CN113159620A - Mine mobile crowd sensing task distribution method based on weighted undirected graph - Google Patents

Mine mobile crowd sensing task distribution method based on weighted undirected graph Download PDF

Info

Publication number
CN113159620A
CN113159620A CN202110512785.0A CN202110512785A CN113159620A CN 113159620 A CN113159620 A CN 113159620A CN 202110512785 A CN202110512785 A CN 202110512785A CN 113159620 A CN113159620 A CN 113159620A
Authority
CN
China
Prior art keywords
task
miners
miner
time
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110512785.0A
Other languages
Chinese (zh)
Other versions
CN113159620B (en
Inventor
江海峰
王梓蒙
肖硕
张宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China University of Mining and Technology CUMT
Original Assignee
China University of Mining and Technology CUMT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China University of Mining and Technology CUMT filed Critical China University of Mining and Technology CUMT
Priority to CN202110512785.0A priority Critical patent/CN113159620B/en
Publication of CN113159620A publication Critical patent/CN113159620A/en
Application granted granted Critical
Publication of CN113159620B publication Critical patent/CN113159620B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06313—Resource planning in a project environment
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Mining

Abstract

The mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph is suitable for mines covered with network signals, and for miners, the miners carry intelligent terminals which can be connected with the network, and the intelligent terminals can detect the heart rate and the downhole state of the miners; the task distribution method comprises the following steps: step 1: establishing a miner credibility model, and step 2: making a non-directional path diagram, and step 3: establishing a task set and a miner state information base, and step 4: emergency task allocation, step 5: non-urgent task allocation. The design not only adopts the weighted undirected graph to calculate the path range, but also classifies the tasks, and different types are distributed in different modes, thereby effectively improving the success rate of task distribution.

Description

Mine mobile crowd sensing task distribution method based on weighted undirected graph
Technical Field
The invention relates to a mine mobile crowd sensing task distribution method based on a weighted undirected graph, which is particularly suitable for intelligent distribution of underground tasks.
Background
The smart mobile device (mobile phone, tablet computer, etc.) can not only be used as a mobile device for daily communication, but also as a sensing unit because of its own embedded sensors, such as an acceleration sensor, a digital compass, a gyroscope Global Positioning System (GPS), a microphone, a camera, etc., sharing of sensed data becomes possible by using these sensors. Mobile crowd sensing is a new type of sensing network that uses sensors of the mobile devices of the ordinary users for sensing.
Currently, research on mobile crowd sensing is mainly focused on the following aspects: (1) research on application systems of crowd sensing in different fields such as cooperative positioning and environment detection; (2) and sensing task distribution. The perception task needs the participants of the perception task to complete, the participants are basic units for completing the perception task, different participants have different abilities, and the perception task is different in completion quality due to different degrees of aggressiveness. Participants are reasonably selected according to the perception task requirements, and task completion efficiency is improved, so that the establishment of a credibility evaluation model in a mobile crowd sensing system is particularly important.
China is a large coal resource country, and the energy structure mainly based on coal is difficult to change in a short time. The underground tunnel structure of the mine is complex, and the working environment of miners is severe. In recent years, with the development of smart phones and mobile communication technologies, how to combine mobile intelligent sensing with traditional coal mine safety monitoring is a research focus. The mine explosion-proof mobile phone taking the smart phone as a core is gradually popularized and applied in the field of underground safety monitoring. The mining explosion-proof mobile phone can realize the voice and video call, the dynamic gas inspection, the operation condition detection of electromechanical equipment and the integrated management of safe production information. In a mine with good western mining conditions and high informatization level, underground miners hold smart phones for coal safety certification in hands. Meanwhile, the intelligent information mining lamp with the stronger function is also popularized and applied in coal mines, and has the functions of environmental parameter detection, conversation, personnel positioning, camera shooting and the like, and also has the function of diagnosing the health state of miners. However, the existing applications based on the mining intelligent terminal are sensing oriented to individual terminals, group attributes of terminal carriers are not considered, large-scale and complex sensing tasks oriented to mine safety cannot be completed, and the intelligent mine construction target of ubiquitous sensing and intelligent decision-making cannot be realized.
With the improvement of the automation and informatization level of a coal mine, particularly the development of an internet of things technology, a wireless transmission technology and a mobile detection technology and the popularization and application in the coal industry, the promotion and the modification of the upgrading and the modification of the coal mine safety monitoring are the core research content of the current coal mine safety monitoring field by means of mobile perception, wireless transmission, downlink transmission of detection data, intelligent management and control and the like, and the mining intelligent mobile terminal taking a smart phone as a core is gradually popularized and applied in different fields of underground safety monitoring. According to the technical scheme, the design and application of a mine gas dynamic inspection and control system [ Lu-shining, mine gas dynamic inspection and control system [ J ]. coal science technology, 2018,46(08):125-129 ], aiming at the problems that manual inspection data cannot be uploaded in real time, data processing results cannot be fed back to the site in time and the like in gas inspection management, the mine gas dynamic inspection and control system based on a smart phone and a multi-parameter gas detector is designed and developed. The document [ Li Ming Jian, mining explosion-proof mobile phone data synchronization technology research and application [ J ]. computer application and software, 2018,35(2):74-79 ] is also applied to coal mine gas monitoring and early warning, and a database synchronization scheme based on Sync Framework is provided for the problem of bidirectional synchronization between an underground explosion-proof mobile phone-based explosion-proof information acquisition system and a ground information database, so that on-demand bidirectional synchronization of an explosion-proof dynamic information database is realized. The situation awareness realization technology research of a wisdom mine service system [ J ] computer research and development, 2014,51(12):2746 plus 2758 ] proposes modeling based on miner situation information, and provides timely and effective information service for miners to avoid potential safety hazards by utilizing various existing information system resources. The Wearable devices are classified in a document [ Mardonova M, Choi Y.review of Wearable Device Technology and Its Applications to the Mining Industry [ J ]. Energies,2018,11(3):547 ], and information acquisition is realized by matching with a smart phone through application of different types of Wearable devices, so that vital signs of miners can be sensed more comprehensively, and the safety of underground coal mine operation is enhanced. However, the existing research does not consider the group attribute of the terminal carrier and the cooperation opportunity brought by the terminal movement, and cannot fully utilize the ever-increasing scale effect of the underground intelligent mobile terminal, so that the large-scale, fine-grained and comprehensive and thorough perception requirements of the smart mine cannot be met. Therefore, mine safety monitoring and mobile crowd sensing need to be combined, underground miners and equipment carrying intelligent mobile terminals are used as sensing sources, sensing is carried out by utilizing wide distribution, flexible mobility and opportunity connectivity of the underground miners and the equipment, and a large-scale and complex sensing task facing the mine safety monitoring is completed.
The proper participant set in the mobile crowd sensing can ensure that the sensing task is efficiently and accurately completed, so a task distribution mechanism is always one of the key problems in the mobile crowd sensing research. The existing algorithm is as follows:
a heuristic algorithm HRA (empirical algorithm) introduces a reference value of the work capacity of the current worker, comprehensively studies the credit and the capability of the worker, the interest and the hobbies of the worker and the position discount function of the worker and the task, and improves the quality of the expected completed task to the maximum extent on the premise of controlling the budget.
The HAT algorithm (Hub-based multi-Task Assignment algorithm) has the main idea that a central node and corresponding subordinate nodes are screened out according to social attributes among the nodes, Task distribution of the subordinate nodes is determined when a Task distributor meets the central node, tasks are distributed by determining the central node and the subordinate nodes thereof, and the central node is responsible for distribution of the tasks and recovery of results.
The MARL algorithm (Multi-agent relevance task algorithm) is combined with the Q-Learning idea to optimize task allocation time according to the heterogeneity of different participants and tasks on the premise of meeting quality constraints, a real-world-oriented Multi-task allocation method based on Multi-subject reinforcement Learning is provided, and an optimization target is defined as pursuing shorter task allocation time on the premise of meeting quality constraints.
Mingchu Li et al propose a pareto optimal particle swarm optimization algorithm to solve the multi-task and multi-optimization target problem under the condition of limited number of people. Wang yangan et al have designed a heuristic MIA algorithm to solve the problem of coverage balancing user selection with budget constraints. Weiping Zhu et al propose a greedy discrete particle swarm optimization algorithm based on a genetic algorithm, fully consider the heterogeneity among users, and maximize the number of known completed tasks while satisfying certain constraints.
The task allocation method is mostly suitable for task allocation under the universal condition, and the actual condition of coal production is not considered. The method has the advantages that more participants are selected under the ubiquitous condition, the willingness of the participants is considered, whether the cost is over-paid or not is judged, the participants can subjectively select tasks which the participants want to participate in, factors of the participants and equipment do not need to be considered, if one participant is selected, the participant can refuse task allocation of a platform, the cost is minimized in the ubiquitous condition as the final aim, and the method has great difference with a task allocation strategy taking coal mine production as the background.
Disclosure of Invention
The invention aims to overcome the problem that the group attribute of a terminal carrier is not considered in task allocation in the prior art, and provides a mine mobile crowd sensing task distribution method based on a weighted undirected graph and distributed according to the group attribute of the terminal carrier.
In order to achieve the above purpose, the technical solution of the invention is as follows:
the mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph is suitable for mines covered with network signals, and for miners, the miners carry intelligent terminals which can be connected with the network, and the intelligent terminals can detect the heart rate and the downhole state of the miners;
the task distribution method comprises the following steps:
step 1: establishing a credibility model of the miners,
defining a participant's reputation as R (u)i) The credit degree is in the range of [0,1]]The credibility value is 0 to indicate that the miner is completely untrustworthy, the credibility value is 0.5 to indicate uncertainty, the credibility value is 1 to indicate complete credibility, and the initial credibility of the participant who participates in the perception task for the first time is 0.5;
according to the similarity of the task completion place, the similarity of the task completion time and the task completion qualityMiner credibility model R (u) established by three kinds of data informationi)；
Step 2: making an undirected path graph, taking each underground roadway as the edge of the undirected graph, taking the connection point of each chamber, the roadway and the roadway as the node of the undirected graph, and acquiring the moving speed of each miner
Figure BDA0003060985810000041
Dis (d)i,dj) Indicating mine adjacent node di,djThe weight w (d) between adjacent nodes is determined according to the undirected graph edge length and the miner speedi,dj),
Figure BDA0003060985810000042
w(di,dj) Represents a neighboring node di,djWeight in between;
and step 3: a task set and a miner state information base are established,
establishing a task set: editing known tasks to be completed into a task set M ═ M1,m2,m3…mnWhere m is the 4 sets of important information for a single task:
Figure BDA0003060985810000043
Figure BDA0003060985810000044
indicates the number of persons required for the task and
Figure BDA0003060985810000045
Figure BDA0003060985810000046
which represents the start time of the task,
Figure BDA0003060985810000047
which indicates the end time of the task,
Figure BDA0003060985810000048
indicating that the system predicted completion of the taskRequired time and
Figure BDA0003060985810000049
Figure BDA00030609858100000410
represents the type of task;
establishing a miner state information base: r participants are provided, and a miner state information base U ═ of the r participants is established1,u2,u3,…,ur) Wherein
Figure BDA00030609858100000411
R(ui) Expressed as a measure of the reputation of the mineworker,
Figure BDA00030609858100000412
expressed as the average moving speed of the miners and
Figure BDA00030609858100000413
(xi,yi) Indicating the miner uiThe position coordinates of the (c) and (d),
Figure BDA00030609858100000414
indicating the miner uiA processed historical task type;
and 4, step 4: the distribution of the urgent tasks is carried out,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after adding a task node, acquiring the type and position information of a task, and screening miners who can meet the requirement of solving the task from all the miners to acquire a candidate set of the miners according to the type of the task and whether the miners have engaged in similar tasks;
because a certain time is needed for completing the task, the time needed for completing the task is subtracted on the basis of the specified task weight of the system:
Figure BDA00030609858100000415
the weight of the final task is obtained; by passing throughThe task node is used as the center, the edge is diffused outwards along the communicated edge, the weight w of the passing edge is continuously superposed in the diffusion process, and the edge weight sum w from the task position node to the diffusion node is obtained in the diffusion process (d)i,dj)；
Edge weight sum w (d) of the path as it passesi,dj) Achieving task requirement weight wmjStopping diffusion, and acquiring a full path plan taking the task node as a center within a specified constraint time; then miners in the candidate set acquired by the miners meeting the task requirements in the path are ranked according to the credibility, the miners with the highest credibility are preferentially selected to execute the emergency task, and a task and a path diagram are sent to the intelligent equipment of the selected miners, and the emergency task allocation is completed at the moment;
and 5: the assignment of the non-urgent tasks is performed,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after the task node is added, acquiring the type and position information of the task, and screening out all miners which can finish the task of the task type in the underground miner set according to the task type and whether the miners engage in similar tasks; according to the weighted undirected graph, starting from a task node, diffusing outwards to obtain a place which can be reached by the task furthest in a task time range, and selecting miners meeting the task requirement in the place range as a candidate set of miners;
firstly, according to information fed back from an intelligent terminal carried by a miner, constructing a fatigue state model of the miner, wherein the fatigue state model is as follows:
Figure BDA0003060985810000051
in the above formula:
Figure BDA0003060985810000052
indicating the miner uiFatigue value of (2), unit: w;uibjindicating the miner uiHeart rate, unit: the times are/min;
Figure BDA0003060985810000053
indicating the miner uiHeight, unit: cm;
Figure BDA0003060985810000054
indicating the miner uiAge, unit: year;
Figure BDA0003060985810000055
indicating the miner uiLength of time to trip, unit: min;
in a fatigue state model
Figure BDA0003060985810000056
According to the acquired electric quantity of the miner portable intelligent terminal
Figure BDA0003060985810000057
A state value model related to miners is constructed, and the formula is as follows (2):
Figure BDA0003060985810000058
in the above formula: st (u)i) For the miner's real-time status value, kmaxMaximum value of real-time fatigue, k, among all minersminMinimum value of real-time fatigue among all miners, EmaxMaximum value of real-time electricity among all miners, EminIs the minimum value of real-time electricity quantity, alpha, in all miners1And alpha2In order to operate the coefficients of the operation,
Figure BDA0003060985810000059
α2＝1-α1and 0 is not more than alpha1≤1,0≤α2≤1；
After the state values of all miners are obtained, the state values are subjected to standardization treatment, wherein the formula is as follows (3):
Figure BDA00030609858100000510
wherein: st (u)i) Indicating the miner uiState value of (st)minMinimum value, st, representing the state of participation in the mineworkermaxMaximum value representing the state of participating miners;
in the miner state evaluation model ST (u)i) On the basis of the method, a miner credibility evaluation model R (u)i) The miner evaluation function RS (u) is constructedi) As in formula (4):
RS(ui)＝αR(ui)+βST(ui) (4)
wherein: alpha and beta are both arithmetic coefficients, 0 < alpha < 1,0 < beta < 1, alpha + beta is 1, RS (u)i) Representing a non-emergency task miner evaluation function;
miners in the candidate set which are obtained by miners meeting the task requirements in the path according to the non-emergency task miner evaluation index RS (u)i) And sequencing, preferentially selecting the miners with the non-emergency tasks to execute the non-emergency tasks, wherein the miners with the evaluation indexes of the non-emergency tasks are front, and sending the tasks and the path diagram to the intelligent equipment of the selected miners, wherein the non-emergency tasks are distributed completely.
The step 1: in the establishment of a model of the credibility of miners,
i, constructing a model of the similarity l of the task completion site, wherein the model is as shown in a formula (5):
Figure BDA0003060985810000061
in the above formula, ui(xi,yi),mj(xj,yj) Respectively representing the location vectors of the miners i and the tasks j, where xi,yiRespectively representing the vertical and horizontal coordinates, x, of the position at which the mineworker i completes the task and submits the dataj,yjThe ordinate and abscissa representing the position required by task j; delta x and delta y are coordinate position variables of task adjustment, and delta x is greater than 0 and delta y is greater than 0;
II, constructing a task completion time similarity p model as shown in a formula (6):
Figure BDA0003060985810000062
in the above formula, the first and second carbon atoms are,
Figure BDA0003060985810000063
for the time that the mineworker i starts the task j,
Figure BDA0003060985810000064
for the time that miner i finishes task j,
Figure BDA0003060985810000065
the start time of task j is required for the system,
Figure BDA0003060985810000066
the end time of task j is required for the system,
Figure BDA0003060985810000067
indicating the degree of difference in user initiation from the initiation of the task request,
Figure BDA0003060985810000068
the difference degree of the latest end time of the user and the end time required by the task is represented, and n represents the difference degree interval divided by miners;
calculating the difference degree between the starting time of the perception user and the starting time of the perception task:
the time difference value of the user participating in the task can be used for measuring the difference degree of the perception user starting to execute the perception task; first, the start time of the task will be perceived
Figure BDA0003060985810000071
And start time of each perceived user
Figure BDA0003060985810000072
Converting into numerical data; then, compute the perception task and use for different perceptionAbsolute difference of start time between households
Figure BDA0003060985810000073
The minimum and maximum difference in absolute difference in start time between a perception task and different perception users is [ theta ]1,θn]This interval is divided on average into n-1 equidistant cells: { [ theta ] theta1,θ2],[θ2,θ3],·,[θn-1,θn]After the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls in a certain cell, sequentially giving a difference degree value {0, 1, …, n-1, n } to each cell to obtain a difference degree value corresponding to the cell in which the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls;
ii, obtaining a difference degree value of the perception user and the perception task deadline by adopting a calculation mode with the same difference degree of the perception user and the perception task starting time;
iii, adding the difference degree of the perception user and the perception task time and the perception user and the perception task deadline time, and dividing the sum by the difference interval to obtain the time difference degree of the miner i to the task j through calculation;
III, constructing a model of data quality reliability q, as shown in formula (8):
Figure BDA0003060985810000074
Figure BDA0003060985810000075
in the above formula:
Figure BDA0003060985810000076
mean value of values, ds (u), representing task ji,mj) Value, ds (u), representing the miner upload task ji,mj)maxRepresents the maximum value of the uploaded data of all the participating miners for completing the task j,
Figure BDA0003060985810000077
total number of people participating in task j;
when the numerical value provided by the miners is closer to the real data, the uploaded data is higher in reliability, and the data quality reliability formula q (u) isi,mj) The closer to 1 the middle expression is, and otherwise, the closer to 0 the middle expression is, the data acquired by the miners are far away from the real data, namely, the deviation from the real data is larger, and the reliability is lower;
IV, constructing a miner credibility model, wherein the final classification result of miners in the model is that the dependent variable y is 1 when the credibility is more than 0.5 or the dependent variable y is 0 when the credibility is less than 0.5, so that the value of the dependent variable y in the model is 0 or 1, the dependent variable y accords with the 0 and 1 distribution, and determining coefficients of three independent variable factors by adopting a logistic regression method to establish the credibility model, wherein the formula (9) shows that:
Figure BDA0003060985810000081
wherein: (x) is a logical function of (f),
Figure BDA0003060985810000082
c(ui,mj) The credibility of the miner i participating in the task j is obtained;
let x be beta0+β1·l(ui,mj)+β2·p(ui,mj)+β3·q(ui,mj) (10)
Substituting conversion and logarithm are taken into the formula (9) to obtain a single task credibility logistic regression model:
Figure BDA0003060985810000083
calculating according to a logistic regression solution method, and enabling E (y)i＝1)＝c(ui,mj) Using maximum likelihoodEstimating, namely calculating task time coefficients beta of different tasks under the mine by introducing three independent variables2Task location coefficient beta1And a task quality coefficient beta3And a constant value of beta0The value of (d);
in order to avoid that the credibility is influenced by the difference of the number of the miners participating in the tasks, the total average value processing is carried out on all the miners credibility, so that the miners credibility formula is as shown in formula (12):
Figure BDA0003060985810000084
where k represents the perceptual user uiTotal number of historically completed tasks, R (u)i) Representing the reputation of the miners.
The step 3: establishing a task set and a miner state information base, and aiming at the task type, adopting Gray code to the task type
Figure BDA0003060985810000085
Dividing, carrying out condition constraint according to the requirement of the perception task on perception time, and drawing a task type table according to the requirement of the task type, wherein the table is as follows:
Figure BDA0003060985810000086
miners with different functions can complete different types of tasks, and the system selects miners with corresponding functions to complete the tasks with corresponding types according to the task type codes.
Compared with the prior art, the invention has the beneficial effects that:
1. the mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph is characterized in that a mine is used as a background, research is carried out aiming at the real-time task distribution problem under the mine, the tasks are divided into urgent and non-urgent tasks according to the mine working environment and the actual production requirements, and different algorithms are adopted for distributing the tasks according to the difference of the urgent degrees. Such an allocation method takes into account the actual situation of the individual tasks. Therefore, the task allocation of the design is more reasonable, and the success rate of the task allocation is effectively improved.
2. The mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph constructs a miner credibility model related to the position of miners, the task time and the task quality, designs and provides an emergency task distribution algorithm based on the weighted undirected graph by taking time as constraint according to the emergency degree of tasks, so that the miners can be matched with the miners which can finish the tasks relatively efficiently by using smaller time overhead, and a new thought is provided for underground task distribution. Therefore, the design utilizes the undirected weighted algorithm to match miners who finish tasks relatively efficiently with less time overhead, and the task allocation time is shortened.
3. According to the mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph, a miner state value model is established according to the workload of miners and the electric quantity of portable equipment, a non-emergency task distribution algorithm is provided, and the aim of considering more relevant factors of the miners and the equipment thereof in the non-emergency task state is achieved. Therefore, the design considers the equipment condition of miners more, and further improves the probability of completing the distribution task.
4. According to the mine mobile crowd sensing task distribution method based on the weighted undirected graph, the mobile crowd sensing utilizes the existing sensing equipment and the deployed communication network, and special deployment is not needed, so that the deployment and maintenance cost is greatly reduced. Meanwhile, as the user carries the mobile terminal such as a mobile phone and the like, the sensing coverage and sensing data transmission can be promoted, the sensing with fine granularity is realized, and more crowd sensing applications are to be mined and explored. Along with the improvement of the informatization and intellectualization level of the coal mine, the configuration scale of the intelligent mobile terminal for the coal mine is continuously enlarged, and the sensing, calculation and communication capabilities of the terminal are continuously enhanced. The combination of mobile crowd sensing and mine safety monitoring can further excavate the sensing potential of the existing mine intelligent mobile terminal, and expand the crowd sensing application based on the intelligent mobile terminal, thereby realizing large-scale and complex sensing tasks. Therefore, the design can be suitable for most underground network environments, and the application range is wide.
Drawings
FIG. 1 is a schematic diagram of a coal mine underground network according to the present invention.
FIG. 2 is a task location similarity diagram of the present invention.
Fig. 3 is a flow chart of the weighted undirected graph algorithm of the present invention.
FIG. 4 is a comparison of the average task allocation times of the four algorithms in example 3.
FIG. 5 is a comparison chart of the average reputation of miners in the four algorithms of example 3.
FIG. 6 is a graph comparing the four algorithms of Miner population versus Emergency assignment power in example 3.
Fig. 7 is a comparison graph of the task number of the four algorithms and the success rate of the emergency task allocation in example 3.
FIG. 8 is a graph comparing the number of miners versus the success rate of non-urgent task assignment for the four algorithms of example 3.
Fig. 9 is a graph comparing the task number of the four algorithms with the success rate of non-urgent task allocation in example 3.
FIG. 10 is a comparison chart of the average state values of four algorithms participating in miners in example 3.
Fig. 11 is a comparison graph of the average values of the remaining capacities of the devices of the selected participants in the four algorithms of example 3.
In the figure: MARL stands for MARL algorithm; HTA denotes HTA algorithm; HRA denotes HRA algorithm; E-WGA represents the E-WGA algorithm of the invention; allocation time represents allocation time; number of Participants indicates the Number of Participants; average reputations represent Average reputations; number of miners indicates the Number of miners; the Allocation success rate represents the Allocation success rate; number of tasks represents the Number of tasks; average state value represents an Average state value; the Average power of the equipment is expressed as the Average capacity of the equipment.
Detailed Description
The present invention will be described in further detail with reference to the following description and embodiments in conjunction with the accompanying drawings.
Referring to fig. 1 to 3, a mine mobile crowd sensing task distribution method based on a weighted undirected graph is suitable for mines covered with network signals, and for miners who all carry intelligent terminals capable of being connected with a network, the intelligent terminals can detect the heart rate and the downhole state of the miners;
the task distribution method comprises the following steps:
step 1: establishing a credibility model of the miners,
defining a participant's reputation as R (u)i) The credit degree is in the range of [0,1]]The credibility value is 0 to indicate that the miner is completely untrustworthy, the credibility value is 0.5 to indicate uncertainty, the credibility value is 1 to indicate complete credibility, and the initial credibility of the participant who participates in the perception task for the first time is 0.5;
establishing a miner credibility model R (u) according to three data information of task completion place similarity, task completion time similarity and task completion qualityi)；
Step 2: making an undirected path graph, taking each underground roadway as the edge of the undirected graph, taking the connection point of each chamber, the roadway and the roadway as the node of the undirected graph, and acquiring the moving speed of each miner
Figure BDA0003060985810000101
Dis (d)i,dj) Indicating mine adjacent node di,djThe weight w (d) between adjacent nodes is determined according to the undirected graph edge length and the miner speedi,dj),
Figure BDA0003060985810000111
w(di,dj) Represents a neighboring node di,djWeight in between;
and step 3: a task set and a miner state information base are established,
establishing a task set: editing known tasks to be completed into a task set M ═ M1,m2,m3…mnWhere m is the 4 sets of important information for a single task:
Figure BDA0003060985810000112
Figure BDA0003060985810000113
indicates the number of persons required for the task and
Figure BDA0003060985810000114
Figure BDA0003060985810000115
which represents the start time of the task,
Figure BDA0003060985810000116
which indicates the end time of the task,
Figure BDA0003060985810000117
indicating that the system predicted the time required to complete the task and
Figure BDA0003060985810000118
Figure BDA0003060985810000119
represents the type of task;
establishing a miner state information base: r participants are provided, and a miner state information base U ═ of the r participants is established1,u2,u3,...,ur) Wherein
Figure BDA00030609858100001110
R(ui) Expressed as a measure of the reputation of the mineworker,
Figure BDA00030609858100001111
expressed as the average moving speed of the miners and
Figure BDA00030609858100001112
(xi,yi) Indicating the miner uiThe position coordinates of the (c) and (d),
Figure BDA00030609858100001113
indicating the miner uiA processed historical task type;
and 4, step 4: the distribution of the urgent tasks is carried out,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after adding a task node, acquiring the type and position information of a task, and screening miners who can meet the requirement of solving the task from all the miners to acquire a candidate set of the miners according to the type of the task and whether the miners have engaged in similar tasks;
because a certain time is needed for completing the task, the time needed for completing the task is subtracted on the basis of the specified task weight of the system:
Figure BDA00030609858100001114
the weight of the final task is obtained; the total weight w (d) of the edges from the task position node to the diffusion node is obtained in the diffusion process by taking the task node as the center, diffusing outwards along the communicated edges and continuously superposing the weight w of the passing edges in the diffusion processi,dj)；
Edge weight sum w (d) of the path as it passesi,dj) Achieving task requirement weight wmjStopping diffusion, and acquiring a full path plan taking the task node as a center within a specified constraint time; then miners in the candidate set acquired by the miners meeting the task requirements in the path are ranked according to the credibility, the miners with the highest credibility are preferentially selected to execute the emergency task, and a task and a path diagram are sent to the intelligent equipment of the selected miners, and the emergency task allocation is completed at the moment;
and 5: the assignment of the non-urgent tasks is performed,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after the task node is added, acquiring the type and position information of the task, and screening out all miners which can finish the task of the task type in the underground miner set according to the task type and whether the miners engage in similar tasks; according to the weighted undirected graph, starting from a task node, diffusing outwards to obtain a place which can be reached by the task furthest in a task time range, and selecting miners meeting the task requirement in the place range as a candidate set of miners;
firstly, according to information fed back from an intelligent terminal carried by a miner, constructing a fatigue state model of the miner, wherein the fatigue state model is as follows:
Figure BDA0003060985810000121
in the above formula:
Figure BDA0003060985810000122
indicating the miner uiFatigue value of (2), unit: w; u. ofibjIndicating the miner uiHeart rate, unit: the times are/min;
Figure BDA0003060985810000123
indicating the miner uiHeight, unit: cm;
Figure BDA0003060985810000124
indicating the miner uiAge, unit: year;
Figure BDA0003060985810000125
indicating the miner uiLength of time to trip, unit: min;
in a fatigue state model
Figure BDA0003060985810000126
According to the acquired electric quantity of the miner portable intelligent terminal
Figure BDA0003060985810000127
A state value model related to miners is constructed, and the formula is as follows (2):
Figure BDA0003060985810000128
in the above formula: st (u)i) For the miner's real-time status value, kmaxMaximum value of real-time fatigue, k, among all minersminMinimum value of real-time fatigue among all miners, EmaxMaximum value of real-time electricity among all miners, EminIs the minimum value of real-time electricity quantity, alpha, in all miners1And alpha2In order to operate the coefficients of the operation,
Figure BDA0003060985810000129
α2＝1-α1and 0 is not more than alpha1≤1,0≤α2≤1；
After the state values of all miners are obtained, the state values are subjected to standardization treatment, wherein the formula is as follows (3):
Figure BDA00030609858100001210
wherein: st (u)i) Indicating the miner uiState value of (st)minMinimum value, st, representing the state of participation in the mineworkermaxMaximum value representing the state of participating miners;
in the miner state evaluation model ST (u)i) On the basis of the method, a miner credibility evaluation model R (u)i) The miner evaluation function RS (u) is constructedi) As in formula (4):
RS(ui)＝αR(ui)+βST(ui) (4)
wherein: alpha and beta are both arithmetic coefficients, 0 < alpha < 1,0 < beta < 1, alpha + beta is 1, RS (u)i) Representing a non-emergency task miner evaluation function;
miners in the candidate set which are obtained by miners meeting the task requirements in the path according to the non-emergency task miner evaluation index RS (u)i) And sequencing, preferentially selecting the miners with the non-emergency tasks to execute the non-emergency tasks, wherein the miners with the evaluation indexes of the non-emergency tasks are front, and sending the tasks and the path diagram to the intelligent equipment of the selected miners, wherein the non-emergency tasks are distributed completely.
The step 1: in the establishment of a model of the credibility of miners,
i, constructing a model of the similarity l of the task completion site, wherein the model is as shown in a formula (5):
Figure BDA0003060985810000131
in the above formula, ui(xi,yi),mj(xj,yj) Respectively representing the location vectors of the miners i and the tasks j, where xi,yiRespectively representing the vertical and horizontal coordinates, x, of the position at which the mineworker i completes the task and submits the dataj,yjThe ordinate and abscissa representing the position required by task j; delta x and delta y are coordinate position variables of task adjustment, and delta x is greater than 0 and delta y is greater than 0;
II, constructing a task completion time similarity p model as shown in a formula (6):
Figure BDA0003060985810000132
in the above formula, the first and second carbon atoms are,
Figure BDA0003060985810000133
for the time that the mineworker i starts the task j,
Figure BDA0003060985810000134
for the time that miner i finishes task j,
Figure BDA0003060985810000135
the start time of task j is required for the system,
Figure BDA0003060985810000136
the end time of task j is required for the system,
Figure BDA0003060985810000137
indicating the degree of difference in user initiation from the initiation of the task request,
Figure BDA0003060985810000138
the difference degree of the latest end time of the user and the end time required by the task is represented, and n represents the difference degree interval divided by miners;
calculating the difference degree between the starting time of the perception user and the starting time of the perception task:
the time difference value of the user participating in the task can be used for measuring the difference degree of the perception user starting to execute the perception task; first, the start time of the task will be perceived
Figure BDA0003060985810000139
And start time of each perceived user
Figure BDA00030609858100001310
Converting into numerical data; then, the absolute difference of the starting time between the perception task and the different perception users is calculated
Figure BDA00030609858100001311
The minimum and maximum difference in absolute difference in start time between a perception task and different perception users is [ theta ]1,θn]This interval is divided on average into n-1 equidistant cells: { [ theta ] theta1,θ2],[θ2,θ3],...,[θn-1,θn]After the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls in a certain cell, sequentially giving a difference degree value {0, 1, …, n-1, n } to each cell to obtain a difference degree value corresponding to the cell in which the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls;
ii, obtaining a difference degree value of the perception user and the perception task deadline by adopting a calculation mode with the same difference degree of the perception user and the perception task starting time;
iii, adding the difference degree of the perception user and the perception task time and the perception user and the perception task deadline time, and dividing the sum by the difference interval to obtain the time difference degree of the miner i to the task j through calculation;
III, constructing a model of data quality reliability q, as shown in formula (8):
Figure BDA0003060985810000141
Figure BDA0003060985810000142
in the above formula:
Figure BDA0003060985810000143
mean value of values, ds (u), representing task ji,mj) Value, ds (u), representing the miner upload task ji,mj)maxRepresents the maximum value of the uploaded data of all the participating miners for completing the task j,
Figure BDA0003060985810000144
total number of people participating in task j;
when the numerical value provided by the miners is closer to the real data, the uploaded data is higher in reliability, and the data quality reliability formula q (u) isi,mj) The closer to 1 the middle expression is, and otherwise, the closer to 0 the middle expression is, the data acquired by the miners are far away from the real data, namely, the deviation from the real data is larger, and the reliability is lower;
IV, constructing a miner credibility model, wherein the final classification result of miners in the model is that the dependent variable y is 1 when the credibility is more than 0.5 or the dependent variable y is 0 when the credibility is less than 0.5, so that the value of the dependent variable y in the model is 0 or 1, the dependent variable y accords with the 0 and 1 distribution, and determining coefficients of three independent variable factors by adopting a logistic regression method to establish the credibility model, wherein the formula (9) shows that:
Figure BDA0003060985810000145
wherein: (x) is a logical function of (f),
Figure BDA0003060985810000151
c(ui,mj) The credibility of the miner i participating in the task j is obtained;
let x be beta0+β1·l(ui,mj)+β2·p(ui,mj)+β3·q(ui,mj) (10)
Substituting conversion and logarithm are taken into the formula (9) to obtain a single task credibility logistic regression model:
Figure BDA0003060985810000152
calculating according to a logistic regression solution method, and enabling E (y)i＝1)＝c(ui,mj) By utilizing maximum likelihood estimation, the task time coefficients beta of different tasks under the mine can be calculated by introducing three independent variables2Task location coefficient beta1And a task quality coefficient beta3And a constant value of beta0The value of (d);
in order to avoid that the credibility is influenced by the difference of the number of the miners participating in the tasks, the total average value processing is carried out on all the miners credibility, so that the miners credibility formula is as shown in formula (12):
Figure BDA0003060985810000153
where k represents the perceptual user uiTotal number of historically completed tasks, R (u)i) Representing the reputation of the miners.
The step 3: establishing a task set and a miner state information base, and aiming at the task type, adopting Gray code to the task type
Figure BDA0003060985810000154
Dividing, carrying out condition constraint according to the requirement of the perception task on perception time, and drawing a task type table according to the requirement of the task type, wherein the table is as follows:
Figure BDA0003060985810000155
miners with different functions can complete different types of tasks, and the system selects miners with corresponding functions to complete the tasks with corresponding types according to the task type codes.
The principle of the invention is illustrated as follows:
MARL algorithm: the Multi-agent discovery task algorithm.
HTA algorithm: the Hub-based multi-Task Assignment algorithm.
HRA algorithm: the Heuristic algorithm.
E-WGA algorithm: the Emergency Weighted undirected Graph Algorithm.
The time difference value of the user participating in the task can be used for measuring the difference degree of the perception user starting to execute the perception task, see: ronghui, fire xu, huchunhua, morningman a collaborative filtering recommendation algorithm based on user similarity [ J ] news bulletins, 2014, 35(02):16-24.
The fatigue degree value of the self state of the miner is shown as follows: the research on the influence of the exercise load experiment on heart rate by the juchen juan, Wang Xiu lan, Liu Weidong, Miner [ J ]. labor protection science and technology, (1997 (06):52-53+ 57).
Example 1:
the mine mobile crowd-sourcing perception task distribution method based on the weighted undirected graph is suitable for mines covered with network signals, and for miners, the miners carry intelligent terminals which can be connected with the network, and the intelligent terminals can detect the heart rate and the downhole state of the miners;
the task distribution method comprises the following steps:
step 1: establishing a credibility model of the miners,
defining a participant's reputation as R (u)i) The credit degree is in the range of [0,1]]The credibility value of 0 indicates that the miner is completely untrustworthy, the credibility value of 0.5 indicates uncertainty and credibilityThe value of 1 represents complete credibility, and the initial credit degree of a participant who participates in the perception task for the first time is 0.5;
establishing a miner credibility model R (u) according to three data information of task completion place similarity, task completion time similarity and task completion qualityi)；
Step 2: making an undirected path graph, taking each underground roadway as the edge of the undirected graph, taking the connection point of each chamber, the roadway and the roadway as the node of the undirected graph, and acquiring the moving speed of each miner
Figure BDA0003060985810000161
Dis (d)i,dj) Indicating mine adjacent node di,djThe weight w (d) between adjacent nodes is determined according to the undirected graph edge length and the miner speedi,dj),
Figure BDA0003060985810000162
w(di,dj) Represents a neighboring node di,djWeight in between;
and step 3: a task set and a miner state information base are established,
establishing a task set: editing known tasks to be completed into a task set M ═ M1,m2,m3…mnWhere m is the 4 sets of important information for a single task:
Figure BDA0003060985810000163
Figure BDA0003060985810000164
indicates the number of persons required for the task and
Figure BDA0003060985810000165
which represents the start time of the task,
Figure BDA0003060985810000166
which indicates the end time of the task,
Figure BDA0003060985810000167
indicating that the system predicted the time required to complete the task and
Figure BDA0003060985810000171
Figure BDA0003060985810000172
represents the type of task;
establishing a miner state information base: r participants are provided, and a miner state information base U ═ of the r participants is established1,u2,u3,...,ur) Wherein
Figure BDA0003060985810000173
R(ui) Expressed as a measure of the reputation of the mineworker,
Figure BDA0003060985810000174
expressed as the average moving speed of the miners and
Figure BDA0003060985810000175
(xi,yi) Indicating the miner uiThe position coordinates of the (c) and (d),
Figure BDA0003060985810000176
indicating the miner uiA processed historical task type;
and 4, step 4: the distribution of the urgent tasks is carried out,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after adding a task node, acquiring the type and position information of a task, and screening miners who can meet the requirement of solving the task from all the miners to acquire a candidate set of the miners according to the type of the task and whether the miners have engaged in similar tasks;
because a certain time is needed for completing the task, the time needed for completing the task is subtracted on the basis of the specified task weight of the system:
Figure BDA0003060985810000177
the weight of the final task is obtained; the total weight w (d) of the edges from the task position node to the diffusion node is obtained in the diffusion process by taking the task node as the center, diffusing outwards along the communicated edges and continuously superposing the weight w of the passing edges in the diffusion processi,dj)；
Edge weight sum w (d) of the path as it passesi,dj) Achieving task requirement weight wmjStopping diffusion, and acquiring a full path plan taking the task node as a center within a specified constraint time; then miners in the candidate set acquired by the miners meeting the task requirements in the path are ranked according to the credibility, the miners with the highest credibility are preferentially selected to execute the emergency task, and a task and a path diagram are sent to the intelligent equipment of the selected miners, and the emergency task allocation is completed at the moment;
and 5: the assignment of the non-urgent tasks is performed,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after the task node is added, acquiring the type and position information of the task, and screening out all miners which can finish the task of the task type in the underground miner set according to the task type and whether the miners engage in similar tasks; according to the weighted undirected graph, starting from a task node, diffusing outwards to obtain a place which can be reached by the task furthest in a task time range, and selecting miners meeting the task requirement in the place range as a candidate set of miners;
firstly, according to information fed back from an intelligent terminal carried by a miner, constructing a fatigue state model of the miner, wherein the fatigue state model is as follows:
Figure BDA0003060985810000178
in the above formula:
Figure BDA0003060985810000181
indicating the miner uiFatigue value of (2), unit: w; u. ofibjIndicating the miner uiHeart rate, unit: the times are/min;
Figure BDA0003060985810000182
indicating the miner uiHeight, unit: cm;
Figure BDA0003060985810000183
indicating the miner uiAge, unit: year;
Figure BDA0003060985810000184
indicating the miner uiLength of time to trip, unit: min;
in a fatigue state model
Figure BDA0003060985810000185
According to the acquired electric quantity of the miner portable intelligent terminal
Figure BDA0003060985810000186
A state value model related to miners is constructed, and the formula is as follows (2):
Figure BDA0003060985810000187
in the above formula: st (u)i) For the miner's real-time status value, kmaxMaximum value of real-time fatigue, k, among all minersminMinimum value of real-time fatigue among all miners, EmaxMaximum value of real-time electricity among all miners, EminIs the minimum value of real-time electricity quantity, alpha, in all miners1And alpha2In order to operate the coefficients of the operation,
Figure BDA0003060985810000188
α2＝1-α1and 0 is not more than alpha1≤1,0≤α2≤1；
After the state values of all miners are obtained, the state values are subjected to standardization treatment, wherein the formula is as follows (3):
Figure BDA0003060985810000189
wherein: st (u)i) Indicating the miner uiState value of (st)minMinimum value, st, representing the state of participation in the mineworkermaxMaximum value representing the state of participating miners;
in the miner state evaluation model ST (u)i) On the basis of the method, a miner credibility evaluation model R (u)i) The miner evaluation function RS (u) is constructedi) As in formula (4):
RS(ui)＝αR(ui)+βST(ui) (4)
wherein: alpha and beta are both arithmetic coefficients, 0 < alpha < 1,0 < beta < 1, alpha + beta is 1, RS (u)i) Representing a non-emergency task miner evaluation function;
miners in the candidate set which are obtained by miners meeting the task requirements in the path according to the non-emergency task miner evaluation index RS (u)i) And sequencing, preferentially selecting the miners with the non-emergency tasks to execute the non-emergency tasks, wherein the miners with the evaluation indexes of the non-emergency tasks are front, and sending the tasks and the path diagram to the intelligent equipment of the selected miners, wherein the non-emergency tasks are distributed completely.
The step 1: in the establishment of a model of the credibility of miners,
i, constructing a model of the similarity l of the task completion site, wherein the model is as shown in a formula (5):
Figure BDA0003060985810000191
in the above formula, ui(xi,yi),mj(xj,yj) Respectively representing the location vectors of the miners i and the tasks j, where xi,yiRespectively representing the vertical and horizontal coordinates, x, of the position at which the mineworker i completes the task and submits the dataj,yjThe ordinate and abscissa representing the position required by task j; Δ x, Δ y are coordinates of task adjustmentPosition variable, and Δ x > 0, Δ y > 0;
II, constructing a task completion time similarity p model as shown in a formula (6):
Figure BDA0003060985810000192
in the above formula, the first and second carbon atoms are,
Figure BDA0003060985810000193
for the time that the mineworker i starts the task j,
Figure BDA0003060985810000194
for the time that miner i finishes task j,
Figure BDA0003060985810000195
the start time of task j is required for the system,
Figure BDA0003060985810000196
the end time of task j is required for the system,
Figure BDA0003060985810000197
indicating the degree of difference in user initiation from the initiation of the task request,
Figure BDA0003060985810000198
the difference degree of the latest end time of the user and the end time required by the task is represented, and n represents the difference degree interval divided by miners;
calculating the difference degree between the starting time of the perception user and the starting time of the perception task:
the time difference value of the user participating in the task can be used for measuring the difference degree of the perception user starting to execute the perception task; first, the start time of the task will be perceived
Figure BDA0003060985810000199
And start time of each perceived user
Figure BDA00030609858100001910
Converting into numerical data; then, the absolute difference of the starting time between the perception task and the different perception users is calculated
Figure BDA00030609858100001911
The minimum and maximum difference in absolute difference in start time between a perception task and different perception users is [ theta ]1,θn]This interval is divided on average into n-1 equidistant cells: { [ theta ] theta1,θ2],[θ2,θ3],...,[θn-1,θn]After the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls in a certain cell, sequentially giving a difference degree value {0, 1, …, n-1, n } to each cell to obtain a difference degree value corresponding to the cell in which the absolute difference value between the task starting time of the sensing user and the task requiring starting time of the sensing task falls;
ii, obtaining a difference degree value of the perception user and the perception task deadline by adopting a calculation mode with the same difference degree of the perception user and the perception task starting time;
iii, adding the difference degree of the perception user and the perception task time and the perception user and the perception task deadline time, and dividing the sum by the difference interval to obtain the time difference degree of the miner i to the task j through calculation;
III, constructing a model of data quality reliability q, as shown in formula (8):
Figure BDA0003060985810000201
Figure BDA0003060985810000202
in the above formula:
Figure BDA0003060985810000203
mean value of values, ds (u), representing task ji,mj) Value, ds (u), representing the miner upload task ji,mj)maxRepresents the maximum value of the uploaded data of all the participating miners for completing the task j,
Figure BDA0003060985810000204
total number of people participating in task j;
when the numerical value provided by the miners is closer to the real data, the uploaded data is higher in reliability, and the data quality reliability formula q (u) isi,mj) The closer to 1 the middle expression is, and otherwise, the closer to 0 the middle expression is, the data acquired by the miners are far away from the real data, namely, the deviation from the real data is larger, and the reliability is lower;
IV, constructing a miner credibility model, wherein the final classification result of miners in the model is that the dependent variable y is 1 when the credibility is more than 0.5 or the dependent variable y is 0 when the credibility is less than 0.5, so that the value of the dependent variable y in the model is 0 or 1, the dependent variable y accords with the 0 and 1 distribution, and determining coefficients of three independent variable factors by adopting a logistic regression method to establish the credibility model, wherein the formula (9) shows that:
Figure BDA0003060985810000205
wherein: (x) is a logical function of (f),
Figure BDA0003060985810000206
c(ui,mj) The credibility of the miner i participating in the task j is obtained;
let x be beta0+β1·l(ui,mj)+β2·p(ui,mj)+β3·q(ui,mj) (10)
Substituting conversion and logarithm are taken into the formula (9) to obtain a single task credibility logistic regression model:
Figure BDA0003060985810000207
by logistic regressionCalculating a solution method, let E (y)i＝1)＝c(ui,mj) By utilizing maximum likelihood estimation, the task time coefficients beta of different tasks under the mine can be calculated by introducing three independent variables2Task location coefficient beta1And a task quality coefficient beta3And a constant value of beta0The value of (d);
in order to avoid that the credibility is influenced by the difference of the number of the miners participating in the tasks, the total average value processing is carried out on all the miners credibility, so that the miners credibility formula is as shown in formula (12):
Figure BDA0003060985810000211
where k represents the perceptual user uiTotal number of historically completed tasks, R (u)i) Representing the reputation of the miners.
Example 2:
example 2 is substantially the same as example 1 except that:
the step 3: establishing a task set and a miner state information base, and aiming at the task type, adopting Gray code to the task type
Figure BDA0003060985810000212
Dividing, carrying out condition constraint according to the requirement of the perception task on perception time, and drawing a task type table according to the requirement of the task type, wherein the table is as follows:
Figure BDA0003060985810000213
miners with different functions can complete different types of tasks, and the system selects miners with corresponding functions to complete the tasks with corresponding types according to the task type codes.
Example 3:
example 3 is substantially the same as example 2 except that:
an example of the algorithm:
comparative demonstration of participant selection mechanism using MATLAB: in order to simulate real sensing task distribution and ensure that a simulation result is close to a fact, a real mine environment is abstracted into a weighted undirected graph with 30 nodes, and N sensing tasks are simulated, and M users are randomly distributed in edges or nodes of the undirected graph. In addition, the effectiveness of the model strategy and the solution method of the invention is verified by setting the upper limit of the total number of users M to 60 and the upper limit of the number of tasks N to 50, and the credit value Ci of each participant and the perception quality of each user are subject to uniform distribution on [0,1 ]. Specific experimental groupings and parameter settings are shown in the following table:
Figure BDA0003060985810000214
Figure BDA0003060985810000221
the objective of the emergency task allocation strategy is to enable the task to be processed and completed as soon as possible, that is, a higher task allocation proportion can be obtained in a shorter time, and the non-emergency task tends to select a less-fatigued participating miner with a more sufficient device power to complete the task. Therefore, it is important for the two tasks to select the miners with different requirements suitable for different task types.
And selecting a MARL algorithm, a Heuristic (HRA) algorithm and an HTA algorithm as experimental reference algorithms of the E-WPA algorithm.
In the simulation process, each type of experiment is carried out for 10 times, and the simulation result is the average value of the 10 experimental results. In the following simulation experiment, the time of task allocation, the number of mobile users, the sum of credit degrees and the related indexes of fatigue degrees are mainly considered.
Simulation analysis:
in the face of the allocation requirement of an emergency task, a participator with higher credibility is required to be found in a shorter time; therefore, the number of tasks N is set to 10, and the distribution time of the distribution algorithm in 4 is respectively observed under different numbers of people, as shown in fig. 4;
referring to fig. 4, in comparison of the present configuration, it can be seen that, in terms of reaching the same computation time, due to the increasing number of miners, the data volume of the number of participating persons to be considered and computed by a Heuristic (HRA) algorithm increases, while the HAT algorithm increases the task allocation time by determining the central node and analyzing the relevant attributes of the subordinate nodes, and the task allocation time of the emergency weighted path algorithm of the present invention is obviously superior to the other three algorithms. Statistics show that the average time required for calculation is shortened by 10% on average when the similar distribution rate is achieved. By increasing the scale, the allocation time of the algorithm to miners of different sizes can be verified.
Referring to fig. 5, another key evaluation index for the emergency mission algorithm is reputation, and for this reason, the average reputation of the selected miners participants under different numbers of people is also verified in the experiment. Experiments show that the average credibility of miners selected by the weighted undirected graph algorithm is approximately maintained at the level of 0.8 according to the increasing number of people, and the credibility of participants is also maintained at a higher level due to the fact that a Heuristic (HRA) algorithm also takes the credibility of the participants as an important optimization index, and as can be seen from fig. 5, the urgent weighted undirected graph algorithm is always superior to other algorithms, and the MAR algorithm and HAT algorithm which pay attention to time and central nodes are relatively lower.
As an algorithm related to task allocation, the success rate of task allocation is also an evaluation index of important consideration, see fig. 6 and 7, and observe the allocation capability of the weighted path algorithm in emergency from two aspects of the number of participants in different scales and the number of tasks in different scales, as can be seen from fig. 6 and 7: the distribution success rate of the weighted undirected graph algorithm is higher than that of the other three algorithms; and as can be seen from fig. 6: with the increase of the number of participated people, the success rate of each distribution algorithm is in an ascending trend and is stabilized to be more than 90%; when the number of participants is small, the weighted undirected path algorithm can still keep the success rate of nearly 50%, and because other algorithms excessively depend on certain indexes such as the position of miners and the interest preference of miners, the HAT algorithm depends on the meeting of a central node and a subordinate node, and the task allocation success rate is obviously low.
The success rate of non-emergency task allocation is also an evaluation index to be considered, see fig. 8 and 9, the allocation capability of the weighted path algorithm under the non-emergency condition is observed from the number of participants with different scales and the number of tasks with different scales, as can be seen from fig. 8 and 9: the task allocation under the non-emergency condition is that the allocation power is higher than that of the other three algorithms; and as can be seen from fig. 8: with the increase of the number of participated people, the success rate of each distribution algorithm is in an ascending trend and is stabilized to be more than 90%; when the number of participants is less, the weighted undirected path algorithm of the invention can still keep the success rate of nearly 40 percent obviously higher than the other three algorithms.
For the weighted path algorithm in the non-emergency task state, after the task allocation success rate is simulated respectively, the average state value of the selected miners under the non-emergency task is compared, and the average electric quantity of the equipment of the selected miners is selected. From fig. 10, it can be seen that under the non-emergency task, the miner status value is maintained at 0.6, and the other three algorithms are all below 0.5, and the invention has nearly 10% improvement on the status value. And it can be seen from fig. 11 that the remaining power of the equipment of the selected participant is increased by 15% on average through the state value, so that it can be seen that the miners selected by the non-emergency algorithm are superior to the other three algorithms in terms of fatigue state and remaining power of the equipment. Therefore, miners with low fatigue and high equipment power can be preferably selected by the algorithm to participate in the task.
Compared with the similar algorithm through experimental simulation, the feasibility of the algorithm is verified: in the emergency task allocation algorithm, the allocation time is reduced under the same allocation rate, and higher average credit degree can be kept under different participation numbers, so that the completion of mine tasks is effectively guaranteed. The comprehensive state value of the selected participants in the non-emergency task allocation algorithm is higher than that of the similar algorithm, so that matching of miners, equipment and task places is optimized.

Claims (3)

1. A mine mobile crowd sensing task distribution method based on a weighted undirected graph is characterized by comprising the following steps:
the task distribution method is suitable for mines covered with network signals, and for miners, the individual miners carry intelligent terminals capable of being connected with the network, and the intelligent terminals can detect the heart rate and the downhole state of the miners;
the task distribution method comprises the following steps:
step 1: establishing a credibility model of the miners,
defining a participant's reputation as R (u)i) The credit degree is in the range of [0,1]]The credibility value is 0 to indicate that the miner is completely untrustworthy, the credibility value is 0.5 to indicate uncertainty, the credibility value is 1 to indicate complete credibility, and the initial credibility of the participant who participates in the perception task for the first time is 0.5;
establishing a miner credibility model R (u) according to three data information of task completion place similarity, task completion time similarity and task completion qualityi)；
Step 2: making an undirected path graph, taking each underground roadway as the edge of the undirected graph, taking the connection point of each chamber, the roadway and the roadway as the node of the undirected graph, and acquiring the moving speed of each miner
Figure FDA00030609858000000114
Dis (d)i,dj) Indicating mine adjacent node di,djThe weight between adjacent nodes is determined according to the undirected graph edge length and the miner speed
Figure FDA0003060985800000011
w(di,dj) Represents a neighboring node di,djWeight in between;
and step 3: a task set and a miner state information base are established,
establishing a task set: editing known tasks to be completed into a task set M ═ M1,m2,m3…mnWhere m is the 4 sets of important information for a single task:
Figure FDA0003060985800000012
Figure FDA0003060985800000013
indicates the number of persons required for the task and
Figure FDA0003060985800000014
Figure FDA0003060985800000015
which represents the start time of the task,
Figure FDA0003060985800000016
which indicates the end time of the task,
Figure FDA0003060985800000017
indicating that the system predicted the time required to complete the task and
Figure FDA0003060985800000018
Figure FDA0003060985800000019
represents the type of task;
establishing a miner state information base: r participants are provided, and a miner state information base U ═ of the r participants is established1,u2,u3,...,ur) Wherein
Figure FDA00030609858000000110
R(ui) Expressed as a measure of the reputation of the mineworker,
Figure FDA00030609858000000111
expressed as the average moving speed of the miners and
Figure FDA00030609858000000112
(xi,yi) Indicating the miner uiPosition ofThe coordinates of the position of the object to be imaged,
Figure FDA00030609858000000113
indicating the miner uiA processed historical task type;
and 4, step 4: the distribution of the urgent tasks is carried out,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after adding a task node, acquiring the type and position information of a task, and screening miners who can meet the requirement of solving the task from all the miners to acquire a candidate set of the miners according to the type of the task and whether the miners have engaged in similar tasks;
because a certain time is needed for completing the task, the time needed for completing the task is subtracted on the basis of the specified task weight of the system:
Figure FDA0003060985800000021
the weight of the final task is obtained; the total weight w (d) of the edges from the task position node to the diffusion node is obtained in the diffusion process by taking the task node as the center, diffusing outwards along the communicated edges and continuously superposing the weight w of the passing edges in the diffusion processi,dj)；
Edge weight sum w (d) of the path as it passesi,dj) Achieving task requirement weight wmjStopping diffusion, and acquiring a full path plan taking the task node as a center within a specified constraint time; then miners in the candidate set acquired by the miners meeting the task requirements in the path are ranked according to the credibility, the miners with the highest credibility are preferentially selected to execute the emergency task, and a task and a path diagram are sent to the intelligent equipment of the selected miners, and the emergency task allocation is completed at the moment;
and 5: the assignment of the non-urgent tasks is performed,
for the distribution of a single emergency task, at the beginning of the algorithm, the emergency task is taken as a new task node to be added into an undirected graph; after the task node is added, acquiring the type and position information of the task, and screening out all miners which can finish the task of the task type in the underground miner set according to the task type and whether the miners engage in similar tasks; according to the weighted undirected graph, starting from a task node, diffusing outwards to obtain a place which can be reached by the task furthest in a task time range, and selecting miners meeting the task requirement in the place range as a candidate set of miners;
firstly, according to information fed back from an intelligent terminal carried by a miner, constructing a fatigue state model of the miner, wherein the fatigue state model is as follows:
Figure FDA0003060985800000022
in the above formula:
Figure FDA0003060985800000023
indicating the miner uiFatigue value of (2), unit: w; u. ofibjIndicating the miner uiHeart rate, unit: the times are/min;
Figure FDA0003060985800000024
indicating the miner uiHeight, unit: cm;
Figure FDA0003060985800000025
indicating the miner uiAge, unit: year;
Figure FDA0003060985800000026
indicating the miner uiLength of time to trip, unit: min;
in a fatigue state model
Figure FDA0003060985800000027
According to the acquired electric quantity of the miner portable intelligent terminal
Figure FDA0003060985800000028
A state value model related to miners is constructed, and the formula is as follows (2):
Figure FDA0003060985800000031
in the above formula: st (u)i) For the miner's real-time status value, kmaxMaximum value of real-time fatigue, k, among all minersminMinimum value of real-time fatigue among all miners, EmaxMaximum value of real-time electricity among all miners, EminIs the minimum value of real-time electricity quantity, alpha, in all miners1And alpha2In order to operate the coefficients of the operation,
Figure FDA0003060985800000032
α2＝1-α1and 0 is not more than alpha1≤1,0≤α2≤1；
After the state values of all miners are obtained, the state values are subjected to standardization treatment, wherein the formula is as follows (3):
Figure FDA0003060985800000033
wherein: st (u)i) Indicating the miner uiState value of (st)minMinimum value, st, representing the state of participation in the mineworkermaxMaximum value representing the state of participating miners;
in the miner state evaluation model ST (u)i) On the basis of the method, a miner credibility evaluation model R (u)i) The miner evaluation function RS (u) is constructedi) As in formula (4):
RS(ui)＝αR(ui)+βST(ui) (4)
wherein: alpha and beta are both arithmetic coefficients, 0 < alpha < 1,0 < beta < 1, alpha + beta is 1, RS (u)i) Representing a non-emergency task miner evaluation function;
the miners meeting the task requirements in the path are acquired into a candidate setThe miners in the interior are rated according to the non-emergency task miners index RS (u)i) And sequencing, preferentially selecting the miners with the non-emergency tasks to execute the non-emergency tasks, wherein the miners with the evaluation indexes of the non-emergency tasks are front, and sending the tasks and the path diagram to the intelligent equipment of the selected miners, wherein the non-emergency tasks are distributed completely.
2. The mine mobile crowd-sourcing aware task distribution method based on weighted undirected graph according to claim 1, wherein:
the step 1: in the establishment of a model of the credibility of miners,
i, constructing a model of the similarity l of the task completion site, wherein the model is as shown in a formula (5):
Figure FDA0003060985800000041
in the above formula, ui(xi,yi),mj(xj,yj) Respectively representing the location vectors of the miners i and the tasks j, where xi,yiRespectively representing the vertical and horizontal coordinates, x, of the position at which the mineworker i completes the task and submits the dataj,yjThe ordinate and abscissa representing the position required by task j; delta x and delta y are coordinate position variables of task adjustment, and delta x is greater than 0 and delta y is greater than 0;
II, constructing a task completion time similarity p model as shown in a formula (6):
Figure FDA0003060985800000042
in the above formula, the first and second carbon atoms are,
Figure FDA0003060985800000043
for the time that the mineworker i starts the task j,
Figure FDA0003060985800000044
for the time that miner i finishes task j,
Figure FDA0003060985800000045
the start time of task j is required for the system,
Figure FDA0003060985800000046
the end time of task j is required for the system,
Figure FDA0003060985800000047
indicating the degree of difference in user initiation from the initiation of the task request,
Figure FDA0003060985800000048
the difference degree of the latest end time of the user and the end time required by the task is represented, and n represents the difference degree interval divided by miners;
calculating the difference degree between the starting time of the perception user and the starting time of the perception task:
the time difference value of the user participating in the task can be used for measuring the difference degree of the perception user starting to execute the perception task; first, the start time of the task will be perceived
Figure FDA0003060985800000049
And start time of each perceived user
Figure FDA00030609858000000410
Converting into numerical data; then, the absolute difference of the starting time between the perception task and the different perception users is calculated
Figure FDA00030609858000000411
The minimum and maximum difference in absolute difference in start time between a perception task and different perception users is [ theta ]1,θn]This interval is divided on average into n-1 equidistant cells: { [ theta ] theta1,θ2],[θ2,θ3],...,[θn-1,θn]When the absolute difference value between the task starting time of the sensing user and the task starting time required by the sensing task falls in a certain cellAfter the time, giving a difference degree value {0, 1, …, n-1, n } to each cell in sequence to obtain a difference degree value corresponding to the cell where the absolute difference value between the task starting time of the perception user and the required starting time of the perception task falls;
ii, obtaining a difference degree value of the perception user and the perception task deadline by adopting a calculation mode with the same difference degree of the perception user and the perception task starting time;
iii, adding the difference degree of the perception user and the perception task time and the perception user and the perception task deadline time, and dividing the sum by the difference interval to obtain the time difference degree of the miner i to the task j through calculation;
III, constructing a model of data quality reliability q, as shown in formula (8):
Figure FDA0003060985800000051
Figure FDA0003060985800000052
in the above formula:
Figure FDA0003060985800000055
mean value of values, ds (u), representing task ji,mj) Value, ds (u), representing the miner upload task ji,mj)maxRepresents the maximum value of the uploaded data of all the participating miners for completing the task j,
Figure FDA0003060985800000056
total number of people participating in task j;
when the numerical value provided by the miners is closer to the real data, the uploaded data is higher in reliability, and the data quality reliability formula q (u) isi,mj) The closer to 1 the middle expression is, and otherwise, the closer to 0 the middle expression is, the data acquired by the miners are far away from the real data, namely, the deviation from the real data is larger, and the reliability is lower;
IV, constructing a miner credibility model, wherein the final classification result of miners in the model is that the dependent variable y is 1 when the credibility is more than 0.5 or the dependent variable y is 0 when the credibility is less than 0.5, so that the value of the dependent variable y in the model is 0 or 1, the dependent variable y accords with the 0 and 1 distribution, and determining coefficients of three independent variable factors by adopting a logistic regression method to establish the credibility model, wherein the formula (9) shows that:
Figure FDA0003060985800000057
wherein: (x) is a logical function of (f),
Figure FDA0003060985800000053
c(ui,mj) The credibility of the miner i participating in the task j is obtained;
let x be beta0+β1·l(ui,mj)+β2·p(ui,mj)+β3·q(ui,mj) (10)
Substituting conversion and logarithm are taken into the formula (9) to obtain a single task credibility logistic regression model:
Figure FDA0003060985800000054
calculating according to a logistic regression solution method, and enabling E (y)i＝1)＝c(ui,mj) By utilizing maximum likelihood estimation, the task time coefficients beta of different tasks under the mine can be calculated by introducing three independent variables2Task location coefficient beta1And a task quality coefficient beta3And a constant value of beta0The value of (d);
in order to avoid that the credibility is influenced by the difference of the number of the miners participating in the tasks, the total average value processing is carried out on all the miners credibility, so that the miners credibility formula is as shown in formula (12):
Figure FDA0003060985800000061
where k represents the perceptual user uiTotal number of historically completed tasks, R (u)i) Representing the reputation of the miners.
3. The mine mobile crowd-sourcing aware task distribution method based on weighted undirected graph according to claim 1 or 2, wherein:
the step 3: establishing a task set and a miner state information base, and aiming at the task type, adopting Gray code to the task type
Figure FDA0003060985800000063
Dividing, carrying out condition constraint according to the requirement of the perception task on perception time, and drawing a task type table according to the requirement of the task type, wherein the table is as follows:
Figure FDA0003060985800000062
miners with different functions can complete different types of tasks, and the system selects miners with corresponding functions to complete the tasks with corresponding types according to the task type codes.
CN202110512785.0A 2021-05-11 2021-05-11 Mine mobile crowd sensing task distribution method based on weighted undirected graph Active CN113159620B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110512785.0A CN113159620B (en) 2021-05-11 2021-05-11 Mine mobile crowd sensing task distribution method based on weighted undirected graph

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110512785.0A CN113159620B (en) 2021-05-11 2021-05-11 Mine mobile crowd sensing task distribution method based on weighted undirected graph

Publications (2)

Publication Number Publication Date
CN113159620A true CN113159620A (en) 2021-07-23
CN113159620B CN113159620B (en) 2023-08-18

Family

ID=76874695

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110512785.0A Active CN113159620B (en) 2021-05-11 2021-05-11 Mine mobile crowd sensing task distribution method based on weighted undirected graph

Country Status (1)

Country Link
CN (1) CN113159620B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108600271A (en) * 2018-05-10 2018-09-28 重庆邮电大学 A kind of method for secret protection of trust state assessment
CN109120523A (en) * 2017-06-23 2019-01-01 阿里巴巴集团控股有限公司 Multi-node path selection method, device, cloud platform resource regulating method and device
CN110825810A (en) * 2019-10-28 2020-02-21 天津理工大学 Block chain-based crowd sensing double privacy protection method
CN111580978A (en) * 2020-05-12 2020-08-25 中国联合网络通信集团有限公司 Edge computing server layout method and task allocation method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109120523A (en) * 2017-06-23 2019-01-01 阿里巴巴集团控股有限公司 Multi-node path selection method, device, cloud platform resource regulating method and device
CN108600271A (en) * 2018-05-10 2018-09-28 重庆邮电大学 A kind of method for secret protection of trust state assessment
CN110825810A (en) * 2019-10-28 2020-02-21 天津理工大学 Block chain-based crowd sensing double privacy protection method
CN111580978A (en) * 2020-05-12 2020-08-25 中国联合网络通信集团有限公司 Edge computing server layout method and task allocation method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
向罗勇;陈文;张陆洋;: "联合区域热度和社交属性感知的移动群智感知参与者选择机制", 电信科学, no. 02, pages 28 - 38 *

Also Published As

Publication number Publication date
CN113159620B (en) 2023-08-18

Similar Documents

Publication Publication Date Title
Capponi et al. A cost-effective distributed framework for data collection in cloud-based mobile crowd sensing architectures
CN105046454A (en) River basin water-based environment management decision support system
KR102586745B1 (en) Method and apparatus of controlling energy consumption
EP2894573B1 (en) Information processing device, information processing method, and program
CN104881800A (en) Mobile-crowd-sourcing-sensing-based motivation system realization method
CN113286315B (en) Load balance judging method, device, equipment and storage medium
CN102904343A (en) State monitoring system and method based on distributed multi-agent system
CN109408228A (en) Intelligent perception method for allocating tasks based on allocation of budget
CN111090631A (en) Information sharing method and device under distributed environment and electronic equipment
CN109902974B (en) Construction scheme generation method and device
CN114019831B (en) Water resource monitoring internet of things platform
CN113159620A (en) Mine mobile crowd sensing task distribution method based on weighted undirected graph
CN114492927A (en) Meteorological information pushing method, storage medium and computer terminal
CN109886581A (en) Participant&#39;s selection method based on the quality of data in a kind of mobile gunz perception task
CN109068271A (en) Control method, device, equipment and the storage medium of controller
Wu et al. A utility-based subcontract method for sensing task in mobile crowd sensing
Wang et al. A task allocation algorithm for coal mine mobile crowd sensing based on weighted undirected graph
CN103702277A (en) Data transmission method and electronic equipment
Liu et al. A hybrid federated learning framework with dynamic task allocation for multi-party distributed load prediction
CN116192906A (en) Industrial park comprehensive energy digital monitoring method based on cloud edge end cooperation
CN114035671A (en) Cloud + multi-terminal cooperative energy-saving task unloading method for multi-working-mode wearable computing system
CN113973339A (en) Load coordination method and device for 4G cell and 5G cell and computer storage medium
Gad-ElRab et al. Multiple criteria-based efficient schemes for participants selection in mobile crowd sensing
CN111475266B (en) Diversity-constrained crowd sensing task allocation method
CN110825500B (en) Crowd sensing task allocation method for three-layer structure

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant