Mobile crowd sensing related multi-task allocation algorithm based on greedy algorithm

Abstract

The invention discloses a mobile crowd sensing related multi-task allocation algorithm based on a greedy algorithm, which utilizes a fuzzy logic algorithm to calculate task-participant space-time distribution to obtain accurate task and participant space-time distribution, thereby defining different multi-task allocation scenes and providing reliable parameter support for a subsequent task allocation algorithm. The task allocation algorithm is more suitable for the scene, and the task allocation rate is improved; based on the calculated task-participant spatial and temporal distribution values, two allocation algorithms are provided for different task-participant spatial and temporal distribution conditions in combination with task correlation, and for the condition of balanced distribution or sufficient participants, the task allocation rate is improved as much as possible by adopting correlation allocation, and the platform overhead is saved; for the condition of participant distribution shortage, the task distribution completion rate is improved as much as possible by means of relevant distribution and by utilizing cost-saving excitation in a previous scene.

Description

Technical Field

[0001] The present invention relates to a multi-task allocation algorithm, and in particular to a mobile crowd-sensing related multi-task allocation algorithm based on a greedy algorithm. Background Art

[0002] With the popularization of smart devices and the rapid development of the Internet, Mobile Crowd Sensing (MCS) has become a new paradigm of perception. MCS uses mobile devices carried by users to collect and process perception data to complete complex perception tasks. Compared with traditional perception methods, MCS can break away from the heavy reliance on professionals and achieve super-time and large-scale wireless perception. In recent years, MCS has been widely used in urban traffic planning, public safety, environmental monitoring, disaster warning and other fields.

[0003] The MCS perception process can be mainly divided into three parts: task allocation, data collection, and perception data analysis. Task allocation is the first step of the MCS perception task and the basis for the successful completion of the perception task. Existing research mainly focuses on task allocation under spatiotemporal conditions, without paying special attention to task allocation strategies in different situations caused by the density of task publishers and participants. Therefore, the allocation strategy may be unsuitable for the current allocation scenario, resulting in low allocation efficiency. In addition, in existing task allocation research, no attention has been paid to the relevance of tasks in task allocation. Summary of the invention

[0004] In view of the problems existing in the above-mentioned prior art, the present invention provides a mobile crowd-sensing related multi-task allocation algorithm based on a greedy algorithm. By calculating the spatiotemporal distribution of tasks and participants, the appropriate allocation algorithm is selected in different scenarios to achieve the best allocation result, save costs and improve task allocation efficiency.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a mobile crowd intelligence perception related multi-task allocation algorithm based on a greedy algorithm, comprising the following steps:

[0006] Calculate the spatiotemporal distribution of tasks using fuzzy logic algorithms T and the spatial and temporal distribution of participants Dp; the time and space of tasks and participants are T n and S m ; The i-th task set is Task = {t1, t2, ..., t i}; the set of j participants is Part = {p1, p2, ..., p j}; The compensation that the task publisher is willing to pay for each i-task set is B i ={b1, b2, ..., bi}; and the cost of participant j for each task i is c i，j , PW i，j is the perceived willingness of participant p j for task t i ; TC(t i , t j ) is used to describe the correlation between tasks; the urgency of the task is ε; for related tasks t i and t j , if the participant has already been assigned a task t i , and then another related task t j is assigned, then the cost for the participant to execute t j will be reduced from c j to a*c j (0 < a < 1), where a is a coefficient;

[0007] In a time T n and space S m , if the spatio-temporal distribution D T of the task and the spatio-temporal distribution D P of the participant are both greater than or equal to the threshold, or the spatio-temporal distribution D T of the task and the spatio-temporal distribution D P of the participant are both less than the threshold, non-incentive processing is performed;

[0008] In a time T n and space S m , if the spatio-temporal distribution D T of the task is less than the threshold, while the spatio-temporal distribution D P of the participant is greater than or equal to the threshold, non-incentive processing is performed;

[0009] In a time T n and space S m , if the spatio-temporal distribution D T of the task is greater than or equal to the threshold, while the spatio-temporal distribution D P of the participant is less than the threshold, incentive processing is performed.

[0010] Furthermore, the PW i，j adopts binary. When PW i，j = 1, it means the participant is willing to perform the sensing task. Conversely, when PW i，j = 0, it means the participant is not willing to perform this task.

[0011] Furthermore, the TC(t i , t j ) is used to describe the correlation between tasks. If task t i and task tj Belonging to the same type of tasks and can be performed in the same time and space, participants perform tasks t i You can also perform tasks t j , then task t i and task t j are related, TC(t i , t j )=1; otherwise, task t i and t j is irrelevant, TC(t i , t j )=0.

[0012] Furthermore, the larger the urgency ε of the task is, the more urgent the task is and it needs to be assigned to the participants as soon as possible; otherwise, it means that the task is of average urgency and can be assigned later.

[0013] Furthermore, the non-incentive processing steps are:

[0014] First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i1 ;

[0015] Second, among all participants, select those who are interested in task t i1 Participants with perceived willingness, among these willing perceiving participants, determine the i1 The participant p with the lowest perceived cost j1 ;

[0016] If the participant p j1 Execute Task i1 The cost of task t is less than or equal to i1 The bid that is willing to be paid, then the task t i1 Assigned to participant p j 1Execute;

[0017] Then filter out the participants p in the task list j1 The tasks you are willing to perform, and then find the tasks in descending order of urgency. i1 Related tasks i2 ;

[0018] At this time, if the task t i2 Assigned to participant p j1 The cost a*c i2，j1 Less than or equal to task t j2 The minimum cost Minc executed by the participant i2 =min(c i2，1 , ..., c i2，j ), and a*ci2，j1 To be less than or equal to the execution task t i2 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i2 Assigned to participant p j1 Execute together;

[0019] In the process of task allocation, define R as the remaining cost of the non-incentive allocation. Where n is the number of times the related task assignment is completed.

[0020] Furthermore, the incentive processing steps are:

[0021] First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i3 ;

[0022] Second, among all participants, select those who are interested in task t i3 Participants with perceived willingness, among these willing perceiving participants, determine the i3 The participant p with the lowest perceived cost j2 ;

[0023] If the participant p j2 Execute Task i3 The cost of task t is less than or equal to i3 The bid that is willing to be paid, then the task t i3 Assigned to participant p j2 implement;

[0024] If the participant p j2 Execute Task i3 Cost of C i3，j2 Greater than task t i3 Willingness to pay i3 , at this time the expenditure gap is R i3 =C i3，j2 -b i3 , the remaining cost R′=RR i3 , if R′≥0, then stimulate R i3 , the task t i3 Assigned to participant p j2 to carry out execution;

[0025] When the task i3 Assigned to participant p j2 When executing, first filter out the participant p in the task list j2 The tasks that you are willing to perform, and then find the tasks in order of urgency from high to low. i3 Related tasks i4, if task t i4 Assigned to participant p j2 The cost a*c i4，j2 Less than or equal to task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 To be less than or equal to the execution task t i4 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i4 Assigned to participant p j2 Execute together, the remaining cost is R'=R'+(1-a)*c i4，j2 ;

[0026] If at this time the task t i4 Assigned to participant p j2 The cost a*c i4,j2 Less than task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 Greater than the task t i4 Willingness to pay i4 , at this time the expenditure gap is R i4 =a*c i4,j2 -b i4 , the remaining cost R″=R′-R i4 , if R″≥0, incentive is given and tasks are assigned to participants for p j2 implement;

[0027] The above process is repeated until all tasks are assigned or the number of participants is exhausted.

[0028] Furthermore, the spatiotemporal distribution D of the task is calculated by the fuzzy logic algorithm. T and the spatial and temporal distribution of participants D P The calculation steps are:

[0029] Firstly, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are determined, and the fuzzy sets of the time periods of task release and participant participation are divided into early trough ET, early peak EWP, equilibrium period BP, late peak LWP and late trough LT; the fuzzy sets of the sub-areas where tasks and participants are located are divided into small amount S, normal NO and large amount B; for the output variables, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are divided into sparse SP, less L, moderate equilibrium ME, more M and dense DE;

[0030] Secondly, regarding time, the 24 hours of a day are divided into l time periods, denoted as T L ={T1, ..., T n , ..., T l}; For space, divide an area into p sub-areas, denoted as S P ={S1, ..., S m , ..., S p}, l = 24, p = 12; therefore, the time T L and space S P As the input of fuzzy logic, we can calculate the spatiotemporal distribution of tasks and participants. After calculation, we can convert the spatiotemporal distribution of tasks into T and the spatial and temporal distribution of participants D P As the output of fuzzy logic;

[0031] The task release time membership function is:

[0032]

[0033] The task release area membership function is:

[0034]

[0035]

[0036] The membership function of the participant time distribution is:

[0037]

[0038] The membership function of the participant regional distribution is:

[0039]

[0040]

[0041] Then the defuzzification is achieved by weighted average method, the formula is:

[0042] Where u is the value of the membership function, and u = min(A(T n ), A(S m ));OW i is the weight, taking the middle value of each set;

[0043] Define the task's temporal and spatial distribution membership function set median sparse OW T-SP =5, less OW T-L =15, medium balance OW T-ME =45, more OW T-M =80, intensive OWT-DE =90;

[0044] Define the time and space distribution of participants Membership function Set median Sparse OW T-SP =10, less OW T-L =20, medium balance OW T-ME =50, more OW T-M =85, intensive OW T-DE =90;

[0045] Through the above calculation, several fuzzy outputs can be converted into an accurate output value OT. The obtained accurate output value OT is the spatiotemporal distribution D of the task. T Or the spatiotemporal distribution of participants Dp.

[0046] Compared with the prior art, the present invention has the following beneficial effects: the fuzzy logic algorithm is used to calculate the temporal and spatial distribution of tasks and participants, and the non-qualitative input is processed by fuzzy logic to obtain the accurate temporal and spatial distribution of tasks and participants, thereby defining different multi-task allocation scenarios, providing reliable parameter support for subsequent task allocation algorithms, making the task allocation algorithm more suitable for the scenario, and improving the task allocation rate;

[0047] Based on the calculated task-participant spatiotemporal distribution values ​​and combined with task relevance, a mobile crowd-sensing related multi-task allocation method based on a greedy algorithm is proposed. In this method, two allocation algorithms are proposed for different task-participant spatiotemporal distribution situations. When the distribution is balanced or there are sufficient participants, related allocation is used to increase the task allocation rate as much as possible and save platform overhead. When there is a shortage of participants, related allocation is used and incentives are provided by using the overhead savings in previous scenarios to increase the task allocation completion rate as much as possible. DETAILED DESCRIPTION

[0048] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] The present invention provides a mobile crowd-sensing related multi-task allocation algorithm based on a greedy algorithm, comprising the following steps:

[0050] The two core factors affecting the definition of the spatiotemporal distribution of tasks and participants are time and space. For example, during working hours and in densely populated areas, the number of tasks posted is relatively large; while during rest time and in sparsely populated areas, the number of tasks posted is relatively small. Therefore, the spatiotemporal distribution D of tasks is calculated by fuzzy logic algorithm.T and the spatial and temporal distribution of participants D P :

[0051] Firstly, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are determined, and the fuzzy sets of the time periods of task release and participant participation are divided into early trough ET, early peak EWP, equilibrium period BP, late peak LWP and late trough LT; the fuzzy sets of the sub-areas where tasks and participants are located are divided into small amount S, normal NO and large amount B; for the output variables, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are divided into sparse SP, less L, moderate equilibrium ME, more M and dense DE.

[0052] Secondly, regarding time, the 24 hours of a day are divided into l time periods, denoted as T L ={T1, ..., T n , ..., T l}; For space, divide an area into p sub-areas, denoted as S P ={S1, ..., S m , ..., S p}, according to the commonly used model, 1 = 24, p = 12; therefore, the time T L and space S P As the input of fuzzy logic, we can calculate the spatiotemporal distribution of tasks and participants. After calculation, we can convert the spatiotemporal distribution of tasks into T and the spatial and temporal distribution of participants D P As the output of fuzzy logic;

[0053] The task release time membership function is:

[0054]

[0055]

[0056] The task release area membership function is:

[0057]

[0058] The membership function of the participant time distribution is:

[0059]

[0060]

[0061] The membership function of the participant regional distribution is:

[0062]

[0063] The rules are defined through fuzzy reasoning. Here, the rules are defined as shown in Table 1:

[0064]

[0065] Table 1: Fuzzy logic rules table

[0066] According to the rules in Table 1, if the task release time period is in the early trough ET, and the sub-region where it is located is a sub-region with a small number of tasks S, it can be obtained that the temporal and spatial distribution of the task is sparse SP. Through the IF-THEN rule, IF <antecedents>THEN <consequent>; Specifically: IF(T n )and(S m ), THEN(D T ), this rule can be expressed as: IF(T n =ET)and(S m =S), THEN(D T =SP); Similarly, if the participant's moving time period is in the late peak LWP, and the sub-region where the participant is located is a sub-region with a large number of tasks B, it can be obtained that the time-space distribution of the participants is dense DE, and the above rule can be expressed as: IF(T n =LWP)and(S m =B), THEN(D P =DE).

[0067] For the calculation rule "and" in the above fuzzy logic, the minimum membership principle is adopted:

[0068] u=min(A(T n ), A(S m ));

[0069] On the basis of the above, defuzzification is then achieved through the weighted average method, the formula is: Where u is the value of the membership function, and u = min(A(T n ), A(S m ));OW i is the weight, taking the middle value of each set;

[0070] Define the task's temporal and spatial distribution membership function set median sparse OW T-SP =5, less OW T-L =15, medium balance OW T-ME =45, more OW T-M =80, intensive OW T-DE =90;

[0071] Define the time and space distribution of participants Membership function Set median Sparse OW T-SP =10, less OW T-L =20, medium balance OW T-ME =50, more OW T-M =85, intensive OW T-DE =90;

[0072] Through the above calculation, several fuzzy outputs can be converted into an available accurate output value OT. The obtained accurate output value OT is the spatiotemporal distribution D of the task. T or the spatial and temporal distribution of participants D P .

[0073] Define the time and space where the task and the participant are located as T n and S m , at time T n and space S m In the example, we define a set of i tasks as Task = {t1, t2, ..., t i }; the set of j participants is Part = {p1, p2, ..., p j }; The compensation that the task publisher is willing to pay for each i-task set is B i ={b1, b2, ..., b i }; and the cost of participant j for each task i is c i,j , In order to describe the participants’ willingness to perform tasks, PW is defined i，j For participants p j For task t i Perceived willingness, PW i，j Using binary, when PW i，j =1, indicating that the participant is willing to perform the perception task; otherwise, PW i，j = 0, indicating that the participant is unwilling to perform this task; we define the parameter TC(t i , t j ) is used to describe the correlation between tasks; in detail, if task t i and task t j Belonging to the same type of tasks and can be performed in the same time and space, participants perform tasks t i You can also perform tasks t j , then task t i and task t j are related, TC(t i , t j )=1; otherwise, task t i and t j is irrelevant, TC(t i , t j ) = 0. For example, Task 1 is to collect car acceleration data, and Task 2 is to collect car driving recorder data. Task 1 and Task 2 belong to the same car data and can be executed in the same time and space, but they are irreplaceable to each other. Then Task 1 and Task 2 are related.

[0074] In the process of assigning tasks, there are differences in the importance of tasks, so the urgency of the task is defined as ε. The larger the ε, the more urgent the task is and it needs to be assigned to the participants as soon as possible. On the contrary, it means that the task is not urgent and can be assigned later. When performing the perception task, since the participants are performing the task t i , it has moved to other space-time regions. If task t j With the task i is relevant. At this time, for the participants, performing task t j No additional space movement is required, so compared to simply executing the task t j , can save a certain amount of time and space costs; at the same time, due to the task t j With the task i For tasks belonging to the same category, participants can perform them at the same time without the need for secondary learning. Therefore, the execution cost of participants is further reduced, and their willingness to perform is enhanced. Therefore, for related tasks t i and t j , the participant has been assigned a task t i Based on this, assign a related task t j , then the participant executes t j The cost will be borne by c j Down to a*c j (0<a<1), a is the coefficient.

[0075] At a time T n and space S m If the task's spatiotemporal distribution D T and the spatial and temporal distribution of participants D P are greater than or equal to the threshold, or the temporal and spatial distribution of the task D T and the spatial and temporal distribution of participants D P If both are less than the threshold, it is considered that the task and the participant's resources can be matched in this area, and no additional incentive is needed to motivate the user to perform the task, so non-incentive treatment can be performed;

[0076] At a time T n and space S m If the task's spatiotemporal distribution D T is less than the threshold, and the spatial and temporal distribution of the participants D P If it is greater than or equal to the threshold, it means that there are few tasks in this area and many participants. It is not necessary to provide additional incentives to participants to motivate users to perform tasks, so non-incentive treatment can be performed;

[0077] At a time T n and space S m If the task's spatiotemporal distribution D T is greater than or equal to the threshold, and the spatial and temporal distribution of the participants D P If it is less than the threshold, it means that in this space-time region, there are more tasks and fewer participants. Some incentives need to be provided to the participants to encourage them to complete as many tasks as possible. Therefore, incentive processing can be performed.

[0078] The non-incentive processing steps are:

[0079] First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i1 ;

[0080] Second, among all participants, select those who are interested in task t i1 Participants with perceived willingness, among these willing perceiving participants, determine the i1 The participant p with the lowest perceived cost j1 ;

[0081] If the participant p j1 Execute Task i1 The cost of task t is less than or equal to i1 The bid that is willing to be paid, then the task t i1 Assigned to participant p j1 implement;

[0082] Then filter out the participants p in the task list j1 The tasks you are willing to perform, and then find the tasks in descending order of urgency. i1 Related tasks i2 ;

[0083] At this time, if the task t i2 Assigned to participant p j1 The cost a*c i2，j1 Less than or equal to task t j2 The minimum cost Minc executed by the participant i2 =min(c i2，1 , ..., c i2，j ), and a*c i2，j1 To be less than or equal to the execution task t i2 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i2 Assigned to participant p j1 Execute together;

[0084] In the process of task allocation, define R as the remaining cost of the non-incentive allocation. Where n is the number of times the related task assignment is completed.

[0085] The steps of incentive processing are:

[0086] First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i3 ;

[0087] Second, among all participants, select those who are interested in task t i3 Participants with perceived willingness, among these willing perceiving participants, determine the i3 The participant p with the lowest perceived cost j2 ;

[0088] If the participant p j2 Execute Task i3 The cost of task t is less than or equal to i3 The bid that is willing to be paid, then the task t i3 Assigned to participant p j2 implement;

[0089] If the participant p j2 Execute Task i3 Cost of C i3，j2 Greater than task t i3 Willingness to pay i3 , at this time the expenditure gap is R i3 =C i3，j2 -b i3 , the remaining cost R′=RR i3 , if R′≥0, then stimulate R i3 , the task t i3 Assigned to participant p j2 to carry out execution;

[0090] When the task i3 Assigned to participant p j2 When executing, first filter out the participant p in the task list j2 The tasks that you are willing to perform, and then find the tasks in order of urgency from high to low. i3 Related tasks i4 , if task t i4 Assigned to participant p j2 The cost a*c i4，j2 Less than or equal to task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 To be less than or equal to the execution task t i4 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i4 Assigned to participant p j2 Execute together, the remaining cost is R'=R'+(1-a)*c i4，j2 ;

[0091] If at this time the task t i4 Assigned to participant p j2 The cost a*c i4，j2 Less than task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 Greater than the task t i4 Willingness to pay i4 , at this time the expenditure gap is R i4 =a*c i4，j2 -b i4 , the remaining cost R″=R′-R i4 , if R″≥0, incentive is given and tasks are assigned to participants for p j 2. Execution;

[0092] The above process is repeated until all tasks are assigned or the number of participants is exhausted.

[0093] The task set Task can be divided into a non-incentive task set Task1 and an incentive task set Task2. The assigned task set is defined as TA, and the initial value is Therefore, the algorithm can be summarized as:

[0094]

[0095]

[0096]

[0097]

[0098] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any marking in a claim should not be considered as limiting the claim to which it relates.

[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any slight modification, equivalent substitution and improvement made to the above embodiment based on the technical essence of the present invention should be included in the protection scope of the technical solution of the present invention.< / consequent> < / antecedents>

Claims

1. A mobile crowd-sensing related multi-task allocation algorithm based on a greedy algorithm, characterized in that: The following steps are involved: Calculate the spatio-temporal distribution D of tasks through a fuzzy logic algorithm T and the spatio-temporal distribution D of participants P ; The time and space where the tasks and participants are located are T n and S m ; The set of i tasks is Task = {t1, t2,..., t i}; The set of j participants is Part = {p1, p2,..., p j}; The compensations that the task publishers are willing to pay for the set of i tasks are B i = {b1, b2,..., b i}; And the cost of participant j for each task i is c i，j , PW i，j is the perceived willingness of participant p j for task t i ; TC(t i , t j ) is used to describe the correlation between tasks; The urgency of the task is ε; For related tasks t i and t j , on the basis that the participant has been assigned a task t i , and then assigned a related task t j , then the cost for the participant to execute t j will be reduced from c j to a * c j (0 < a < 1), where a is a coefficient; At a time T n and space S m If the task's spatiotemporal distribution D T and the spatial and temporal distribution of participants D P are greater than or equal to the threshold, or the temporal and spatial distribution of the task D T and the spatial and temporal distribution of participants D P All are less than the threshold value, and non-excitation treatment is performed; At a time T n and space S m If the task's spatiotemporal distribution D T is less than the threshold, and the spatial and temporal distribution of the participants D P If it is greater than or equal to the threshold, non-excitation processing is performed; At a time T n and space S m If the task's spatiotemporal distribution D T is greater than or equal to the threshold, and the spatial and temporal distribution of the participants D P If the value is less than the threshold, excitation processing is performed.

2. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The PW i，j Using binary, when PW i，j =1, indicating that the participant is willing to perform the perception task; otherwise, PW i，j =0, indicating that the participant is unwilling to perform the task.

3. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The TC (t i , t j ) is used to describe the correlation between tasks. If task t i and task t j Belonging to the same type of tasks and can be performed in the same time and space, participants perform tasks t i You can also perform tasks t j , then task t i and task t j are related, TC(t i , t j )=1; otherwise, task t i and t j is irrelevant, TC(t i , t j )=0.

4. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The larger the urgency ε of the task is, the more urgent the task is and it needs to be assigned to the participants as soon as possible; otherwise, it means that the task is of average urgency and can be assigned later.

5. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The non-incentive processing steps are: First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i1 ; Second, among all participants, select those who are interested in task t i1 Participants with perceived willingness, among these willing perceiving participants, determine the i1 The participant p with the lowest perceived cost j1 ; If the participant p j1 Execute Task i1 The cost of task t is less than or equal to i1 The bid that is willing to be paid, then the task t i1 Assigned to participant p j1 implement; Then filter out the participants p in the task list j1 The tasks you are willing to perform, and then find the tasks in descending order of urgency. i1 Related tasks i2 ; At this time, if the task t i2 Assigned to participant p j1 The cost a*c i2，j1 Less than or equal to task t j2 The minimum cost Minc executed by the participant i2 =min(c i2，1 , ..., c i2，j ), and a*c i2，j1 To be less than or equal to the execution task t i2 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i2 Assigned to participant p j1 Execute together; In the process of task allocation, define R as the remaining cost of the non-incentive allocation. Where n is the number of times the related task assignment is completed.

6. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The incentive processing steps are: First, sort the tasks from high to low in terms of urgency and determine the most urgent task. i3 ; Second, among all participants, select those who are interested in task t i3 Participants with perceived willingness, among these willing perceiving participants, determine the i3 The participant p with the lowest perceived cost j2 ; If the participant p j2 Execute Task i3 The cost of task t is less than or equal to i3 The bid that is willing to be paid, then the task t i3 Assigned to participant p j2 implement; If the participant p j2 Execute Task i3 Cost of C i3，j2 Greater than task t i3 Willingness to pay i3 , at this time the expenditure gap is R i3 =C i3，j2 -b i3 , the remaining cost R′=RR i3 , if R′≥0, then stimulate R i3 , the task t i3 Assigned to participant p j2 to carry out execution; When the task i3 Assigned to participant p j2 When executing, first filter out the participant p in the task list j2 The tasks that you are willing to perform, and then find the tasks in order of urgency from high to low. i3 Related tasks i4 , if task t i4 Assigned to participant p j2 The cost a*c i4，j2 Less than or equal to task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 To be less than or equal to the execution task t i4 If the user is willing to pay the fee, relevant tasks will be assigned and the task t i4 Assigned to participant p j2 Execute together, the remaining cost is R'=R'+(1-a)*c i4，j2 ; If at this time the task t i4 Assigned to participant p j2 The cost a*c i4，j2 Less than task t j4 The minimum cost Minc executed by the participant i4 =min(c i4，1 , ..., c i4，j ), and a*c i4，j2 Greater than the task t i4 Willingness to pay i4 , at this time the expenditure gap is R i4 =a*c i4，j2 -b i4 , the remaining cost R″=R′-R i4 , if R″≥0, then incentive is given and tasks are assigned to participants for p j2 implement; The above process is repeated until all tasks are assigned or the number of participants is exhausted.

7. According to the greedy algorithm-based mobile crowd intelligence perception related multi-task allocation algorithm of claim 1, it is characterized in that: The spatiotemporal distribution D of the task is calculated by the fuzzy logic algorithm. T and the spatial and temporal distribution of participants D P The calculation steps are: Firstly, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are determined, and the fuzzy sets of the time periods of task release and participant participation are divided into early trough ET, early peak EWP, equilibrium period BP, late peak LWP and late trough LT; the fuzzy sets of the sub-areas where tasks and participants are located are divided into small amount S, normal NO and large amount B; for the output variables, the fuzzy sets of the spatiotemporal distribution of tasks and the spatiotemporal distribution of participants are divided into sparse SP, less L, moderate equilibrium ME, more M and dense DE; Secondly, for time, divide the 24 hours of a day into l time periods, denoted as T L ={T1, ..., T n , ..., T l }; For space, divide an area into p sub-areas, denoted as S P ={S1, ..., S m , ..., S p }, l = 24, p = 12; Therefore, the time T L and space S P As the input of fuzzy logic, we can calculate the spatiotemporal distribution of tasks and participants. After calculation, we can convert the spatiotemporal distribution of tasks into T and the spatial and temporal distribution of participants Dp as the output of fuzzy logic; then The task release time membership function is: The task release area membership function is: The membership function of the participant time distribution is: The membership function of the participant regional distribution is: Then the defuzzification is achieved by weighted average method, the formula is: Where u is the value of the membership function, and u = min(A(T n ),A(S m ));OW i is the weight, taking the middle value of each set; Define the task's temporal and spatial distribution membership function set median sparse OW T-SP =5, less OW T-L =15, medium balance OW T-ME =45, more OW T-M =80, intensive OW T-DE =90; Define the time and space distribution of participants Membership function Set median Sparse OW T-SP =10, less OW T-L =20, medium balance OW T-ME =50, more OW T-M =85, intensive OW T-DE =90; Through the above calculation, several fuzzy outputs can be converted into an accurate output value OT. The obtained accurate output value OT is the spatiotemporal distribution D of the task. T or the spatial and temporal distribution of participants D P .

Images