historical data set; the sample set capacity is equal to the average number of orders within a set period of time in the historical data set; Constructing a threshold sample set according to the sample set capacity and the historical data set; the threshold sample set includes the revenue increments corresponding to a set number of order data randomly sampled from the historical data set; Using the shortest path planning algorithm, path planning is performed under the assumption that the target order is accepted to obtain candidate paths; Using the intelligent vehicle freight model, the corresponding revenue increment is calculated according to the candidate path; If the revenue increment corresponding to the candidate path is greater than the maximum revenue increment in the threshold sample set, the target order is accepted and the subsequent path is updated; If the revenue increment corresponding to the candidate path is less than or equal to the maximum revenue increment value in the threshold sample set, the target order is rejected and the original path is maintained.

5. The online user recruitment method for space crowdsourcing according to claim 4, characterized in that: Using the shortest path planning algorithm, we perform path planning under the assumption that the target order is accepted, and obtain candidate paths, including: Obtaining a current route set of the smart car, and adding the pickup point and delivery point of the target order to the current route set; All possible paths in the current route set are traversed, and the path with the shortest distance that meets the set constraints is selected as a candidate path.

6. The online user recruitment method for space crowdsourcing according to claim 5, characterized in that: The set constraints include: the pickup point of the order is before the delivery point of the order, the number of orders in the current route set is less than or equal to the current capacity of the smart car, and the revenue of the target order is greater than the additional cost added by the new route.

7. The online user recruitment method for space crowdsourcing according to claim 6, characterized in that: The set number is equal to the sample set capacity minus 1.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the online user recruitment method for spatial crowdsourcing according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the online user recruitment method for spatial crowdsourcing described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the online user recruitment method for spatial crowdsourcing described in any one of claims 1 to 7 is implemented.

Images