Mobile crowdsourcing implicit dynamic excitation method and system based on anchoring effect

Abstract

The invention relates to the technical field of mobile crowdsourcing, in particular to a mobile crowdsourcing implicit dynamic excitation method and system based on an anchoring effect. The method comprises the following steps: constructing an LSTM task price prediction model, namely a TP-LSTM model; task price prediction and service quality prediction are carried out based on tasks by using a TP-LSTM model; performing task information anchoring based on a prediction result; and performing worker recruitment optimal incentive based on task information anchoring, wherein the worker recruitment optimal incentive comprises the optimal incentive of the initial recruitment process and the optimal incentive of the emergency recruitment process. The TP-LSTM model is utilized to significantly improve the accuracy of task price prediction, a time sequence mode in data can be identified and utilized, prediction errors are effectively reduced, and more reliable decision support is provided for a platform.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile crowdsourcing, and in particular to an implicit dynamic incentive method and system for mobile crowdsourcing based on anchoring effect. Background Art

[0002] Crowdsourcing platforms rely on the participation of a large number of workers and high-quality perception data, but in real scenarios, workers may be reluctant to participate in perception tasks. On the one hand, since participating in perception tasks consumes certain resources, participants will not actively participate unless they receive sufficient incentives to compensate for the cost of their investment. On the other hand, for some location-based tasks, the collected data may involve the user's location privacy information. Once this sensitive information is leaked, it will pose a hidden danger to the user's safety. Therefore, in order to stimulate workers' willingness to participate, the platform needs to provide effective incentives and appropriately compensate workers.

[0003] In mobile crowdsourcing, the sensing platform has clear requirements on the number of workers required for the task and the quality of data. The platform needs enough workers to complete the task and has strict standards for the quality of data submitted by the workers. However, there are two main problems in previous studies. First, existing studies often assume that the data quality provided by all workers is the same. This assumption is too idealistic and ignores the complexity of the problem. In fact, factors such as the ability, equipment performance, geographical location, and task completion attitude of different workers will significantly affect the data quality. Second, for the task requester, the goal is to achieve the maximum net utility, which is determined by the task cost and service quality. At present, most studies tend to fix the expected service quality and only adjust the task cost to select workers. This method fails to fully consider the changes in the expected service quality and the need for dynamic pricing during the phased recruitment process, and may have certain limitations in practical applications.

[0004] As a key concept in behavioral economics, the anchoring effect shows a significant difference from traditional economics. The anchoring effect can be applied to mobile crowdsourcing incentives. When posting a task, the platform can create a psychological anchor point by setting a lower initial bid for the task. Later, if the actual payment price of the task increases compared to the initial price, workers may still feel that this is a reasonable or even favorable remuneration because their reference point has been fixed on the initial lower price. This strategy can help the platform maintain worker engagement and satisfaction while controlling costs.

[0005] Although existing technologies have supported mobile crowdsourcing task recruitment to a certain extent, when faced with large-scale and long-term crowdsourcing, workers’ participation is often low and the cost of recruitment is uncontrollable. How to effectively improve worker participation and control task costs is still a challenge. In addition, the traditional explicit incentive mechanism has many drawbacks. It is necessary to continuously add and expand incentives to recruit self-purchased workers, which may lead to problems such as low worker satisfaction and uncontrollable task costs. Summary of the invention

[0006] In order to solve the above-mentioned problems, the present invention provides a mobile crowdsourcing implicit dynamic incentive method and system based on the anchoring effect. By introducing the anchoring effect in behavioral economics, the workers' psychological expectations of tasks are reduced through the anchoring effect, and then the task attributes are adjusted through a multi-stage recruitment method. The final strategy is determined through a bargaining game in each stage.

[0007] In a first aspect, the present invention provides a mobile crowdsourcing implicit dynamic incentive method based on the anchoring effect, which adopts the following technical solution: An implicit dynamic incentive method for mobile crowdsourcing based on anchoring effect, comprising: Get task information and worker information; Construct an LSTM task price prediction model, namely the TP-LSTM model; Use the TP-LSTM model to predict task price and service quality based on tasks; Based on the prediction results, task information is anchored; The optimal incentive for worker recruitment is determined based on task information anchoring, which includes the optimal incentive for the initial recruitment process and the optimal incentive for the emergency recruitment process.

[0008] Furthermore, the construction of the LSTM task price prediction model includes constructing an LSTM task price prediction model TP-LSTM with a combined multi-gating mechanism based on a long short-term memory network LSTM and a gating network. By combining multiple gating networks including a forgetting gate, an input gate, an output gate and a self-attention gate, data information is allowed to flow between units. At the same time, through the self-attention gating network, combined with the time series data of the output gate, different weights are fused according to the correlation at the previous moment to solve the problem of accuracy deviation caused by the input of multi-dimensional structured data.

[0009] Furthermore, the TP-LSTM model is used to predict the task price and service quality based on the task, including using the TP-LSTM model to obtain the benchmark price at the time of task release, and then adjusting the task benchmark price according to the task difficulty and task urgency, and obtaining the final predicted price after the adjustment is completed. , the adjustment formula is: in, is the weight of the task difficulty, is the weight of the task urgency, and .

[0010] Furthermore, the use of the TP-LSTM model to predict task prices and service quality based on tasks also includes calculating the maximum service quality that a worker can submit through the worker's reputation and ability value, calculating the benchmark service quality of the candidate worker set by performing statistical analysis on the maximum service quality value, and taking the benchmark service quality as a benchmark, considering the difficulty and urgency of the specific task, to obtain the predicted required service quality of the final task.

[0011] Furthermore, the task information is anchored based on the prediction results, including taking a dynamic method to set the anchoring, and calculating the final task price and required service quality by calculating the anchoring factor, which is expressed as: The specific formula is: in is the basic anchoring factor of the task price, ε is the basic anchoring factor of the required service quality, k is used to adjust the impact of task difficulty on the dynamic anchor point, and m is used to adjust the impact of task urgency on the dynamic anchor point.

[0012] Furthermore, the optimal incentive for the initial recruitment process includes adopting an infinite bargaining game under complete information, limiting the required service quality of the tasks in each stage to remain unchanged, and in the infinite bargaining game between the platform and the workers in each stage, the platform bids first, and the platform's discount factor is related to the task completion status. The task urgency is determined by the remaining perceived time of the task, and the task completion status is determined by the number of workers recruited at the current stage. The task urgency and completion status jointly affect the platform's discount factor. According to the discount factors of the platform and the workers, the utility functions of both parties are obtained in the mth round of the game; for the infinite bargaining game, the Nash equilibrium solution of the subgame is solved by the method of reverse induction of a finite game.

[0013] Furthermore, the optimal incentive for the emergency recruitment process includes starting the emergency recruitment process after the initial recruitment stage ends when the task price exceeds the predicted actual price and the required service quality is lower than the predicted task service quality, wherein during the emergency recruitment process, the platform increases the reward for the workers, and in each stage, the platform still bargains with the workers participating in the recruitment at each stage to determine the final remuneration of the workers; in the bargaining game during the emergency recruitment process, the workers The lowest acceptable price becomes the actual price predicted for the task , and the highest price the platform can accept In each stage of adjustment, the urgency of the task and the amount of task completion gradually increase; the discount factor of the platform and the worker remains and , the game interval is: .

[0014] In the second aspect, a mobile crowdsourcing implicit dynamic incentive system based on anchoring effect includes: The data acquisition module is configured to acquire task information and worker information; The model building module is configured to build an LSTM task price prediction model, namely, the TP-LSTM model; The prediction module is configured to use the TP-LSTM model to perform task price prediction and service quality prediction based on the task; The anchoring module is configured to anchor task information based on the prediction result; The incentive module is configured to perform optimal incentives for worker recruitment based on task information anchoring, including optimal incentives for the initial recruitment process and optimal incentives for the emergency recruitment process.

[0015] In a third aspect, the present invention provides a computer-readable storage medium storing a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor of a terminal device, for example, a mobile crowdsourcing implicit dynamic incentive method based on the anchoring effect.

[0016] In a fourth aspect, the present invention provides a terminal device, comprising a processor and a computer-readable storage medium, wherein the processor is used to implement various instructions; and the computer-readable storage medium is used to store multiple instructions, wherein the instructions are suitable for being loaded by the processor and executing the described method.

[0017] In summary, the present invention has the following beneficial technical effects: (1) Improving price prediction accuracy: The TP-LSTM model significantly improves the accuracy of task price prediction. It can identify and utilize time series patterns in data, effectively reduce prediction errors, and provide more reliable decision support for the platform.

[0018] (2) Improved worker satisfaction: The implicit incentive mechanism of the present invention is based on the anchoring effect. By appropriately adjusting the phased task information, workers can perceive that the platform has compromised on task price and service quality requirements, thereby increasing the workers' perceived benefits and significantly improving worker satisfaction.

[0019] (3) Increased system participation: Through a multi-stage dynamic adjustment mechanism, the present invention increases the attractiveness of tasks, helps the platform recruit more workers, and enhances the overall activity of the system.

[0020] (4) Effectively control task costs: The present invention introduces a bargaining game mechanism, which allows the platform to increase workers' perceived benefits while controlling service quality, increase workers' willingness to participate, speed up task recruitment, and reduce the urgency of tasks during the recruitment process. Costs can be effectively controlled at each recruitment stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a schematic diagram of the method architecture of Embodiment 1 of the present invention; Figure 2 is a comparison chart of the model prediction value and the true value of Example 1 of the present invention; Figure 3 is a comparison diagram of stage changes in worker satisfaction in Example 1 of the present invention; Figure 4 is a graph showing how the task allocation rate changes as the number of workers to be recruited for a task increases in Example 1 of the present invention; Figure 5 This is a bar chart comparing the average task price of Example 1 of the present invention as the number of workers required to be recruited for the task increases. DETAILED DESCRIPTION

[0022] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0023] Example 1: Reference Figure 1 , a mobile crowdsourcing implicit dynamic incentive method based on anchoring effect in this embodiment includes: Get task information and worker information; Construct an LSTM task price prediction model, namely the TP-LSTM model; Use the TP-LSTM model to predict task price and service quality based on tasks; Based on the prediction results, task information is anchored; The optimal incentive for worker recruitment is determined based on task information anchoring, which includes the optimal incentive for the initial recruitment process and the optimal incentive for the emergency recruitment process.

[0024] Specifically: like Figure 1 , the present invention demonstrates a system architecture of implicit dynamic incentive mechanism based on anchoring effect. The system firstly optimizes the time series data processing capability of LSTM model, significantly improving the prediction accuracy of task price fluctuations in mobile crowdsourcing platform. At the same time, the system combines multi-gated network and multi-layer attention mechanism to adapt to the complex application scenarios of mobile crowdsourcing. Figure 2The prediction performance of the TP-LSTM model proposed in the present invention on task prices in mobile crowdsourcing scenarios is demonstrated, showing extremely high accuracy. The present invention introduces dynamic anchoring of task information when tasks are released, such as appropriately reducing the initial price of tasks or improving service quality requirements, so as to form an anchor point for workers' psychological expectations. In the task recruitment stage, the recruitment process is subdivided into two processes: initial recruitment and emergency recruitment, and each process is divided into several stages. In each stage, the system dynamically adjusts the task price and service quality requirements according to the task recruitment progress and task urgency, ensuring that the regulation of service quality is not lower than the predicted level, so as to ensure the reliability of data quality. At the same time, through the bargaining game model of each stage, the present invention can achieve the best strategy between the platform and the workers. The algorithm of the present invention (AE-IDIM) is compared with two ablation experiment algorithms (Nogame algorithm and No anchoring algorithm), a random algorithm (random) and a multi-stage dynamic recruitment algorithm (IC-PS) on multiple indicators such as worker satisfaction, task completion rate and task cost. Figure 3 , Figure 4 and Figure 5 It demonstrates excellent performance in worker satisfaction, task allocation rate and task cost indicators respectively.

[0025] The method proposed by the present invention specifically comprises the following steps: S1: Task benchmark price prediction: Considering the relationship between task price and task release time, the present invention proposes an LSTM task price prediction model (TP-LSTM) with a combined multi-gating mechanism by combining a long short-term memory network LSTM and a gating network. By combining multiple gating networks including a forget gate, an input gate, an output gate and a self-attention gate, the flow of information between units is controlled, and the benchmark price of a task in a certain time period is predicted by learning historical data, and then the predicted task price is adjusted according to the attributes of each task such as difficulty and urgency.

[0026] The TP-LSTM model proposed in the present invention adds a self-attention gating network, combines the time series data of the output gate, and performs fusion of different weights according to the correlation degree of the previous moment. The problem of accuracy deviation caused by inputting multi-dimensional structured data is solved. The time series data can be selectively updated through the self-attention gating network to increase the robustness of the network model. This model can calculate the benchmark price of the task at this moment through the task release time.

[0027] S2: Task price prediction: After obtaining the benchmark price at the time of task release through the TP-LSTM model proposed in the present invention, the benchmark price of the task is adjusted according to the task difficulty and task urgency, and the final predicted price is obtained after the adjustment is completed. The adjustment formula is: in is the weight of the task difficulty, is the weight of the task urgency, and .

[0028] S3: Task service quality prediction: In mobile crowdsourcing, the task service quality is closely related to the reputation and ability of the worker, that is, the greater the reputation and ability of the worker, the better the task service quality he can provide. This paper calculates the maximum service quality that the worker can submit based on the reputation and ability of the worker. Down: To evaluate the average service quality of the candidate worker set of a task, the present invention adopts a benchmark service quality calculation method based on the maximum service quality of the candidate workers. Specifically, firstly, the maximum service quality value provided by each worker is selected from the candidate worker set of the task. These maximum values ​​reflect the highest service quality that each worker can provide under ideal circumstances.

[0029] Next, by performing statistical analysis on these maximum service quality values, the baseline service quality of the candidate worker set is calculated. . Benchmark Quality of Service Provide a reliable benchmark value for predicting the service quality of specific tasks. The calculation is as follows: in, The number of workers in the candidate worker set.

[0030] Get a baseline quality of service Finally, in order to predict the service quality of a specific task, the present invention takes the benchmark service quality as a benchmark, considers the difficulty and urgency of the specific task, and obtains the final task The prediction requires service quality. The specific calculation formula is as follows: S4: Task information anchoring: The present invention adopts a dynamic method to set the anchoring, and calculates the final task price and required service quality by calculating the anchoring factor. The adjustment factor of the anchoring effect is affected by the task difficulty and task urgency. The specific formula is: in It is the basic anchoring factor of the task price. It is the basic anchoring factor for service quality requirements, k is used to adjust the impact of task difficulty on the dynamic anchor point, and m is used to adjust the impact of task urgency on the dynamic anchor point.

[0031] The final task price anchored by the platform and service quality requirements From the following formula: .

[0032] S5. Optimal incentives for worker recruitment: The worker recruitment process can be divided into two main processes: the initial recruitment process and the emergency recruitment process. In each process, the task allocation will be divided into multiple stages, and the time of each stage is the same. In each stage, the platform will bargain with the participating workers to determine the final remuneration of the workers. After each stage, the platform will adjust the published task price and required service quality to attract workers to participate according to the increase in task urgency and the change in worker demand. The condition for workers to participate in recruitment is that their ability value exceeds the required service quality of the platform's current stage tasks.

[0033] S6. Initial recruitment process: In the initial recruitment process, due to the anchoring effect, the task price released by the platform is lower than the actual price, and the required service quality released by the platform is higher than the predicted required service quality. The present invention adopts an infinite bargaining game under complete information, and limits the required service quality of the task in each stage to be unchanged. In the game of the initial recruitment process, workers The psychological minimum price The price of the task, which represents the minimum price that the worker can accept for the task. The maximum price that the platform can accept is the predicted task price. The lowest price that workers can accept at the next stage and the highest price the platform can accept for: The bargaining process can be seen as the platform and the worker dividing the difference between the highest price of the task and the lowest price of the worker. The difference between the two is the game range: The platforms and workers are divided according to the proportion, assuming that the proportion of the platform is The proportion of people is Next, bargaining was carried out on the division, and finally a result acceptable to both the platform and the workers was reached.

[0034] In the infinite bargaining game between the platform and the workers at each stage, the platform makes the first bid. If the worker accepts, the game ends; if the worker refuses, the worker makes a bid This process is repeated until both parties accept it. In each round of the game, the passage of time affects the decision-making of both parties, because delaying the agreement will reduce the present value of future benefits. The extent to which the passage of time affects the decision-making of both parties is determined by their respective discount factors. Suppose the discount factors of the platform and the worker are and The discount factor reflects the patience of both parties in the game.

[0035] Among them, the worker's discount factor is related to the worker's credibility. , the platform's discount factor is related to the task completion status. The task urgency is determined by the remaining perceived time of the task, and the task completion status is determined by the number of workers recruited at the current stage. Task urgency and completion status jointly affect the platform's discount factor, giving it different levels of patience at different stages. Platform discount factor The specific calculation is: in, The remaining perceived time for the task, is the total perceived time of the task, is the number of workers recruited, and .

[0036] After obtaining the discount factors of the platform and workers, in the game Round, we can get the utility function of both parties as: Infinite bargaining games can be solved by refining them into Nash equilibrium subgames. Given that in an infinite bargaining game, a subgame starting from any round of bidding on the platform is equivalent to the entire game starting from the first round, the present invention adopts a finite-stage backward induction method to solve the Nash equilibrium solution of the subgame.

[0037] For the first round of the game wheel , the maximum available ratio of the platform No. Wheel In the The round profit ratio is , workers in The bids for the round must meet Therefore, the workers will accept The bid for round , the platform will accept it. round, the maximum proportion of workers available No. Wheel , the platform is in The round profit ratio is , the platform is in The bids for the round must meet Therefore, the platform The bid for round .

[0038] When the game tends to be balanced, the platform's quotes in odd rounds will tend to be stable. In the first round, the bid ratio given by the platform is the same as that of the first The situation is the same for the wheel, so we have: Finally, the maximum obtainable ratio of the platform in the infinite bargaining game is: For the first round of the game wheel , the minimum obtainable proportion of the platform have to: Therefore, in the infinite bargaining game, there exists a unique subgame Nash equilibrium solution: According to the Nash equilibrium solution of the infinite bargaining game, the best reward given by the platform to the workers in this game is: After each stage, the platform will adjust the task price and service quality requirements for the next stage based on the current task urgency and task completion status. Task urgency is determined by the perceived time remaining for the task, while task completion status is determined by the number of workers recruited in the current stage. The specific update formula is as follows: in, is the update coefficient.

[0039] S7: For some tasks that cannot recruit enough workers within the effective time in the initial recruitment phase, the present invention designs an emergency recruitment process to deal with the situation where the task urgency is too high and the task completion rate is low. The emergency recruitment process will be started after the initial recruitment phase to ensure that these tasks can be completed smoothly.

[0040] The present invention provides that once after stage adjustment, the task price exceeds the predicted actual price and the required service quality is lower than the predicted task service quality, the task will enter the emergency recruitment process. After entering the emergency recruitment process, the task price and required service quality will be updated to their predicted values. And during the stage adjustment in the emergency recruitment process, the platform will no longer further adjust the task price and service quality requirements, but will remain stable to ensure that the task can be completed quickly and meet the expected service quality.

[0041] During the emergency recruitment process, the platform will increase the rewards for workers. In each stage, the platform will still bargain with the workers participating in the recruitment to determine the final remuneration of the workers. In the bargaining game during the emergency recruitment process, the workers The lowest acceptable price becomes the actual price predicted for the task , and the highest price the platform can accept In each stage of adjustment, the urgency of the task and the amount of task completion gradually increase. The discount factor for the platform and the worker remains and , the game interval is: Same as the initial recruitment process, the platform accounts for , the proportion of workers is ,but , then bargain over the division, and finally reach a result acceptable to both the platform and the workers.

[0042] Similar to the infinite bargaining game stage of the initial recruitment process, the emergency recruitment process also has a unique subgame Nash equilibrium solution: According to the Nash equilibrium solution of the infinite bargaining game in the emergency recruitment process, the best reward given by the platform to the workers in this game is: Task recruitment ends when the task exceeds the deadline or the number of recruited workers reaches the requirement.

[0043] (1) This paper proposes a LSTM-based task price prediction model (TP-LSTM). This method builds a model that can estimate task prices by learning the task price rules in historical data sets. After the model training is completed, its parameters are uploaded to the mobile crowdsourcing platform. The platform evaluates and classifies different models based on the data processing capabilities and model effects of crowdsourcing workers. It provides pricing services based on accurate historical data analysis for data requesters.

[0044] (2) This invention introduces the anchoring effect in behavioral psychology and proposes an implicit incentive mechanism method. This method anchors the character attributes when the task is released, and by setting a lower price anchor point and a higher service quality anchor point without significantly increasing the price or reducing the expected service quality, workers can be effectively motivated through staged adjustments of task attributes.

[0045] (3) In view of the different time periods when workers participate in tasks, the present invention divides the task recruitment process into an initial recruitment process and an emergency recruitment process. Each process is divided into multiple stages. At each stage, the task price and service quality requirements are adjusted according to the task recruitment progress and task urgency. The adjustment of service quality will not be lower than the predicted service quality, ensuring the reliability of service quality. At each stage, the optimal strategy of the platform and the workers is obtained through bargaining games.

[0046] Embodiment 2: This embodiment provides a mobile crowdsourcing implicit dynamic incentive system based on anchoring effect, including: The data acquisition module is configured as follows: A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor of a terminal device, for a mobile crowdsourcing implicit dynamic incentive method based on an anchoring effect.

[0047] A terminal device includes a processor and a computer-readable storage medium, wherein the processor is used to implement various instructions; the computer-readable storage medium is used to store multiple instructions, wherein the instructions are suitable for being loaded and executed by the processor to implement the mobile crowdsourcing implicit dynamic incentive method based on the anchoring effect.

[0048] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A mobile crowdsourcing implicit dynamic incentive method based on anchoring effect, characterized in that: include: Get task information and worker information; Construct an LSTM task price prediction model, namely the TP-LSTM model; Use the TP-LSTM model to predict task price and service quality based on tasks; Based on the prediction results, task information is anchored; The optimal incentive for worker recruitment is determined based on task information anchoring, which includes the optimal incentive for the initial recruitment process and the optimal incentive for the emergency recruitment process.

2. According to claim 1, a mobile crowdsourcing implicit dynamic incentive method based on anchoring effect is characterized in that: The LSTM task price prediction model is constructed, including constructing an LSTM task price prediction model TP-LSTM with a combined multi-gating mechanism based on a long short-term memory network LSTM and a gating network. By combining multiple gating networks including a forgetting gate, an input gate, an output gate and a self-attention gate, data information is allowed to flow between units. At the same time, through the self-attention gating network, combined with the time series data of the output gate, different weights are fused according to the correlation at the previous moment, so as to solve the problem of accuracy deviation caused by inputting multi-dimensional structured data.

3. According to claim 2, a mobile crowdsourcing implicit dynamic incentive method based on anchoring effect is characterized in that: The TP-LSTM model is used to predict the task price and service quality based on the task, including using the TP-LSTM model to obtain the benchmark price at the time of task release, and then adjusting the task benchmark price according to the task difficulty and task urgency, and obtaining the final predicted price after the adjustment is completed. , the adjustment formula is: in, is the weight of the task difficulty, is the weight of the task urgency, and .

4. According to claim 3, a mobile crowdsourcing implicit dynamic incentive method based on anchoring effect is characterized in that: The method of using the TP-LSTM model to predict task prices and service quality based on tasks also includes calculating the maximum service quality that a worker can submit based on the worker's reputation and ability value, calculating the benchmark service quality of the candidate worker set by performing statistical analysis on the maximum service quality value, and taking the benchmark service quality as a benchmark, considering the difficulty and urgency of the specific task, to obtain the predicted required service quality of the final task.

5. According to claim 4, a mobile crowdsourcing implicit dynamic incentive method based on anchoring effect is characterized in that: Based on the prediction results, task information is anchored, including taking a dynamic method to set the anchoring, and calculating the final task price and required service quality by calculating the anchoring factor, which is expressed as: The specific formula is: in is the basic anchoring factor of the task price, ε is the basic anchoring factor of the required service quality, k is used to adjust the impact of task difficulty on the dynamic anchor point, and m is used to adjust the impact of task urgency on the dynamic anchor point.

6. The method for implicit dynamic incentive of mobile crowdsourcing based on anchoring effect according to claim 5, characterized in that: The optimal incentive for the initial recruitment process includes adopting an infinite bargaining game under complete information, limiting the required service quality of the tasks in each stage to remain unchanged. In the infinite bargaining game between the platform and the workers in each stage, the platform bids first, and the platform's discount factor is related to the task completion status. The task urgency is determined by the remaining perceived time of the task, and the task completion status is determined by the number of workers recruited at the current stage. The task urgency and completion status jointly affect the platform's discount factor. According to the discount factors of the platform and the workers, the utility functions of both parties are obtained in the mth round of the game. For the infinite bargaining game, the Nash equilibrium solution of the subgame is solved by the method of reverse induction of the finite game.

7. The method for implicit dynamic incentive of mobile crowdsourcing based on anchoring effect according to claim 6 is characterized in that: The optimal incentive of the emergency recruitment process includes starting the emergency recruitment process after the initial recruitment stage ends when the task price exceeds the predicted actual price and the required service quality is lower than the predicted task service quality. In the emergency recruitment process, the platform increases the reward to the workers. In each stage, the platform still bargains with the workers participating in the recruitment to determine the final remuneration of the workers. In the bargaining game during the emergency recruitment process, the workers The lowest acceptable price becomes the actual price predicted for the task , and the highest price the platform can accept In each stage of adjustment, the urgency of the task and the amount of task completion gradually increase; the discount factor of the platform and the worker remains and , the game interval is: 。 8. A mobile crowdsourcing implicit dynamic incentive system based on anchoring effect, characterized in that: include: The data acquisition module is configured to acquire task information and worker information; The model building module is configured to build an LSTM task price prediction model, namely, the TP-LSTM model; The prediction module is configured to use the TP-LSTM model to perform task price prediction and service quality prediction based on the task; The anchoring module is configured to anchor task information based on the prediction result; The incentive module is configured to perform optimal incentives for worker recruitment based on task information anchoring, including optimal incentives for the initial recruitment process and optimal incentives for the emergency recruitment process.

9. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the method according to claim 1 .

10. A terminal device, comprising a processor and a computer-readable storage medium, wherein the processor is used to implement each instruction; and the computer-readable storage medium is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the method as claimed in claim 1 .

Images