A method and system for mobile swarm intelligence sensing task allocation based on fuzzy clustering and dynamic scheduling

Abstract

This invention proposes a mobile swarm intelligence sensing task allocation method and system based on fuzzy clustering and dynamic scheduling. First, fuzzy clustering is used to mine the latent spatial structure of tasks, and tasks are classified into fuzzy clusters with membership attributes through soft partitioning. After obtaining the spatiotemporal distribution characteristics of tasks, a benefit matrix including spatiotemporal feasibility and time cost is constructed, and the Hungarian algorithm is used to achieve a globally optimal match between worker and task clusters. For the online execution phase, a dynamic scheduling mechanism including a primary task pool and a collaborative task pool is designed. Workers select tasks based on a hybrid priority index combining time urgency, spatial proximity, and cluster membership. This invention solves the problem of balancing global efficiency and local flexibility in large-scale task allocation through a hierarchical mechanism of "macro-partitioning-meso-matching-micro-scheduling," significantly improving the overall task completion rate while effectively reducing the system's time resource cost.

Description

Technical Field

[0001] This invention belongs to the field of mobile crowd sensing and resource scheduling technology, specifically relating to a mobile crowd sensing task allocation method and system based on fuzzy clustering and dynamic scheduling. Background Technology

[0002] Mobile crowdsensing, as an emerging data collection paradigm, utilizes ordinary users carrying smart mobile devices as "workers" to sense and collect environmental data. The core challenge of operating a mobile crowdsensing platform lies in how to effectively allocate massive amounts of tasks with spatiotemporal constraints to workers with varying abilities and dynamically changing movement trajectories.

[0003] Existing research primarily focuses on local optimization strategies, such as path planning for individual workers or on-demand task matching based on greedy heuristics. While these methods are computationally efficient, they often fail to achieve an ideal balance between global task completion rate and overall time efficiency. In particular, the spatial distribution of tasks typically exhibits natural clustering characteristics, while worker decision-making behaviors are diverse and dynamic, making the design of a framework capable of achieving globally optimal resource allocation extremely complex. Simply pursuing the globally optimal solution leads to an exponential increase in computational complexity with the number of tasks and workers, making it difficult to meet real-time requirements. Conversely, relying solely on local greedy approaches can easily result in uneven task allocation, with a large backlog of tasks in some areas, thus increasing the overall time resource cost of the system. Therefore, a task allocation scheme that can balance macro-level global planning with micro-level dynamic scheduling is needed. Summary of the Invention

[0004] The purpose of this invention is to propose a mobile swarm intelligence sensing task allocation method and system based on fuzzy clustering and dynamic scheduling. This method and system uses a hierarchical mechanism to mine the task space structure and perform optimal matching at the macro level using fuzzy clustering, and performs dynamic scheduling through hybrid priority at the micro level, thereby significantly reducing the overall time resource cost of the task while ensuring a high task completion rate.

[0005] To achieve the above objectives, the technical solution of the present invention is: a mobile swarm intelligence sensing task allocation method based on fuzzy clustering and dynamic scheduling. The method aims to maximize the overall task completion rate and minimize the overall task time and resource cost, and allocates tasks to a given task set and worker set, comprising the following three stages:

[0006] The first stage is the fuzzy task partitioning stage, which uses a fuzzy clustering algorithm to divide tasks with geographical location attributes into different clustering regions, and obtains the membership matrix of each task to each cluster and the cluster center.

[0007] The second stage is the optimal worker-cluster assignment stage. Based on the membership matrix, a benefit matrix between workers and clusters is constructed. The maximum weight matching algorithm is used to determine the one-to-one mapping relationship between workers and task clusters, and each worker is assigned a responsible cluster region.

[0008] The third stage is the hybrid priority scheduling stage, which enters the online execution process. Workers construct task pools based on the clustering regions they are responsible for, and dynamically select and execute tasks according to hybrid priority indicators that include time, space, and membership.

[0009] Preferably, the step of using a fuzzy clustering algorithm to divide tasks with geographical location attributes into different clustering regions, and obtaining the membership matrix and cluster centers of each task to each cluster, specifically includes the following steps:

[0010] Initialize the membership matrix ; Update the cluster center set iteratively and membership matrix Until convergence;

[0011] Update the membership matrix elements in The calculation formula is:

[0012]

[0013] in, For the task location, and All are cluster centers. For fuzzy coefficients, The total number of clusters, This is an index variable used to iterate through the cluster centers;

[0014] Update cluster center The calculation formula is:

[0015]

[0016] in, The total number of tasks. This is an index variable used for traversing tasks;

[0017] When the change in the membership matrix is ​​less than the convergence threshold Alternatively, iteration can stop when the maximum number of iterations is reached, and the final membership matrix can be output. and cluster center set .

[0018] Preferably, the step of constructing a benefit matrix between workers and clusters based on the membership matrix, and using the maximum weight matching algorithm to determine the one-to-one mapping relationship between workers and task clusters, is as follows:

[0019] First, construct the benefit matrix. Benefit matrix elements in Indicates that workers Assigned to cluster The expected benefits are calculated using the following formula:

[0020]

[0021] in, For workers Execute the task The estimated time cost, It is a local minimum introduced to prevent the denominator from being zero. For spatiotemporal feasibility assessment function:

[0022]

[0023] in, Indicates the current time. Indicates task Expiration date Indicates workers The deadline for returning to the initial location. Indicates workers movement speed, Indicates task Location and workers The distance between the initial positions;

[0024] Next, the benefit matrix Transform into a cost matrix Cost matrix elements in The calculation formula is as follows:

[0025]

[0026] The Hungarian algorithm is then used to solve the cost matrix to obtain the optimal matching scheme that maximizes the total benefit.

[0027] Preferably, in the hybrid priority scheduling phase, each worker Maintain the primary task pool and collaborative task pool :

[0028] The primary task pool Includes the clusters responsible for workers in all unassigned tasks. The task with the highest membership level;

[0029] The collaborative task pool Includes clustering in all unassigned tasks Its membership degree is not the highest, but it exceeds the preset collaboration threshold. The task;

[0030] When workers select tasks from a candidate list consisting of a primary task pool and a collaborative task pool, they do so based on a hybrid priority metric. Sort and combine priority metrics Based on time priority Space priority Membership priority Weighted summation yields:

[0031]

[0032] in: Calculated based on the remaining effective time of the task. For the task The remaining valid time, and These represent the maximum and minimum remaining effective time for all tasks in the task pool, respectively. Based on the distance calculation between the task and the worker, This represents the straight-line distance between two locations. Indicates workers Location, and Each task and worker is represented in the task pool. Distance between The maximum and minimum values ​​in; Based on task Clustering membership degree calculate; These are the weighting coefficients.

[0033] The hybrid priority scheduling phase employs a dynamic scheduling strategy, and the specific execution steps are as follows:

[0034] Step A1: When the worker When in an idle state, update workers based on the current set of unassigned tasks. The primary task pool and collaborative task pool And merge them into a candidate list;

[0035] Step A2: Calculate the mixed priority for each task in the candidate list. The task is to sort the values ​​in ascending order.

[0036] Step A3: Check the sorted task list sequentially, using the feasibility function. Judgment Task Does it satisfy the spatiotemporal constraints?

[0037] Step A4: Select the first task in the list that meets the constraints as the next execution target, update the status of the selected task to "assigned", and remove the selected task from the set of unassigned tasks.

[0038] Preferably, the overall task time resource cost in the optimization objective incorporates a time penalty term for unassigned tasks;

[0039] The total task time resource cost is the average time cost of all tasks, including both assigned and unassigned tasks.

[0040] For tasks that have been successfully assigned, calculate the actual completion time cost of the assigned task. ;

[0041] For tasks that are ultimately not assigned, calculate the penalty time cost for all unassigned tasks. The penalty time cost Defined as:

[0042]

[0043] in It represents the maximum straight-line distance between any two task locations among all unassigned tasks in the task region. Minimum moving speed for workers, This is the minimum processing speed for workers.

[0044] A mobile crowd sensing task allocation system based on fuzzy clustering and dynamic scheduling includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs any of the steps in the above-mentioned mobile crowd sensing task allocation method.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] This invention employs a hierarchical mechanism to mine the spatial structure of tasks and perform optimal matching using fuzzy clustering at the macro level, and to dynamically schedule tasks using hybrid priorities at the micro level. This approach ensures a high task completion rate while significantly reducing the overall time and resource costs of the tasks. Attached Figure Description

[0047] Figure 1 This is a framework diagram for implementing the method according to an embodiment of the present invention. Detailed Implementation

[0048] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0049] This invention proposes a mobile swarm intelligence sensing task allocation method based on fuzzy clustering and dynamic scheduling. First, the system uses fuzzy task partitioning to mine the potential spatial distribution structure of tasks and obtain the membership degree of each task to different cluster centers, thus achieving soft task partitioning. Second, in the optimal worker-cluster allocation stage, the system constructs a benefit matrix quantifying the degree of worker-cluster matching based on task membership, worker mobility, and spatiotemporal constraints. The Hungarian algorithm is then used to solve for the globally optimal match, assigning each worker a primary task cluster region. Finally, in the online execution stage, hybrid priority scheduling is employed. Workers construct a primary task pool containing core tasks and a collaborative task pool containing surrounding tasks based on their assigned clusters. Tasks are dynamically selected based on a hybrid priority index combining time urgency, spatial proximity, and cluster membership, achieving real-time atomic task allocation.

[0050] In this embodiment, the mobile crowd sensing system includes a task set. Gather with workers Each task in the task set Includes location coordinates and expiration time; each worker in the worker set This includes the current location, movement speed, processing speed, and task deadline. The system performs task allocation and scheduling according to the following steps:

[0051] The first step is for the system to perform fuzzy task partitioning. To reveal the potential spatial structure of the tasks, fuzzy clustering is used to partition them. The tasks are divided into This step involves clustering. Unlike traditional hard partitioning, this step computes the task... Belongs to clustering membership degree Specifically, initialize the membership matrix. By iteratively updating the cluster centers membership degree Until convergence. Among them, the cluster centers... The weighted centroids are based on membership degree. Then, the distance from the task to each cluster center and its fuzzy weighted index are used. Decision made. This step outputs the final membership matrix. and cluster center set It preserves the fuzzy boundary information of the task space distribution, providing a data foundation for subsequent flexible collaboration.

[0052] The second step involves the system performing optimal worker-cluster assignment. At this stage, the system determines which region each worker is macroscopically responsible for. First, a benefit matrix is ​​constructed. , of which elements Representatives will work Assigned to cluster The overall expected benefits. This benefit value is proportional to the task membership degree. The power of the factor is inversely proportional to the estimated time cost for workers to complete the task. At the same time, a feasibility function is introduced. A hard constraint check is performed on task expiration time and worker deadlines; if these are not feasible, the benefit is zero. After construction, the benefit matrix is ​​generated. Convert to cost matrix (That is, take the maximum value minus the current value), and apply the Hungarian algorithm to solve for the minimum cost (i.e., maximum benefit) matching. The final matching result is... A one-to-one correspondence between workers and clusters was established, that is, each worker... Assigned to be responsible for a specific cluster .

[0053] The third step involves the system executing mixed priority scheduling. This is the online dynamic execution phase. When workers... When in an idle state, based on the cluster it is responsible for Together with the current set of unassigned tasks, construct two task pools:

[0054] Priority Pool : Contains all pairs of clusters The unassigned task with the highest degree of subordination represents the worker's core area of ​​responsibility.

[0055] Collaborative task pool : Includes members whose membership degree is not the highest but exceeds a certain threshold Unassigned tasks represent the surrounding area where workers can assist. The two pools are merged into a candidate list, and the mixed priority for each task is calculated. This indicator is composed of three weighted parts:

[0056] (1) Time priority This reflects the urgency of the remaining time for the task;

[0057] (2) Spatial priority This reflects the distance between the task and the workers;

[0058] (3) Membership priority This reflects whether the task falls within the worker's core area of ​​responsibility. The specific calculation formula is as follows: All values ​​are normalized, and smaller values ​​have higher priority. These are the weighting coefficients.

[0059] The fourth step is for the system to complete task assignment. Workers select from the candidate list... The values ​​are sorted in ascending order, and the spatiotemporal feasibility of each task is checked sequentially. Once a task that meets the criteria is found... The system performs an atomic operation, updates the task status to "assigned," and assigns it to a worker. The worker is then removed from the unassigned task set. If there are no currently feasible tasks, the worker either waits or moves to the cluster center.

[0060] The fifth step is to evaluate the overall system performance indicators. After all tasks are completed or all workers have left, the system calculates the overall task completion rate and the overall task time and resource cost as evaluation indicators. The overall task time and resource cost not only includes the actual time spent on completed tasks but also introduces a penalty time item for unassigned residual tasks. To comprehensively measure the effectiveness of the allocation plan.

[0061] The above are preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for allocating mobile swarm sensing tasks based on fuzzy clustering and dynamic scheduling, characterized in that, The method aims to maximize the overall task completion rate and minimize the overall task time and resource cost. It allocates tasks to a given set of tasks and workers, and includes the following three stages: The first stage is the fuzzy task partitioning stage, which uses a fuzzy clustering algorithm to divide tasks with geographical location attributes into different clustering regions, and obtains the membership matrix of each task to each cluster and the cluster center. The second stage is the optimal worker-cluster assignment stage. Based on the membership matrix, a benefit matrix between workers and clusters is constructed. The maximum weight matching algorithm is used to determine the one-to-one mapping relationship between workers and task clusters, and each worker is assigned a responsible cluster region. The third stage is the hybrid priority scheduling stage, which enters the online execution process. Workers construct task pools based on the clustering regions they are responsible for, and dynamically select and execute tasks according to hybrid priority indicators that include time, space and membership. The step of constructing a benefit matrix between workers and clusters based on the membership matrix, and using the maximum weight matching algorithm to determine the one-to-one mapping relationship between workers and task clusters, is as follows: First, construct the benefit matrix. Benefit matrix elements in Indicates that workers Assigned to cluster The expected benefits are calculated using the following formula: in, Membership matrix The elements in For fuzzy coefficients, For workers Execute the task The estimated time cost, This value is introduced to prevent the denominator from being 0. For spatiotemporal feasibility assessment function: in, Indicates the current time. Indicates task Expiration date Indicates workers The deadline for returning to the initial location. Indicates workers movement speed, Indicates task Location and workers The distance between the initial positions; Next, the benefit matrix Transform into a cost matrix Cost matrix elements in The calculation formula is as follows: The Hungarian algorithm is then used to solve the cost matrix to obtain the optimal matching scheme that maximizes the total benefit.

2. The mobile swarm sensing task allocation method based on fuzzy clustering and dynamic scheduling according to claim 1, characterized in that, The step involves using a fuzzy clustering algorithm to divide tasks with geographic location attributes into different cluster regions, obtaining the membership matrix of each task to each cluster and the cluster centers. Specific steps include: Initialize the membership matrix The cluster center set is updated by iteratively minimizing the objective function of the weighted sum of squared distances. and membership matrix Until convergence; Update the membership matrix elements in The calculation formula is: in, For the task location, and All are cluster centers. The total number of clusters, This is an index variable used to iterate through the cluster centers; Update cluster center The calculation formula is: in, The total number of tasks. This is an index variable used for traversing tasks; The iteration stops when the change in the membership matrix is ​​less than the convergence threshold ε or the maximum number of iterations is reached, and the final membership matrix is ​​output. and cluster center set .

3. The mobile swarm sensing task allocation method based on fuzzy clustering and dynamic scheduling according to claim 1, characterized in that, In the hybrid priority scheduling phase, each worker Maintain the primary task pool and collaborative task pool : The primary task pool Includes the clusters responsible for workers in all unassigned tasks. The task with the highest membership level; The collaborative task pool Includes clustering in all unassigned tasks Its membership degree is not the highest, but it exceeds the preset collaboration threshold. The task; When workers select tasks from a candidate list consisting of a primary task pool and a collaborative task pool, they do so based on a hybrid priority metric. Sort and combine priority metrics Based on time priority Space priority Membership priority Weighted summation yields: in Calculated based on the remaining effective time of the task. For the task The remaining valid time, and These represent the maximum and minimum remaining valid time for all tasks in the task pool, respectively. It is a value introduced to prevent the denominator from being 0; Based on the distance calculation between the task and the worker, This represents the straight-line distance between two locations. Indicates workers Location, For the task location, and Each task and worker is represented in the task pool. Distance between The maximum and minimum values ​​in; Based on task Clustering membership degree calculate; These are the weighting coefficients.

4. The mobile swarm sensing task allocation method based on fuzzy clustering and dynamic scheduling according to claim 3, characterized in that, The hybrid priority scheduling phase employs a dynamic scheduling strategy, and the specific execution steps are as follows: Step A1: When the worker When in an idle state, update workers based on the current set of unassigned tasks. The primary task pool and collaborative task pool And merge them into a candidate list; Step A2: Calculate the mixed priority for each task in the candidate list. , and according to The task is to sort the values ​​in ascending order. Step A3: Check the sorted task list sequentially, using the feasibility function. Judgment Task Does it satisfy the spatiotemporal constraints? Step A4: Select the first task in the list that satisfies the constraints as the next execution target, atomically update the status of the selected task to "assigned", and remove the selected task from the set of unassigned tasks.

5. The mobile swarm sensing task allocation method based on fuzzy clustering and dynamic scheduling according to claim 1, characterized in that, The overall task time resource cost in the optimization objective introduces a time penalty term for unassigned tasks. The total task time resource cost is the average time cost of all tasks, including both assigned and unassigned tasks. For tasks that have been successfully assigned, calculate the actual completion time cost of the assigned task. ; For tasks that are ultimately not assigned, calculate the penalty time cost for all unassigned tasks. The penalty time cost Defined as: in It represents the maximum straight-line distance between any two task locations among all unassigned tasks in the task region. Minimum moving speed for workers, This is the minimum processing speed for workers.

6. A mobile swarm intelligence sensing task allocation system based on fuzzy clustering and dynamic scheduling, characterized in that, It includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the mobile crowd sensing task allocation method as described in any one of claims 1-5.

Images