Crowd sensing associated task allocation method based on minimum residual task quantity priority

Abstract

The invention belongs to the technical field of mobile crowd sensing, and particularly relates to a crowd sensing associated task allocation method based on minimum residual task number priority, which is characterized in that a plurality of task requesters exist in a mobile crowd sensing system, and each task requester issues a task set comprising a plurality of associated sensing tasks, only when all the tasks in the set are completed, it is considered that the tasks issued by the requester are effectively completed. According to the method, different priorities are set for the remaining tasks in the task sets of different requesters, so that the tasks in the task sets of the requesters with fewer remaining uncompleted tasks have higher priorities, and thus the tasks can be completed preferentially in the task allocation process, and the effective completed number of the associated tasks is expected to be increased.

Description

Technical Field

[0001] The present invention belongs to the technical field of mobile crowd sensing, and in particular relates to a crowd sensing associated task allocation method based on the priority of the minimum number of remaining tasks. Background Art

[0002] In the mobile crowd intelligence perception system, the task requester submits tasks to the service platform according to his own needs. The service platform is responsible for recruiting mobile users and assigning perception tasks to different users. The task allocation method is the key to affecting the performance of the mobile crowd intelligence perception system. An effective task allocation scheme can recruit as few users as possible to complete more perception tasks. However, when the task requester publishes multiple interrelated perception tasks at the same time, due to the correlation between these tasks, in the application considered by this patent, the completion of these tasks can only be considered valid when all interrelated tasks are completed. When allocating tasks, the number of remaining unfinished tasks issued by each requester needs to be considered. The goal of the present invention is to design a crowd intelligence perception related task allocation method based on the priority of the minimum number of remaining tasks. Summary of the invention

[0003] The present invention designs a method for allocating associated tasks based on crowd intelligence perception with priority given to the minimum number of remaining tasks. The method adopts an inverse function to set different priorities for the remaining tasks in the task sets of different requesters, so that the tasks in the task sets of requesters with fewer remaining unassigned tasks have higher priorities, so that they can be completed first in the task allocation process. When the platform recruits users, it first calculates the priorities of all tasks to be assigned based on the task execution status in the task set published by each requester. On this basis, it calculates the utility values ​​of all users to be recruited based on the task execution status and travel costs of each user to be recruited, and then recruits users and selects paths based on the utility values ​​of the users to be recruited. The goal of this method is to increase the number of effective completions of associated tasks under limited budget constraints.

[0004] The application environment of the present invention is as follows:

[0005] This method is aimed at a mobile crowd-sensing system consisting of a service platform, multiple task requesters and multiple mobile users. Each task requester publishes a task set including multiple related perception tasks. The completion of these tasks can only be considered valid when all tasks in the set are completed; each mobile user provides one or more alternative travel paths to reach his destination. The platform can recruit mobile users and select a travel path for them based on the execution of tasks on different paths, and pay corresponding travel compensation. The platform has a certain amount of budget for user recruitment. Each perception task can be completed by any mobile user and only needs to be completed once. This method assumes that the tasks published by all requesters are homogeneous and that all mobile users have the same perception capabilities. The goal is to maximize the number of effective completions of related tasks. This method includes the following steps:

[0006] A) Generate system architecture:

[0007] The mobile crowd sensing system consists of a platform, multiple task requesters and multiple mobile users. The set of task requesters in the system is denoted as R, and each requester r j ∈R, 1≤j≤|R|, publish a set T containing multiple related perception tasks j , only when all the tasks in the set are completed can the task issued by the requester be considered to be effectively completed, and partial completion is meaningless; each perception task t i ∈T j , 1≤i≤|T j |, can be completed by any mobile user and only needs to be completed once. When the task-bearing user passes by the task location, the task is considered to be completed, that is, the task completion only requires some simple execution actions without additional execution time; the set of mobile users in the system is denoted as W, and each mobile user w k ∈W, 1≤k≤|W|, reports one or more alternative travel routes to the platform to reach its destination. The platform can recruit new mobile users and select a travel route for them based on the task execution status on different alternative routes, and pay corresponding travel compensation. For a newly recruited user, all pending tasks on its mobile route will be assigned to the user; the platform has a certain amount of budget for user recruitment;

[0008] B) Calculate task priority:

[0009] Let the jth requester r j Submitted associated task set T j The total number of tasks in is recorded as When T j The number of tasks assigned in When any task t∈T is yet to be assigned j, whose priorities are:

[0010]

[0011] The above expression sets the priority of each task to be assigned to the reciprocal of the total number of unassigned tasks in the task set of the requester to which it belongs. That is, the smaller the number of unassigned tasks in the same task set, the higher the priority of the task. In addition, for a task that has been assigned, its priority is 0.

[0012] C) Recruiting mobile users and determining their movement paths:

[0013] 1) Calculate the priority of all tasks to be assigned based on the current execution status of the task set of each task requester;

[0014] 2) For each candidate user w k ∈CW, where CW is the candidate user set, that is, the user set that has not yet been assigned a task. It is virtualized into multiple users according to the number of its alternative travel paths. Each path corresponds to a virtual user. Represents user w k The virtual user generated when selecting the i-th path is then calculated The utility value is:

[0015] The ratio of the sum of the priorities of the tasks yet to be assigned to the tasks to their travel costs is as follows:

[0016]

[0017] in, Represents user w k The set of all tasks yet to be assigned that can be executed when the i-th path is selected,

[0018] Represents user w k The travel cost when choosing the i-th path is the travel compensation that the platform needs to provide to the user, and its value is greater than 0;

[0019] 3) According to the utility values ​​of the virtual users to be calculated in step 2), the actual user corresponding to the virtual user with the largest utility value greater than 0 is recruited from all these virtual users, and its moving path is determined. All the perception tasks to be assigned on the path are assigned to the user, and the user recruited in the current step is deleted from the candidate user set CW;

[0020] 4) Repeat the above steps 1)-3) until all users are recruited, all tasks are assigned, there are no more virtual users with a useful value greater than 0, or the budget is exhausted.

Claims

1. A method for allocating group intelligence-sensing related tasks based on the priority of the minimum number of remaining tasks, characterized by: A) Generate system architecture: The mobile crowd sensing system consists of a platform, multiple task requesters and multiple mobile users. The set of task requesters in the system is denoted as R, and each requester r j ∈R, 1≤j≤|R|, publish a set T containing multiple related perception tasks j , only when all the tasks in the set are completed can the task issued by the requester be considered to be effectively completed, and partial completion is meaningless; each perception task t i ∈T j , 1≤i≤|T j |, can be completed by any mobile user and only needs to be completed once. When the task-bearing user passes by the task location, the task is considered to be completed, that is, the task completion only requires some simple execution actions without additional execution time; the set of mobile users in the system is denoted as W, and each mobile user w k ∈W, 1≤k≤|W|, reports one or more alternative travel routes to the platform to reach its destination. The platform can recruit new mobile users and select a travel route for them based on the task execution status on different alternative routes, and pay corresponding travel compensation. For a newly recruited user, all pending tasks on its mobile route will be assigned to the user; the platform has a certain amount of budget for user recruitment; B) Calculate task priority: Let the jth requester r j Submitted associated task set T j The total number of tasks in is recorded as When T j The number of tasks assigned in When any task t∈T j , whose priorities are: The above expression sets the priority of each task to be assigned to the reciprocal of the total number of unassigned tasks in the task set of the requester to which it belongs. That is, the smaller the number of unassigned tasks in the same task set, the higher the priority of the task. In addition, for a task that has been assigned, its priority is 0. C) Recruiting mobile users and determining their movement paths: 1) Calculate the priority of all tasks to be assigned based on the current execution status of the task set of each task requester; 2) For each candidate user w k ∈CW, where CW is the candidate user set, that is, the user set that has not yet been assigned a task. It is virtualized into multiple users according to the number of its alternative travel paths. Each path corresponds to a virtual user. Represents user w k The virtual user generated when selecting the i-th path is then calculated The utility value is: The ratio of the sum of the priorities of the tasks yet to be assigned to the tasks to their travel costs is as follows: in, Represents user w k The set of all tasks yet to be assigned that can be executed when the i-th path is selected, Represents user w k The travel cost when choosing the i-th path is the travel compensation that the platform needs to provide to the user, and its value is greater than 0; 3) According to the utility values ​​of the virtual users to be calculated in step 2), the actual user corresponding to the virtual user with the largest utility value greater than 0 is recruited from all these virtual users, and its moving path is determined. All the perception tasks to be assigned on the path are assigned to the user, and the user recruited in the current step is deleted from the candidate user set CW; 4) Repeat the above steps 1)-3) until all users are recruited, all tasks are assigned, there are no more virtual users with a useful value greater than 0, or the budget is exhausted.

Images