Group Intelligence Perception Task Allocation Method, System and Device Based on Weiner Attribution

Abstract

The present invention discloses a method, system and device for crowd-sensing task allocation based on Weiner attribution. The method includes determining the attribution bias factor of participants; obtaining the optimal weight matrix for participants according to the effect of the attribution bias factor, and determining the optimal task completion order of participants according to the magnitudes of the task weights in the optimal weight matrix; obtaining the reward uncertainty preference weight according to the preference degree of participants under the interaction of rewards and uncertainty; calculating the optimal task completion quantity for participants by combining the optimal weight matrix and the reward uncertainty preference weight; determining the final reward distribution, and after evaluating and weighing the expected completion rate and the expected reward, determining the reward compensation factor for each region, and the determination of the reward compensation factor is based on the compensation weights of each region. This solution can improve the coverage rate of task completion without increasing the rewards of tasks, and improve the overall completion quality of crowd-sensing tasks.

Description

Technical Field

[0001] The present invention relates to the technical field of crowd sensing data acquisition, and in particular to a crowd sensing task allocation method, system and device based on Weiner attribution. Background Technique

[0002] Crowd sensing is a new mode that combines the idea of crowdsourcing and the sensing ability of mobile devices. It can form a sensing network through existing mobile devices (including mobile phones, tablets, etc.), and can collect sensing data without laying a large number of sensing devices over a large area.

[0003] The effectiveness of crowd sensing services and the quality of the final completion are affected by the regional coverage rate of task completion, where the coverage rate is the ratio of the number of tasks completed in each region to the number of task requirements. For example, real-time environmental data across the city is helpful for urban planning and construction. However, to achieve the planning goal, the required data set needs to cover the entire city. Data with a low coverage rate will lead to one-sided decision-making in the city and have adverse consequences. Therefore, improving the coverage rate is an urgent problem to be solved in crowd sensing. Due to the uneven population distribution in most regions and the fact that users are accustomed to moving in their resident regions, there are fewer participants in tasks in remote areas with sparse populations, resulting in a very low coverage rate in these regions. Therefore, designing an effective task allocation method and changing the task selection behavior of participants accordingly so that the distribution of participants is consistent with the task distribution can improve the final task coverage rate.

[0004] Existing crowd-sensing task allocation methods aiming to improve task coverage can be divided into two categories: monetary allocation methods and non-monetary allocation methods. Monetary task allocation methods usually refer to the platform using means such as rewards to change the decision-making behavior of participants in choosing tasks, and thus achieving the platform's goals. Non-monetary task allocation methods refer to the platform using other means than monetary rewards to change the decision-making behavior of participants, and thus achieving the platform's goals. However, the current task allocation methods do not consider the impact of uncertainty on participants' decisions. Uncertainty means that the decision-maker cannot know for sure the distribution range and state of future gains and losses. Currently, the vast majority of studies assume that the costs and benefits involved in the decision-making of participants are determinable when designing task allocation methods. Then, these studies calculate the final benefits of each plan based on these determined values, and thus determine the decision-making plan that maximizes the benefits. However, when participants make decisions using rewards and costs, if there is uncertainty, the uncertainty will affect the decision-making behavior of participants as an implicit cost, causing their decisions to deviate from the optimal decision. Therefore, the decision-making assumption based on determined costs and benefits will cause the final effect of the task allocation method to deviate from the expected effect. And when uncertainty is considered in the behavioral decision-making of nodes, the platform's requirements can be more efficiently met. In addition to the above-mentioned majority of studies on task allocation methods that do not consider uncertainty, there are currently very few studies that discuss task allocation methods based on probability rather than on determined costs and benefits. But these studies assume that participants can completely evaluate the expected probabilities and expected benefits of all choices and make decisions based on this. However, the probability weighting theory points out that in the presence of uncertainty, relative weights need to be assigned to each option when discussing participants' decisions. This is because different participants will have different choice probabilities for the uncertainty caused by probability. Under the influence of uncertainty, even when faced with two tasks with the same reward, different participants will have different choice probabilities for these two tasks. Research shows that if corresponding probability weights are assigned to each decision for different uncertainties, participants' decision-making behaviors can be more accurately explained. And it is found that there is uncertainty in crowd-sensing. Summary of the Invention

[0005] Aiming at the deficiencies of the above-mentioned existing technologies, the present invention provides a crowd-sensing task allocation method, system and device based on Weiner attribution. This solution quantifies the impact of tasks in different regions on participants' task selection decisions using uncertainty. First, through the attribution bias factor of participants, the role of uncertainty is weakened. Then, according to the Weiner attribution theory, a reward payment algorithm is designed to prompt the task distribution finally selected by participants to be consistent with the requirements, so as to improve the final task completion coverage rate.

[0006] In a first aspect, a crowd-sensing task allocation method based on Weiner attribution is provided, including:

[0007] Determine the attribution bias factor of the participant, calculate the corresponding attribution bias factor according to the relevant parameters of the participant's attribution bias factor, and update the attribution bias factor according to the completion situation of the participant at the end of each task;

[0008] Determine the task completion order, evaluate the evaluation weight of each task for the participant according to the effect of the attribution bias factor, obtain the optimal weight matrix for the participant, and determine the optimal task completion order of the participant according to the size of the task weights in the optimal weight matrix;

[0009] Determine the reward uncertainty preference weight, obtain the reward uncertainty preference weight according to the preference degree of the participant under the interaction of reward and uncertainty, and use the reward uncertainty preference weight as one of the bases for task selection;

[0010] Determine the optimal task completion quantity, combine the optimal weight matrix and the reward uncertainty preference weight, and calculate the optimal task completion quantity for the participant according to the optimal selectable items and related preferences of the participant;

[0011] Determine the final reward distribution. After evaluating and weighing the expected completion rate and the expected reward, determine the reward compensation factor for each region. The determination of the reward compensation factor is based on the compensation weight of each region; the principle for determining the compensation weight is to make the task demand distribution and the participant distribution in each region more consistent.

[0012] Essentially, this solution classifies the tasks in different regions and measures them by uncertainty. On this basis, an attribution bias factor is introduced, and the attribution bias factor of the participant is continuously updated through the compensation reward factor, making it more inclined to tasks with higher uncertainty. And the reward uncertainty preference weight of the participant is introduced to determine the optimal task selection set of the participant. This improves the overall task selection coverage rate. At the same time, compared with the existing mechanism (DA-based mechanism, DA-based is a traditional location-based task allocation method in crowd sensing, taking the profit that can be obtained as the consideration for the participant's behavior decision), this solution can effectively change the tasks selected by the participant without a substantial increase in the reward, enabling the participant to go to remote areas with greater uncertainty and making the final task coverage more uniform.

[0013] Furthermore, the attribution bias factor of the participant is one of the main influencing factors for the participant to select the task order in the optimal situation. When designing the attribution bias factor, Weiner attribution is affected by past completion experience, that is, participant g iThe update of the attribution bias factor for the t-th task performance is determined by the task completion of the (t - 1)-th task performance and the tasks before the (t - 1)-th task performance.

[0014] The definition formula of the attribution bias factor of the participant is:

[0015]

[0016] where, represents the participant g i the attribution bias factor for the t-th task performance; α, β, γ ∈ (0, 1) and α + β + γ = 1, and α, β, γ respectively represent the benefits of the previous task performance the attribution bias factor of the previous task performance the upper limit of the attribution bias factor the three influence weights on represents the lower limit of the attribution bias factor;

[0017] where, the upper limit of the attribution bias factor is calculated according to the following formula:

[0018]

[0019] where, R i represents the benefit of each task selection set, and the task selection set is composed of a weighted combination of any number of tasks; B represents the set of basic rewards for all tasks, that is M is the number of tasks; V is the covariance set between two tasks, that is V = (Vj 1j2 ) j1,j2 (j1, j2 ∈ (0, M]); represents the lower limit of the uncertainty of the task selection set;

[0020] where R i is calculated according to the following formula:

[0021]

[0022] where, represents the basic reward, w m represents the weight of the task in the task selection set, and M represents the total number of tasks.

[0023] Furthermore, in order to determine the optimal selection order of tasks, it is necessary to define task weights to evaluate the priority of task completion, and combine the attribution bias factor of the participant to calculate the optimal task weight. According to the attribution bias factor calculate the optimal weight matrix of the participant when performing the task for the t-th time and is expressed as follows:

[0024]

[0025] is the optimal weight matrix and is the weight matrix Q i a specific matrix among them, where where w m represents the weight of each selectable real task. The optimization principles that the calculation of the optimal weight matrix needs to satisfy are: under the same uncertainty, the weight matrix with the largest reward needs to be selected; the weight matrix with the largest difference between the reward and the attribution deviation factor needs to be selected. These two principles are expressed as follows:

[0026]

[0027] where represents the uncertainty of the task selection set. Since the essence of task uncertainty is the standard deviation, and because there is a correlation between tasks pairwise, this makes the uncertainty of the task selection set need to be obtained by adding pairwise combinations of tasks, which can be expressed by the following formula:

[0028]

[0029] where represents the correlation between task m1 and task m2, represents participant g i for the uncertainty of task m1, represents participant g i for the uncertainty of task m2, respectively represent the weights of task m1 and task m2 in the task selection set.

[0030] Furthermore, the optimal weight matrix can obtain the optimal task selection order of the participant. However, the optimal weight matrix contains the weights of all tasks, and it is difficult for the participant to select and complete them simultaneously. Therefore, it is necessary to determine, according to the participant's own preferences regarding the reward R i and uncertainty , the optimal number of task selections for participant g i under the action of a specific attribution deviation factor

[0031] The greater the benefit, the greater the probability that the participant will spend more time on the task, so the participant's preference is positively correlated with the benefit; while the greater the uncertainty, the greater the probability that the participant may lose, so the preference is negatively correlated with uncertainty. Therefore, the preference weight of reward uncertainty U i It is expressed as follows:

[0032]

[0033] Among them, τ i Indicates that participants sensitivity; Indicates participants' uncertainty The preference coefficient of the above formula is simple. It can be seen that the benefit is proportional to the preference, while the uncertainty and preference decrease nonlinearly. The nonlinearity is the participant's response to the preference of uncertainty. Uncertainty preference weight U i With R i and The relationship satisfies the concave function characteristics.

[0034] Furthermore, since the weights in the weight matrix are So a weight matrix can be understood as a task. When the participants determine the optimal weight matrix and its corresponding uncertainty When the participants are uncertain, When the participants determine the uncertainty of the task set that should be selected when its value is optimal Later, the optimal number of tasks for participants can be determined. and reward uncertainty preference weight U i Under the joint action of It is expressed as follows:

[0035]

[0036] Furthermore, the number of task choices that optimize the participants' preferences is calculated as:

[0037]

[0038] Among them, ξ i For participants g i The number of tasks that can be completed under normal work intensity, represents the optimal weight matrix The corresponding uncertainty can be expressed by the following formula:

[0039]

[0040] According to the definition of the attribution bias factor, it can be known that represents the remuneration that can be earned in other fields if the participant does not participate in the platform tasks. Therefore, the prerequisite for the participant to participate in the platform tasks is always holds, so always holds. Therefore, the final attribution bias factor and the optimal weight matrix The corresponding uncertainty The relationship between them is as follows expression.

[0041]

[0042] Furthermore, the remuneration received by the participant is divided into two parts in total. The first part is the basic remuneration which is the average value of the real cost of the participant to complete the task That is The second part is the additional reward given by the platform to the participant Among them, the basic remuneration can be understood as the task volume of the task, and θ is the unit additional reward factor of the task volume.

[0043] When the participant selects tasks, they will accept multiple tasks at the same time. Since the real cost of the task is unknown, only after the participant completes a task will they know its real cost. Therefore, if the task completed by the participant is not the last one, and the cumulative real cost after completing this task > the remuneration, the participant will not continue to complete the subsequent tasks. The cumulative income of the participant after completing the t-th task is as follows formula:[[]]END]]

[0044]

[0045] Among them, p represents the probability that the participant obtains a positive income after participating in the task. Increasing the additional reward factor θ can increase the number of tasks completed, but it will increase the remuneration paid by the platform.

[0046] Further, sort the attribution bias factor update factor (α + γ) provided by the participants. The larger (α + γ) is, the greater the probability that the participants will go to remote areas and the greater the effect of the reward. Therefore, the participants are ranked in ascending order of (α + γ) as candidates for the regional set. Since the increase in the participants' attribution bias factor is a continuous process, during the process where the initial task completion location of the participants tends to be in remote areas, the number of tasks completed at the initial task completion location is much more than the number of tasks completed in remote areas. Therefore, after evaluating and weighing the relationship between the expected payment reward and the expected number of tasks completed, in order to make up for the imbalance in the number of tasks completed between regions, it is necessary to assign different compensation weights to different regions. The formula for the compensation weight is defined by the following formula:

[0047]

[0048] where w d represents the compensation weight of the d-th region, λ represents the weight adjustment factor, D represents the number of regions divided, and d represents the d-th region. All regions are equally divided into D regions according to the degree of remoteness to form a set

[0049] Second, a crowd-sensing task allocation system based on Weiner attribution is provided, including:

[0050] An attribution bias factor determination module for participants, which is used to calculate the corresponding attribution bias factor according to the relevant parameters of the participants' attribution bias factor, and update the attribution bias factor according to the completion situation of the participants at the end of each task;

[0051] A task completion order determination module, which is used to evaluate the evaluation weight of each task for the participants according to the effect of the attribution bias factor, obtain the optimal weight matrix for the participants, and determine the optimal task completion order for the participants according to the magnitudes of the weights of each task in the optimal weight matrix;

[0052] A reward uncertainty preference weight determination module, which is used to obtain the reward uncertainty preference weight according to the preference degree of the participants under the interaction of the reward and uncertainty;

[0053] An optimal task completion quantity determination module, which is used to calculate the optimal task completion quantity for the participants by combining the optimal weight matrix and the reward uncertainty preference weight according to the optimal options and relevant preferences of the participants;

[0054] A final reward allocation determination module, which is used to determine the reward compensation factor for each region after evaluating and weighing the expected completion rate and the expected reward, and the determination of the reward compensation factor is based on the compensation weights of each region.

[0055] In a third aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the above-described method for crowd sensing task allocation based on Weiner attribution is implemented.

[0056] In a fourth aspect, a crowd sensing task allocation device based on Weiner attribution is provided, including a sensor, a memory, and a processor. The sensor is used to collect crowd sensing task data, the memory is used to store the computer program and the collected crowd sensing task data, and the processor is used to execute the computer program to implement the above-described method for crowd sensing task allocation based on Weiner attribution.

[0057] Beneficial effects

[0058] The present invention proposes a method, system and device for crowd sensing task allocation based on Weiner attribution. Compared with the existing technologies, the existing task allocation methods are designed based on the objective probability assumptions of traditional decision models, while the present invention takes into account the influence of Weiner attribution theory on participants, and establishes an environment that can affect the task selection of crowd sensing participants. By using uncertainty to measure the historical frequency of task completion in the environment, the influence of uncertainty on participants' decisions is considered, and by designing an attribution bias factor to change the preference factor of participants for uncertainty, the optimal decisions of participants are gradually inclined to remote areas where tasks are less completed. During the task completion process, the attribution bias factor of participants is continuously cultivated and regulated through rewards, so that the distribution of participants is more consistent with the task demand distribution. And considering the preference trade-off between participants for rewards and uncertainty, the optimal number of task completions is determined, improving the final task completion coverage rate. At the same time, compared with the existing mechanisms, this solution can improve the overall coverage rate of tasks on the premise of maximizing the utility of participants without paying more rewards to participants, and improves the quality of crowd sensing task completion. Description of the drawings

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0060] Figure 1 It is a schematic diagram of the crowd sensing task allocation process provided by an embodiment of the present invention; where (a) is a schematic diagram of the process without a mechanism, and (b) is a schematic diagram of the process under the mechanism of the present invention;

[0061] Figure 2 It is the flowchart of the crowd-sensing task allocation method based on Weiner attribution provided by the embodiment of the present invention;

[0062] Figure 3 In (a), it is the distribution map of the completion positions of crowd-sensing tasks without using the allocation method of the present invention, and in (b), it is the distribution map of the completion positions of crowd-sensing tasks using the allocation method of the present invention under the same data;

[0063] Figure 4 In (a), it is the schematic diagram of the distribution of the completion positions of crowd-sensing tasks using the DA-base mechanism, and in (b), it is the schematic diagram of the distribution of the completion positions of crowd-sensing tasks using the allocation method of the present invention under the same data. Detailed implementation manners

[0064] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other implementation manners obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope protected by the present invention.

[0065] Embodiment 1

[0066] As Figure 2 shown, this embodiment provides a crowd-sensing task allocation method based on Weiner attribution, including the following steps:

[0067] Step 1: Establish the crowd-sensing system environment

[0068] As Figure 1 shown, the crowd-sensing task allocation mechanism includes three parties: the task requester, the platform, and the participants G = {g1, g2,..., g i ,..., g N}, where g i represents the i-th participant. The platform will publish M tasks Q = [q1, q2,..., q j ,..., q M , where q j represents the j-th task. When the platform publishes tasks, it will also publish some information about the tasks, including: the offer price of each task the uncertainty of each task evaluated by the platform the reward given by the platform to the participants is where is the additional reward, is the basic reward, is the real cost of completing the task.

[0069] Perform a simulation experiment on the present invention. The specific parameters are shown in Table 1. Set the uncertainty sensitivity factor τ (i.e., the sensitivity of the participant to uncertainty) to 2, and the uncertainty preference coefficient is set to 3. The uncertainty of the task is set to a uniform distribution of (1, 3), and the true cost C is set to a logarithmic distribution with a mean of 1 and a variance of 2. In the system, the number of participants is 20 and the number of tasks is 30.

[0070] Table 1 Simulation parameters

[0071]

[0072] Step 2: Define the revenue of the task selection set

[0073] Set the relationship between the true cost and the uncertainty , and the definition formula of the cost is as follows:

[0074]

[0075] Among them, and are the parameters in the logarithmic distribution of the cost.

[0076] The weights of all tasks constitute the weight matrix where w m represents the weight of each selectable true task. And define different weight matrices Q i corresponding to a task selection set, and the task selection set is a set formed by weighting all tasks according to the weight matrix.

[0077] According to the definition of the task selection set, calculate the uncertainty and the revenue R i of the task selection set. The formula is as follows:

[0078]

[0079] Among them, represents the correlation between task m1 and task m2, represents the uncertainty of participant g i for task m1, represents the uncertainty of participant g i for task m2, respectively represent the weights of task m1 and task m2 in the task selection set. The calculation formula of

[0080]

[0081] Among them, represents the mean value of the cost of task m1, represents the mean value of the cost of task m2, represents the standard deviation of the cost of task m1, represents the standard deviation of the cost of task m2, and E[] represents the calculation of the mean value.

[0082] Step 3: Design the attribution bias factor

[0083] First, calculate the upper and lower limits of the uncertainty of the task selection set and then define the expression of the attribution bias factor within this range. Through the following inequality, the range satisfied by the uncertainty is:

[0084]

[0085] Among them, λ m is the eigenvalue of the covariance matrix of the lower limit of the uncertainty of the task, the upper limit of the uncertainty of the task. After finding the upper and lower limits of the uncertainty of the task selection set , next, as long as the relationship between the attribution bias factor and is found, the upper and lower limits of the attribution bias factor can be known. Through the following formula, the relationship between the attribution bias factor and the uncertainty corresponding to the optimal weight matrix can be obtained:

[0086]

[0087] Next, substitute the upper and lower limits of the uncertainty obtained above into the relationship between the attribution bias factor and the uncertainty corresponding to the optimal weight matrix , and the upper and lower limits of the following participant attribution bias factor can be obtained:

[0088]

[0089] Among them, represents the upper limit of the uncertainty of the task selection set, Represents the lower limit of the uncertainty of the task selection set. By determining the upper and lower limits of the attribution bias factor, the attribution bias factor of the participant when performing the task for the t-th time can be updated according to the definition formula of the attribution bias factor:

[0090]

[0091] where α, β, γ ∈ (0, 1) and α + β + γ = 1, and α, β, and γ respectively represent the benefits of the previous task performance The attribution bias factor of the previous task performance The upper limit of the attribution bias factor The three of them The influence weights on Represents the lower limit of the attribution bias factor.

[0092] Step 4: Determine the optimal task selection set

[0093] Design the optimal weight matrix The conditions to be satisfied:

[0094]

[0095] By solving this optimal linear programming problem, the optimal weight matrix can be obtained

[0096]

[0097] where B represents the set of the basic rewards of all tasks, that is V is the covariance set between two tasks, that is V = (V j1j2 ) j1,j2 (j1, j2 ∈ (0, M]). According to the magnitudes of the weights of each task in the optimal weight matrix, determine the optimal task completion order of the participant.

[0098] Step 5: Determine the optimal number of selected tasks

[0099] Calculate the reward uncertainty preference weight of the participant according to the following formula:

[0100]

[0101] where τ i Represents the sensitivity of the participant to ; Represents the preference coefficient of the participant for the uncertainty .

[0102] According to the uncertainty preference weight U of the participant i and R i and The relationship, calculating the conditions for the participant to obtain the maximum preference:

[0103]

[0104] Next, according to and The relationship between them, determine that the number of task selections that make the participant's preference optimal is:

[0105]

[0106] where ξi is the number of tasks that participant g i can complete under normal working intensity.

[0107] Step 6: Evaluate the participant's task completion

[0108] The total remuneration received by the participant is divided into two parts. The first part is the basic remuneration which is the average value of the actual cost of the participant to complete the task that is The second part is the additional reward given by the platform to the participant Among them, the basic remuneration can be understood as the task volume of the task, and θ is the unit additional reward factor of the task volume. Therefore, the cumulative income of the participant after completing the t-th task is calculated as follows:

[0109]

[0110] By simplification, use the following formula to replace as follows:

[0111]

[0112] where p represents the probability that the participant obtains a positive income after participating in the task;

[0113] The cumulative income of the participant can be calculated using the following formula:

[0114]

[0115] Step 7: Determine the payment of remuneration

[0116] Divide all regions into D equal regions according to the degree of remoteness, forming a set Within each region, the number of tasks to be completed is N = {∈1, ∈2, …, ∈ D}。Sort according to the Weiner attribution update factor (α + γ) provided by the participants. The larger (α + γ) is, the greater the probability that the participant will go to a more remote area. The participants are sequentially used as candidates for the regional set in ascending order of (α + γ). Therefore, after evaluating and weighing the relationship between the expected payment reward and the expected number of tasks completed, in order to make up for the imbalance in the number of tasks completed between regions, different compensation weights need to be assigned to different regions. The formula for the compensation weight is defined by the following formula:

[0117]

[0118] Among them, λ is the weight adjustment factor, and d represents the d-th region. Under the action of this weight coefficient, the target number p of each region can be calculated d as:

[0119]

[0120] Step 8: Evaluate the platform target

[0121] The number of tasks completed by the final participants should be the sum of the numbers of all possible results. The formula is:

[0122]

[0123] Among them, f(n) is the number of possible results at each layer, and len(Φ(k)) represents the number of tasks selected by the participants.

[0124] The income of the participants is the sum of the incomes of all possible results multiplied by their weights:

[0125]

[0126] Among them, g(m) is the income of each possible result.

[0127] In this embodiment, first, the influence of this solution on task selection in the environment is evaluated. Figure 3 (a) shows the probability density map of task selection without using this solution. It is found that the uncertainty of the participants completing tasks is mainly concentrated in the space of (0, 3), and tasks with high uncertainty (3, 4) are less selected, and there are still tasks in remote areas that are not selected, with a low coverage rate; while 3(b) shows the distribution of task selection using this solution. It can be found that the uncertainty of task selection is evenly distributed in (0, 4), and the coverage rate of tasks is high. And it can be clearly found that Figure 3 (a) The points relative to Figure 3 (b) The points are significantly shifted to the left. It shows that under this solution, the participants will increase the uncertainty of the finally selected tasks, thereby increasing the coverage rate of tasks in remote areas.

[0128] To further evaluate the performance of the present invention, the present invention is also compared with other currently representative methods (DA-base), as Figure 4 shown. The verification results show that the performance of the present invention is superior to other currently state-of-the-art methods. As Figure 4 (a) shows, under the mechanism of DA-base, in order to maximize profits, participants will preferentially select tasks in popular regions. And as Figure 4 (b) shows, under the action of this solution, due to the role of Weiner attribution, some participants eliminate part of the uncertainty and select tasks in remote regions without the platform increasing the remuneration. At the same time, it can be found that Figure 4 (b), the tasks in the participant concentration area are generally completed more than those in the sparse area, because the increase in the participant attribution bias factor is a continuous process. During the process of the increase in the attribution bias factor of some participants, they will still complete tasks in popular regions.

[0129] Embodiment 2

[0130] This embodiment provides a crowd-sensing task allocation system based on Weiner attribution, including:

[0131] A participant attribution bias factor determination module, configured to calculate the corresponding attribution bias factor according to the relevant parameters of the participant attribution bias factor, and update the attribution bias factor according to the completion situation of the participant at the end of each task;

[0132] A task completion order determination module, configured to evaluate the evaluation weight of each task for the participant according to the action of the attribution bias factor, obtain the optimal weight matrix for the participant, and determine the optimal task completion order for the participant according to the magnitudes of the task weights in the optimal weight matrix;

[0133] A reward uncertainty preference weight determination module, configured to obtain the reward uncertainty preference weight according to the preference degree of the participant under the interaction of reward and uncertainty;

[0134] An optimal task completion quantity determination module, configured to combine the optimal weight matrix and the reward uncertainty preference weight, and calculate the optimal task completion quantity for the participant according to the optimal selectable items and relevant preferences of the participant;

[0135] A final reward distribution determination module, configured to determine the reward compensation factor for each region after evaluating and weighing the expected completion rate and the expected reward, and the determination of the reward compensation factor is based on the compensation weight of each region.

[0136] It should be understood that the functional unit modules in the various embodiments of the present invention may be concentrated in one processing unit, or each unit module may exist physically alone, or two or more unit modules may be integrated into one unit module, and may be implemented in the form of hardware or software.

[0137] Embodiment 3

[0138] This embodiment provides a crowd intelligence perception task allocation device based on Weiner attribution, including a sensor, a memory, and a processor; the sensor is used to collect crowd intelligence perception task data, the memory is used to store computer programs and the collected crowd intelligence perception task data, and the processor is used to execute the computer program to implement the crowd intelligence perception task allocation method based on Weiner attribution as described above.

[0139] Embodiment 4

[0140] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the crowd intelligence perception task allocation method based on Weiner attribution as described in Embodiment 1.

[0141] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0142] This application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in one Figure 1 one or more flows and / or blocks Figure 1 or multiple blocks.

[0143] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means that implement the functions specified in one or more of the flow Figure 1 one or more of the flows and / or blocks Figure 1 specified in one or more of the blocks.

[0144] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the flow Figure 1 one or more of the flows and / or blocks Figure 1 specified in one or more of the blocks.

[0145] It will be understood that the same or similar parts in the above embodiments may be referred to each other, and the content not described in detail in some embodiments may be referred to the same or similar content in other embodiments.

[0146] Any process or method description shown in the flowchart or described otherwise herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations in which functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0147] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and those of ordinary skill in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for allocating crowd-sourcing sensing tasks based on Weiner attribution, characterized in that Including: Determine the attribution bias factor of the participant, calculate the corresponding attribution bias factor according to the relevant parameters of the participant's attribution bias factor, and update the attribution bias factor according to the participant's completion situation at the end of each task; Determine the task completion order, evaluate the evaluation weight of each task for the participant according to the effect of the attribution bias factor, obtain the optimal weight matrix for the participant, and determine the optimal task completion order for the participant according to the magnitudes of the weights of each task in the optimal weight matrix; Determine the reward uncertainty preference weight, and obtain the reward uncertainty preference weight according to the preference degree of the participant under the interaction of reward and uncertainty; Determine the optimal task completion quantity, combine the optimal weight matrix and the reward uncertainty preference weight, and calculate the optimal task completion quantity for the participant according to the participant's optimal options and relevant preferences; Determine the final reward distribution, after evaluating and weighing the expected completion rate and the expected reward, determine the reward compensation factor for each region, and the determination of the reward compensation factor is based on the compensation weights of each region; The definition formula of the participant's attribution bias factor is: ； Among them, represents the attribution bias factor of the participant in the t-th task; and , respectively represent the reward of the previous task , the attribution bias factor of the previous task , the upper limit of the attribution bias factor among the three influence weights; represents the lower limit of the attribution bias factor; Among them, the upper limit of the attribution bias factor is calculated according to the following formula: ； Among them, represents the revenue of each task selection set, represents the set of basic rewards for all tasks, is the covariance set between two tasks, represents the lower limit of the uncertainty of the task selection set; Among them It is calculated according to the following formula: ； Among them, represents the basic remuneration, represents the weight of the task in the task selection set, represents the total number of tasks; According to the attribution bias factor , calculate the optimal weight matrix of the participant when performing the task for the t-th time , and it is expressed as follows: ； Calculate the optimal weight matrix The optimization principle to be satisfied when... is expressed as follows: ； Among them, represents the uncertainty of the task selection set, which can be expressed by the following formula: ； Among them, represents the correlation between task and task represents the uncertainty of participant for task represents the uncertainty of participant for task respectively represent the weights of task and task in the task selection set; Reward uncertainty preference weight It is expressed as follows: ； Among them, represents the sensitivity of the participant to ; represents the preference coefficient of the participant for uncertainty .

2. The method for allocating crowdsourcing sensing tasks based on Weiner attribution according to claim 1, wherein Under the optimal weight matrix and the reward uncertainty preference weight jointly, the uncertainty of the optimal selection task set is expressed as follows: 。 3. The method for allocating crowdsensing tasks based on Weiner attribution according to claim 2, characterized in that The optimal task completion quantity for the participant is: ； Among them, is the number of tasks that can be completed by the participants under normal working intensity, indicating the uncertainty corresponding to the optimal weight matrix.

4. The method for allocating crowdsourcing perception tasks based on Weiner attribution according to claim 2, characterized in that, After evaluating and weighing the relationship between the expected paid reward and the expected task completion quantity, in order to make up for the imbalance of task completion quantities between regions, different compensation weights need to be assigned to different regions, and the formula of the compensation weight is defined by the following formula: ； Among them, represents the compensation weight of the th region, represents the weight adjustment factor, represents the number of regions divided, represents the th region.

5. A crowd intelligence perception task allocation system based on Weiner attribution, characterized in that, Including: A participant's attribution bias factor determination module, configured to calculate the corresponding attribution bias factor according to the relevant parameters of the participant's attribution bias factor, and update the attribution bias factor according to the participant's completion situation at the end of each task; A task completion order determination module, configured to evaluate the evaluation weight of each task for the participant according to the effect of the attribution bias factor, obtain the optimal weight matrix for the participant, and determine the optimal task completion order for the participant according to the magnitudes of the weights of each task in the optimal weight matrix; A reward uncertainty preference weight determination module, configured to obtain the reward uncertainty preference weight according to the preference degree of the participant under the interaction of reward and uncertainty; An optimal task completion quantity determination module, configured to combine the optimal weight matrix and the reward uncertainty preference weight, and calculate the optimal task completion quantity for the participant according to the participant's optimal options and relevant preferences; A final reward distribution determination module, configured to determine the reward compensation factor for each region after evaluating and weighing the expected completion rate and the expected reward, and the determination of the reward compensation factor is based on the compensation weights of each region; The definition formula of the participant's attribution bias factor is: ； Among them, represents the attribution bias factor of the participant in the t-th task; and , respectively represent the benefit of the previous task , the attribution bias factor of the previous task , the upper limit of the attribution bias factor among the three influence weights; represents the lower limit of the attribution bias factor; Among them, the upper limit of the attribution bias factor is calculated according to the following formula: ； Among them, represents the revenue of each task selection set, represents the set of basic rewards for all tasks, is the covariance set between two tasks, represents the lower limit of the uncertainty of the task selection set; Among them It is calculated according to the following formula: ； Among them, represents the basic remuneration, represents the weight of the task in the task selection set, represents the total number of tasks; According to the attribution bias factor , calculate the optimal weight matrix of the participant when performing the task for the t-th time, and it is expressed as follows: ； Calculate the optimal weight matrix The optimization principle to be satisfied is expressed as follows: ； Among them, represents the uncertainty of the task selection set, which can be expressed by the following formula: ； Among them, represents the correlation between tasks and tasks represents the uncertainty of the participant for task represents the uncertainty of the participant for task respectively represent the weights of tasks and tasks in the task selection set; Reward uncertainty preference weight It is expressed as follows: ； Among them, indicates the sensitivity of the participant to ; indicates the preference coefficient of the participant for uncertainty .

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the group intelligence perception task allocation method based on Weiner attribution described in any one of claims 1 to 4.

7. A crowd-sensing task allocation device based on Weiner attribution, characterized in that, It includes a sensor, a memory, and a processor; the sensor is used to collect crowd-sensing task data, the memory is used to store computer programs and the collected crowd-sensing task data, and the processor is used to execute the computer programs to implement the crowd-sensing task allocation method based on Wiener attribution as described in any one of claims 1 to 4.

Images