CN-122027581-A - Collaborative decision-making method and system in crowd sensing network based on artificial intelligence

CN122027581ACN 122027581 ACN122027581 ACN 122027581ACN-122027581-A

Abstract

The invention relates to the technical field of network decision, in particular to a collaborative decision method and a collaborative decision system in a crowd sensing network based on artificial intelligence, wherein the collaborative decision method comprises the steps of obtaining task characteristic information of each sensing terminal in the crowd sensing network and obtaining resource environment information of the crowd sensing network; and determining initial matching value evaluation values of all tasks in the current task set to be allocated aiming at different sensing terminals based on the task-resource matching evaluation model, and determining a target task allocation strategy according to the strategy values respectively corresponding to the current strategy value matrix and various task allocation strategies. The task-resource matching evaluation model is constructed, so that the matching degree between the task and the perception terminal can be accurately evaluated, task allocation is optimized, resource waste can be reduced, and the task completion efficiency and success rate are improved.

Inventors

ZHANG LI
Shi Fanfeng
WANG FUJUN
WANG LEI
ZUO HAO

Assignees

扬州工业职业技术学院

Dates

Publication Date: 20260512
Application Date: 20251222

Claims (10)

1. The collaborative decision-making method in the crowd sensing network based on artificial intelligence is characterized by comprising the following steps: task characteristic information of each sensing terminal in the crowd sensing network is obtained, and resource environment information of the crowd sensing network is obtained; Determining initial matching value evaluation values of each task in a current task set to be allocated aiming at different sensing terminals based on the task characteristic information and the resource environment information, and determining a target task allocation strategy according to the strategy values respectively corresponding to the current strategy value matrix and various task allocation strategies; Performing allocation operation on tasks in the current task set to be allocated by utilizing a target task allocation strategy and based on the initial matching value evaluation value to obtain a new current task allocation state; screening a collaborative execution sample from a new current task allocation state based on a task-resource matching evaluation model, and calculating a current collaborative execution rate based on the collaborative execution sample; updating the current strategy value matrix in response to the current cooperative execution rate being smaller than a preset cooperative threshold value, and returning to the step of determining initial matching value evaluation values of each task aiming at different sensing terminals in the current task set to be allocated based on the task-resource matching evaluation model; and determining the new current task allocation state as an optimal collaborative decision scheme to guide the crowd-sourced network to operate in response to the current collaborative execution rate not being smaller than a preset collaborative threshold.
2. The collaborative decision-making method in an artificial intelligence based crowd-sourced awareness network according to claim 1, wherein prior to constructing the task-resource matching assessment model, the method further comprises: Respectively distributing at least one task distribution strategy for each task in the initial task set based on a plurality of task distribution strategies; And reasonably distributing and simulating corresponding tasks by utilizing the task distribution strategy under the conditions of network delay, energy consumption and resource availability so as to initialize an initial task set, and determining the initialized task set as a current task set to be distributed.
3. The collaborative decision-making method in an artificial intelligence based crowd-sourced awareness network of claim 2 wherein the task characteristic information includes task complexity, task priority, and task relevance; based on the task characteristic information and the resource environment information, constructing a task-resource matching evaluation model, comprising: Carrying out complexity layering processing on the task characteristic information according to the task complexity in the task characteristic information to obtain task characteristic information with different complexity levels; carrying out availability classification processing on the resource environment information according to the resource availability in the resource environment information to obtain resource environment information with different availability grades; Searching and matching resource environment information with different availability levels by taking task characteristic information with different complexity levels as indexes, and determining a preliminary task-resource matching topology network; Sequencing tasks in the preliminary task-resource matching topological network according to the sequence of the task priorities from high to low to obtain a task-resource matching topological network with the sequenced priorities; Performing association optimization on the task-resource matching topology network with the ordered priorities based on the task association degree, and determining the optimized task-resource matching topology network; performing matching degree quantitative analysis on the optimized task-resource matching topological network, and determining a task-resource matching degree set; Acquiring a preset task-resource matching evaluation index set, and performing parameter adjustment on the evaluation index set based on the task-resource matching degree set to determine a target evaluation index; And constructing a task-resource matching evaluation model according to the target evaluation index, and dynamically evaluating the matching condition of the task and the resource by using the model.
4. The collaborative decision-making method in an artificial intelligence based crowd-sourced awareness network of claim 3 wherein retrieving and matching resource environment information of different availability levels with task characteristic information of different complexity levels as an index, determining a preliminary task-resource matching topology network, comprises: Extracting task characteristic information of a first complexity level and task characteristic information of a second complexity level from task characteristic information of different complexity levels, searching and matching according to the corresponding relation between task complexity and resource availability, and determining a first preliminary matching sub-topology network; Extracting task characteristic information of a third complexity level from task characteristic information of different complexity levels, searching and matching the task characteristic information with the first preliminary matching sub-topology network, and determining a second preliminary matching sub-topology network; and carrying out iterative search matching on task characteristic information of different complexity levels and resource environment information of different availability levels according to the second preliminary matching sub-topology network until matching of all the complexity levels and the availability levels is completed, and obtaining the preliminary task-resource matching topology network.
5. The collaborative decision-making method in an artificial intelligence based crowd-sourced aware network according to claim 4, wherein performing a matching degree quantitative analysis on an optimized task-sourced matching topology network to determine a set of task-sourced matching degrees comprises: Extracting a plurality of matching degree index values of a plurality of tasks and resource matching pairs in the optimized task-resource matching topological network; performing standardization processing on the plurality of matching degree index values, and determining the standardized matching degree index values; And taking the average value of the normalized matching degree index values as a quantized analysis starting point, and carrying out iterative analysis on a plurality of normalized matching degree index values according to a preset quantized bandwidth to determine a task-resource matching degree set.
6. The collaborative decision-making method in an artificial intelligence based crowd-sourced awareness network according to claim 5, wherein determining a target task allocation policy based on the current policy value matrix and policy values corresponding to the plurality of task allocation policies, respectively, comprises: Selecting a target task allocation strategy from the plurality of task allocation strategies based on the selection probabilities respectively corresponding to the plurality of task allocation strategies in the current strategy selection matrix, wherein the selection probabilities respectively corresponding to the plurality of task allocation strategies in the current strategy selection matrix are determined according to the strategy values respectively corresponding to the plurality of task allocation strategies in the current strategy value matrix; Wherein after updating the current policy value matrix, the method further comprises: And updating the current strategy selection matrix according to the updated current strategy value matrix.
7. The collaborative decision-making method in an artificial intelligence based crowd-sourced sensory network according to claim 6, wherein assigning tasks in a current set of tasks to be assigned using a target task assignment policy and based on an initial matching value evaluation value to obtain a new current task assignment state, comprising: dividing a task set with high matching value and a task set with low matching value from the task set to be allocated according to the initial matching value evaluation value; Performing allocation operation on the low-matching-value task set by utilizing a target task allocation strategy, and determining a new current task allocation state by combining the high-matching-value task set and the low-matching-value task set after the allocation operation; The method for distributing the low-matching-value task set by utilizing the target task distribution strategy comprises the following steps: determining the current dynamic adjustment probability based on initial matching value evaluation values respectively corresponding to all tasks in the low matching value task set; And performing task reassignment operation and/or task optimization assignment operation on the task set with low matching value according to the current dynamic adjustment probability by utilizing the target task assignment strategy.
8. The collaborative decision-making method in an artificial intelligence based crowd-sourced sensory network according to claim 7, wherein partitioning a set of high-and low-matching-value tasks from a set of current tasks to be allocated based on an initial matching-value evaluation value comprises: Selecting a plurality of tasks with minimum differences between the initial matching value evaluation value and the preset value demarcation value from the current task set to be allocated according to the preset high matching value proportion to form a high matching value task set; Selecting a plurality of tasks with maximum differences between the initial matching value evaluation value and the preset value demarcation value from the current task set to be allocated according to the preset low matching value proportion to form a third matching value task set; and eliminating tasks in the high-matching-value task set and the third matching-value task set from the current task set to be allocated, and forming a low-matching-value task set by the residual tasks.
9. The collaborative decision-making method in an artificial intelligence based crowd sensing network according to claim 8 wherein the collaborative execution samples are used to characterize a combination of tasks and sensing terminals where actual execution matches expected execution; Screening collaborative execution samples from a new current task allocation state based on a task-resource matching evaluation model, and calculating a current collaborative execution rate based on the collaborative execution samples, including: determining expected execution conditions respectively corresponding to each task and perception terminal combination in a new current task allocation state according to a task-resource matching evaluation model, and screening collaborative execution samples from the new current task allocation state according to the consistency of the expected execution conditions and actual execution conditions; and calculating the current cooperative execution rate corresponding to the new current task allocation state according to the total number of the samples of the cooperative execution samples.
10. An artificial intelligence based collaborative decision-making system in a crowd sensing network, which is adapted to the artificial intelligence based collaborative decision-making method in a crowd sensing network according to any one of claims 1-9, comprising: The data acquisition unit is used for acquiring task characteristic information of each sensing terminal in the crowd sensing network and acquiring resource environment information of the crowd sensing network; The system comprises a first allocation unit, a target task allocation strategy, a first allocation unit, a second allocation unit and a third allocation unit, wherein the first allocation unit is used for constructing a task-resource matching evaluation model based on task characteristic information and resource environment information; The second allocation unit is used for allocating the tasks in the current task set to be allocated by utilizing a target task allocation strategy and based on the initial matching value evaluation value to obtain a new current task allocation state; the task screening unit is used for screening a collaborative execution sample from the new current task allocation state based on the task-resource matching evaluation model, and calculating the current collaborative execution rate based on the collaborative execution sample; the secondary evaluation unit is used for updating the current strategy value matrix in response to the fact that the current cooperative execution rate is smaller than a preset cooperative threshold value, and returning to the step of determining initial matching value evaluation values of all tasks in the current task set to be distributed aiming at different sensing terminals based on the task-resource matching evaluation model; and the cooperative decision unit is used for determining the new current task allocation state as an optimal cooperative decision scheme to guide the crowd sensing network to operate in response to the fact that the current cooperative execution rate is not smaller than a preset cooperative threshold value.

Description

Collaborative decision-making method and system in crowd sensing network based on artificial intelligence Technical Field The invention relates to the technical field of network decision making, in particular to a collaborative decision making method and a collaborative decision making system in a crowd sensing network based on artificial intelligence. Background With the development of information technology and internet, the information volume in each field of society is rapidly increased, the acquisition, processing and analysis of data become more complex, and the conventional decision method is difficult to cope with challenges brought by mass data, so that the decision efficiency and accuracy are required to be improved by relying on artificial intelligence technology. At present, the traditional method uses a fixed evaluation model, can not update the matching degree between a task and a perception terminal in real time, so that when the resource environment or the task characteristics change, the matching degree can not be quickly adapted and adjusted, thereby influencing the task execution efficiency, and the task is often distributed by adopting a simple priority ordering or polling mode, so that the dynamic changes of the task complexity and the resource availability can not be considered, the resource is easy to idle or overload, and the efficiency of the whole system is reduced. In addition, the traditional method is generally focused on the distribution of a single task, ignores the cooperative action among a plurality of perception terminals, causes poor overall execution effect, cannot fully utilize the advantages of each terminal, depends on manual experience to make decisions, is easily influenced by subjective judgment of individuals, causes the decisions to be not scientific enough, and increases the risk of errors. Disclosure of Invention In order to achieve the above purpose, the invention provides the following technical scheme that the collaborative decision method in the crowd sensing network based on artificial intelligence comprises the following steps: task characteristic information of each sensing terminal in the crowd sensing network is obtained, and resource environment information of the crowd sensing network is obtained; Determining initial matching value evaluation values of each task in a current task set to be allocated aiming at different sensing terminals based on the task characteristic information and the resource environment information, and determining a target task allocation strategy according to the strategy values respectively corresponding to the current strategy value matrix and various task allocation strategies; Performing allocation operation on tasks in the current task set to be allocated by utilizing a target task allocation strategy and based on the initial matching value evaluation value to obtain a new current task allocation state; screening a collaborative execution sample from a new current task allocation state based on a task-resource matching evaluation model, and calculating a current collaborative execution rate based on the collaborative execution sample; updating the current strategy value matrix in response to the current cooperative execution rate being smaller than a preset cooperative threshold value, and returning to the step of determining initial matching value evaluation values of each task aiming at different sensing terminals in the current task set to be allocated based on the task-resource matching evaluation model; and determining the new current task allocation state as an optimal collaborative decision scheme to guide the crowd-sourced network to operate in response to the current collaborative execution rate not being smaller than a preset collaborative threshold. Preferably, before constructing the task-resource matching assessment model, the method further comprises: Respectively distributing at least one task distribution strategy for each task in the initial task set based on a plurality of task distribution strategies; And reasonably distributing and simulating corresponding tasks by utilizing the task distribution strategy under the conditions of network delay, energy consumption and resource availability so as to initialize an initial task set, and determining the initialized task set as a current task set to be distributed. Preferably, the task characteristic information comprises task complexity, task priority and task association degree, and the resource environment information comprises resource availability and resource performance; based on the task characteristic information and the resource environment information, constructing a task-resource matching evaluation model, comprising: Carrying out complexity layering processing on the task characteristic information according to the task complexity in the task characteristic information to obtain task characteristic information with different complexity levels; c