CN114493106B - Space crowdsourcing task allocation method based on geographical division - Google Patents

Space crowdsourcing task allocation method based on geographical division Download PDF

Info

Publication number
CN114493106B
CN114493106B CN202111588397.7A CN202111588397A CN114493106B CN 114493106 B CN114493106 B CN 114493106B CN 202111588397 A CN202111588397 A CN 202111588397A CN 114493106 B CN114493106 B CN 114493106B
Authority
CN
China
Prior art keywords
task
delivery point
follows
graph
distribution
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.)
Active
Application number
CN202111588397.7A
Other languages
Chinese (zh)
Other versions
CN114493106A (en
Inventor
郑凯
陈轩磊
叶冠宇
赵艳
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.)
Higher Research Institute Of University Of Electronic Science And Technology Shenzhen
Original Assignee
Higher Research Institute Of University Of Electronic Science And Technology Shenzhen
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 Higher Research Institute Of University Of Electronic Science And Technology Shenzhen filed Critical Higher Research Institute Of University Of Electronic Science And Technology Shenzhen
Priority to CN202111588397.7A priority Critical patent/CN114493106B/en
Publication of CN114493106A publication Critical patent/CN114493106A/en
Application granted granted Critical
Publication of CN114493106B publication Critical patent/CN114493106B/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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Strategic Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Educational Administration (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Remote Sensing (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a space crowdsourcing task allocation method based on geographical division. According to the method, a given geographic space is divided by using an algorithm based on a voronoi diagram and an algorithm based on a voronoi diagram with adaptive weighting, and a task delivery point corresponding to each distribution center is obtained. Then, feature vectors of a distribution center and a task delivery point are obtained by using a graph neural network (CMPNN), and then an attention mechanism is used to obtain feature vectors of workers by combining the feature vectors. And finally, finding the optimal task allocation by using a strategy-based reinforcement learning algorithm. The invention obviously improves the effectiveness and efficiency of task allocation in spatial crowdsourcing through the verification of a real data set and a manually synthesized data set.

Description

Space crowdsourcing task allocation method based on geographical division
Technical Field
The invention relates to the technical field of information processing, in particular to a space crowdsourcing task allocation method based on geographical division.
Background
The development of GPS-enabled smart devices and communication technologies has prompted a further expansion of the spatial crowdsourcing market. In spatial crowdsourcing, task requesters may send spatial tasks to a spatially crowdsourced server, and the server treats the carrier of the smart device as a worker who needs to travel to a specified location and complete the spatial tasks.
In recent years, space crowdsourcing technology has attracted much attention, and therefore many task allocation methods are proposed for different application scenarios. The multi-target task allocation method has the main idea that an optimal solution is searched according to target conflicts by utilizing a multi-target particle swarm optimization algorithm and a sequencing strategy algorithm. But this approach ignores the geographic information that is a prerequisite for task allocation in the spatial dimension. In some real-life scenarios, there are many distribution centers, each of which is responsible for task distribution in a specific area only, and each worker works for only one distribution center. The task allocation method of multi-worker cooperation ignores time information of workers and tasks although geographical information is taken into consideration. The task allocation method based on dynamic rewards also does not take into account the end time of the task.
Recently, a task allocation method based on reinforcement learning has been proposed, which is disadvantageous to the balance of task allocation by setting different priorities for different tasks. And it merely divides the geographic space into grids without considering the importance of workers to the task delivery point.
Disclosure of Invention
The invention aims to provide a space crowdsourcing task allocation method based on geographical division.
In order to achieve the purpose, the invention is implemented according to the following technical scheme:
the invention comprises the following steps:
s1: dividing a given geographic space by using a voronoi diagram-based algorithm and a voronoi diagram-based algorithm based on self-adaptive weighting to obtain a task delivery point corresponding to each distribution center;
s2: obtaining the feature vectors of a distribution center and a task delivery point by using a graph neural network, and then obtaining the feature vectors of workers by combining the feature vectors and using an attention mechanism;
s3: and finding the optimal task distribution by using a reinforced learning algorithm based on a strategy.
The invention has the beneficial effects that:
compared with the prior art, the space crowdsourcing task allocation method based on geographic division divides a complex space problem into a plurality of sub-problems based on geographic information, and can obtain an accurate task allocation result more easily. According to the method, a given geographic space is divided by using a Weino graph-based algorithm and a self-adaptive weighting-based algorithm of the Weino graph, and a task delivery point corresponding to each distribution center is obtained. Then, feature vectors of the distribution center and the task delivery point are obtained by using a graph neural network (CMPNN), and then the feature vectors of workers are obtained by combining the feature vectors by using an attention mechanism. And finally, finding the optimal task distribution by using a reinforced learning algorithm based on a strategy. The invention obviously improves the effectiveness and efficiency of task allocation in spatial crowdsourcing through the verification of a real data set and a manually synthesized data set.
Drawings
FIG. 1 is an exemplary diagram of task allocation of the present invention;
FIG. 2 is the present invention experiment (1) the Gmission data set;
in fig. 2: (a) CPU time; (b) total number of assigned tasks; (c) evenly distributing the task differences;
FIG. 3 is a SYN dataset of experiment (1) of the present invention;
in FIG. 3: (a) CPU time; (b) total number of assigned tasks; (c) evenly distributing the task differences;
FIG. 4 is the present invention's experimental (2) Gmision dataset;
in fig. 4: (a) CPU time; (b) total number of assigned tasks; (c) evenly distributing the task differences;
FIG. 5 is an experimental (2) SYN dataset of the present invention;
in FIG. 5: (a) CPU time; (b) total number of assigned tasks; and (c) evenly distributing the task differences.
Detailed Description
The invention will be further described with reference to the drawings and specific embodiments, which are illustrative of the invention and are not to be construed as limiting the invention.
As shown in fig. 1: the implementation of the whole task allocation of the invention can be divided into two stages: first, geographical partitioning is performed, and second, task allocation is performed. The optimization target is that the number of distributed tasks is maximum, and the difference of the average number of distributed tasks is minimum.
Task allocation involves the definition of some initial data: distribution center, task delivery point, worker. The distribution center may be represented as c = (l, P, S, W), where l represents a location, P represents a set of task delivery points corresponding to the distribution center, S represents a set of assignable tasks, and W represents a set of workers. A task delivery point may be denoted as p = (l, S), where l denotes location and S denotes the set of tasks that the distribution center allocates for. The worker may be represented as w = (l, c), where l represents the location and c represents the distribution center at which the worker works. In addition to this, a set of valid delivery points is defined, which may be denoted as VPS (w). Each worker corresponds to an effective delivery point set, and the workers can complete the tasks in the effective delivery point set in time in the shortest time.
The first stage in the invention is to divide the geographic space, and to obtain the task delivery point corresponding to each distribution center, a voronoi diagram-based algorithm (VDA) and an adaptive weighted voronoi diagram-based algorithm (AWVDA) are used. For the algorithm based on the Voronoi diagram, the region is divided by taking a vertical bisector as a dividing line through an violence method. The inputs to the algorithm are a set of distribution centers C and a set of task delivery points P. And allocating the task delivery points according to the space distance, and allocating the task delivery points to the distribution center closest to the space distance. Finally, the delivery point sets C.P corresponding to all the distribution centers are output. Because the dynamic changes of tasks and workers are not considered by the VDA, the invention also provides an algorithm of a Voronoi diagram based on self-adaptive weighting, and the original space distance is changed into the self-adaptive distance as follows:
ad(p,c i )＝aw i ·d(p,c i )
Figure BDA0003428856000000041
wherein, ad (p, c) i ) Indicating delivery point p to distribution centre c i Adaptive distance of, aw i Indicating a distribution center c i Adaptive weight of d (p, c) i ) Represents the spatial distance, | c i S | represents the number of tasks,|c i w | number of workers. The assignment process is similar to the voronoi diagram.
The second stage in the present invention is the task allocation. Firstly, obtaining the characteristic vector of the distribution center and the task delivery point. Converting each part of divided geographic space into a graph g c The vertices include the distribution center and its task delivery point. The feature vectors of the vertices are then learned from the graph using a combined messaging neural network. The feature vector dimension of each vertex is 2, each edge has m categories, and the corresponding feature vector dimension is m. Set of z neighbor vertices Ne (v) based on graph structure information and vertex nearest i ) Obtaining the vertex v after n +1 rounds of iteration i New feature vector f of i n+1 As follows:
Figure BDA0003428856000000042
where Agg is an aggregation operator ed ij Representing a vertex v i And v j The connection between, k is a translation invariant global kernel,
Figure BDA0003428856000000043
d ed representing the dimension of the edge. To better predict the type of edge, the model can be further optimized as follows:
Figure BDA0003428856000000044
wherein h is ed Is a neural network, k ori And k nei Are two different translation invariant global kernels. After N iterations, the feature vector for each vertex contains information for its N nearest neighbors. Then, the feature vector of the worker is obtained through an attention mechanism, and the formula is as follows:
Figure BDA0003428856000000045
wherein W (W, v) i ) Is the normalized sum of attention scores
Figure BDA0003428856000000046
Is worker w and task delivery point v i Value in between. In computing query Q w When introducing graph context feature vectors to better weigh the importance of task delivery points to workers, as follows:
Figure BDA0003428856000000051
Figure BDA0003428856000000052
wherein [;]it is the operation of the connection that is,
Figure BDA0003428856000000053
is the feature vector of graph gc, f 1 Is the feature vector of the first vertex. And finally, performing task allocation by using strategy-based reinforcement learning. Obtaining importance weight of worker to task delivery point using attention mechanism in combination with feature vectors of worker and graph vertex
Figure BDA0003428856000000054
Importance represents the likelihood of a worker being assigned to an effective task delivery point, and the distribution of task assignments is the product of all possible assignments, as follows:
Figure BDA0003428856000000055
where w belongs to the set of workers, v i Belonging to the set of valid task delivery points. The optimization goal of reinforcement learning is to maximize the expected reward, as shown below:
θ * ＝argmax L Rc (θ)
Figure BDA0003428856000000056
wherein, re (A) c ) Is the reward per assignment that combines two optimization objectives as follows:
Figure BDA0003428856000000057
|A c s | is the number of assigned tasks, A c Dis is the average distance traveled by the worker, A c .S dif Is the difference in the average number of tasks allocated. Finally, to reduce the variance during training, the reward is improved using a merit function, as follows:
Figure BDA0003428856000000058
in the graph gc, SA task assignments are randomly generated in order to increase the convergence rate. And finally, outputting a final task distribution result according to the trained model.
The invention has the advantages that: the self-adaptive weighted Voronoi diagram algorithm is used for geographic division, the problem of uneven resource distribution caused by dynamic change of workers and tasks is solved, and a proper task delivery point is matched for each task distribution center. The feature vectors (distribution centers and their task delivery points) of the vertices are then obtained using a combined graph-based neural network, the message-passing neural network (CMPNN), which fully uses the information to the nearest k vertices in an iterative process to enhance feature richness. Subsequently, an artificial feature vector is obtained using an attention mechanism. And generating keys and values by using the graph vertex characteristic vectors respectively, and connecting the graph vertex characteristic vectors and the graph characteristic vectors to generate a query so as to embody the knowledge of workers on the information of the distribution center, thereby improving the effectiveness of task distribution. And finally, the task allocation is realized by using strategy-based reinforcement learning, so that faster convergence can be realized, and the optimal task allocation is obtained.
Experiment:
the experimental part uses two datasets, one is the real dataset gmission and the other is a manually generated dataset (SYN). The indexes recorded and compared in the experiment are the difference of CPU time, the total number of the distributed tasks and the average number of the distributed tasks. A genetic algorithm is introduced in the design of the comparison experiment, the simulated annealing algorithm is used for task allocation in the space after geographical division, and six comparison methods are designed.
1) Voronoi Diagram (VDA) + Genetic Algorithm (GA): voronoi diagrams for geography and genetic algorithms for task assignment
2) Voronoi Diagram (VDA) + simulated annealing algorithm (SA): voronoi diagram for geography division and simulated annealing algorithm for task allocation
3) Voronoi Diagram (VDA) + Reinforcement Learning (RL): voronoi diagram for geography division and simulated annealing algorithm for task allocation
4) Adaptive Weighted Voronoi Diagram (AWVDA) + Genetic Algorithm (GA): adaptive weighted voronoi diagram for geography partitioning and genetic algorithm for task allocation
5) Adaptive Weighted Voronoi Diagram (AWVDA) + simulated annealing algorithm (SA): adaptive weighted voronoi diagrams for geography partitioning and simulated annealing algorithms for task allocation
6) Adaptive Weighted Voronoi Diagram (AWVDA) + Reinforcement Learning (RL): adaptive weighted voronoi diagram for geography partitioning, reinforcement learning for task allocation
By controlling variables
1) Only the number of the task delivery points was changed to make ten sets of experiments. The gmission data sets are respectively set to 100, 200, 300, 400 and 500, the SYN data sets are respectively set to 1000, 2000, 3000, 4000 and 5000, and the experimental results are shown in fig. 2 and 3;
by changing the number of task delivery points, the AWVDA + RL achieves the best effect on both the gmission and syn data sets, the total number of distributed tasks is the largest, the difference of the average number of distributed tasks is the smallest, and the cpu time is relatively more but within the acceptable range.
2) Only the number of workers was changed to perform ten experiments. The gmission data sets are respectively set to 100, 200, 300, 400 and 500, the SYN data sets are respectively set to 1000, 2000, 3000, 4000 and 5000, and the experimental results are shown in fig. 3 and 4;
the task distribution by the reinforcement learning is more than the total number of the distributed tasks distributed by other methods, and the difference of the average task distribution number is smaller. The AWVDA reduces the difference in the average number of task assignments more than the VDA. The overall performance exhibited by the AWVDA + RL method is still the best.
The technical solution of the present invention is not limited to the limitations of the above specific embodiments, and all technical modifications made according to the technical solution of the present invention fall within the protection scope of the present invention.

Claims (2)

1. A space crowdsourcing task allocation method based on geographic partitioning is characterized by comprising the following steps:
s1: dividing a given geographic space by using a voronoi diagram-based algorithm and a voronoi diagram-based algorithm based on self-adaptive weighting to obtain a task delivery point corresponding to each distribution center;
the algorithm based on the voronoi diagram is as follows: dividing the areas by taking a vertical bisector as a boundary line through a brute force method, wherein the input of an algorithm is a distribution center set C and a task delivery point set P; distributing the task delivery points according to the spatial distance, and distributing the divided areas to distribution centers with the shortest distances; finally, outputting a delivery point set C.P corresponding to all distribution centers;
the algorithm of the voronoi diagram based on the self-adaptive weighting is as follows: changing the spatial distance into an adaptive distance as shown in the following formula:
ad(p,c i )＝aw i ·d(p,c i )
Figure FDA0003852268340000011
wherein, ad (p, c) i ) Indicating delivery point p to distribution center c i Adaptive distance of (aw) i Indicating a distribution center c i Adaptive weight of d (p, c) i ) Representing spatial distance, < >c i S | represents the number of tasks, | c i W | number of watchmen; the distribution process is the same as the algorithm based on the voronoi diagram;
s2: obtaining the feature vectors of a distribution center and a task delivery point by using a graph neural network, and then obtaining the feature vectors of workers by combining the feature vectors and using an attention mechanism;
the method for obtaining the feature vector of the worker by combining the feature vectors by using an attention mechanism comprises the following steps:
s2.1: converting each part of divided geographic space into a graph g c The top point comprises a distribution center and a task delivery point thereof;
s2.2: learning feature vectors of vertices from the graph using a combined messaging neural network; the feature vector dimension of each vertex is 2, each edge has m categories, and the corresponding feature vector dimension is m; set of z neighbor vertices Ne (v) based on graph structure information and vertex recency i ) Obtaining the vertex v after n +1 iteration i New feature vector of
Figure FDA0003852268340000012
As follows:
Figure FDA0003852268340000021
where Agg is an aggregation operator, d ij Representing a vertex v i And v j The connection between, k is a translation invariant global kernel,
Figure FDA0003852268340000022
d ed representing dimensions of edges
S2.3: the model was further optimized as follows:
Figure FDA0003852268340000023
wherein h is ed Is a neural network, and the network is a neural network,k ori and k nei The two different translation invariant global kernels are adopted, and after N iterations, the feature vector of each vertex comprises the information of N nearest vertices;
s2.4: the worker's feature vector is obtained by the attention mechanism, and the formula is as follows:
Figure FDA0003852268340000024
wherein W (W, v) i ) Is the normalized sum of attention scores
Figure FDA0003852268340000025
Is worker w and task delivery point v i A value in between; at computation of query Q w When introducing graph context feature vectors to better weigh the importance of task delivery points to workers, as follows:
Figure FDA0003852268340000026
Figure FDA0003852268340000027
wherein, [;]it is the operation of the connection that is,
Figure FDA0003852268340000028
is the feature vector of graph gc, f 1 Is the feature vector of the first vertex;
s3: and finding the optimal task distribution by using a reinforced learning algorithm based on a strategy.
2. The method for spatial crowd-sourced task allocation based on geographic partitioning according to claim 1, wherein: the step S3 specifically comprises the following steps:
obtaining by attention mechanism by combining feature vectors of workers and graph vertexesWeighting importance of workers to a task delivery point
Figure FDA0003852268340000029
The importance represents the likelihood of a worker being assigned to an effective task delivery point, and the distribution of task assignments is the product of all possible assignments, as follows:
Figure FDA00038522683400000210
where w belongs to the set of workers, v i Belonging to an effective task delivery point set; the optimization goal of reinforcement learning is to maximize the expected reward, as follows:
θ * ＝argmax L Rc (θ)
Figure FDA0003852268340000031
wherein, re (A) c ) Is the reward per assignment that combines two optimization objectives as follows:
Figure FDA0003852268340000032
|A c s | is the number of assigned tasks, A c Dis is the average distance travelled by the worker, A c .S dif Is the difference in the average number of assigned tasks; finally, to reduce the variance during training, the reward is improved using a merit function, as follows:
Figure FDA0003852268340000033
in order to accelerate the convergence rate, SA task assignments are randomly generated in the graph gc; and finally, outputting a final task distribution result according to the trained model.
CN202111588397.7A 2021-12-23 2021-12-23 Space crowdsourcing task allocation method based on geographical division Active CN114493106B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111588397.7A CN114493106B (en) 2021-12-23 2021-12-23 Space crowdsourcing task allocation method based on geographical division

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111588397.7A CN114493106B (en) 2021-12-23 2021-12-23 Space crowdsourcing task allocation method based on geographical division

Publications (2)

Publication Number Publication Date
CN114493106A CN114493106A (en) 2022-05-13
CN114493106B true CN114493106B (en) 2022-11-22

Family

ID=81493763

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111588397.7A Active CN114493106B (en) 2021-12-23 2021-12-23 Space crowdsourcing task allocation method based on geographical division

Country Status (1)

Country Link
CN (1) CN114493106B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115694877B (en) * 2022-08-30 2023-08-15 电子科技大学长三角研究院(衢州) Space crowdsourcing task allocation method based on federal preference learning

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110110545A (en) * 2019-03-25 2019-08-09 湖州师范学院 The space crowdsourcing Quality Control Model detected based on location privacy protection and tricker
CN113191416A (en) * 2021-04-26 2021-07-30 杭州电子科技大学 Large-scale geographic point data-oriented spatial attribute associated Voronoi diagram generation method
CN113822463A (en) * 2021-08-09 2021-12-21 武汉理工大学 Voronoi diagram and improved particle swarm algorithm based electric vehicle charging station location and volume fixing method

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10281279B2 (en) * 2016-10-24 2019-05-07 Invensense, Inc. Method and system for global shape matching a trajectory
US11295248B2 (en) * 2019-08-01 2022-04-05 Gcp Applied Technologies Inc. Coordinating concrete delivery and placement

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110110545A (en) * 2019-03-25 2019-08-09 湖州师范学院 The space crowdsourcing Quality Control Model detected based on location privacy protection and tricker
CN113191416A (en) * 2021-04-26 2021-07-30 杭州电子科技大学 Large-scale geographic point data-oriented spatial attribute associated Voronoi diagram generation method
CN113822463A (en) * 2021-08-09 2021-12-21 武汉理工大学 Voronoi diagram and improved particle swarm algorithm based electric vehicle charging station location and volume fixing method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
一种典型任务的多无人机协同任务分配算法研究;仲筱艳等;《自动化技术与应用》;20160825(第08期);全文 *
空间众包环境下的任务定价模型研究――以"拍拍赚"平台为例;童天语等;《赤峰学院学报(自然科学版)》;20181125(第11期);全文 *

Also Published As

Publication number Publication date
CN114493106A (en) 2022-05-13

Similar Documents

Publication Publication Date Title
CN113705610B (en) Heterogeneous model aggregation method and system based on federal learning
Tsai et al. Particle swarm optimization with selective particle regeneration for data clustering
CN110765320B (en) Data processing method, device, storage medium and computer equipment
CN104516785B (en) A kind of cloud computing resources dispatch system and method
Ye et al. Task allocation with geographic partition in spatial crowdsourcing
CN114902273A (en) System and method for optimizing resource allocation using GPU
CN114493106B (en) Space crowdsourcing task allocation method based on geographical division
CN114385376B (en) Client selection method for federal learning of lower edge side of heterogeneous data
CN113190041B (en) Unmanned aerial vehicle cluster online target distribution method based on constraint relaxation technology
CN112071122A (en) Unmanned aerial vehicle group path planning method for information acquisition
CN116777170A (en) Multi-robot task allocation method based on chaotic self-adaptive dung beetle optimization algorithm
CN111767977A (en) Group particle gradient descent algorithm based on improved genetic algorithm
Gupta et al. Multi-UAV deployment for NOMA-enabled wireless networks based on IMOGWO algorithm
CN114641006A (en) Frequency spectrum allocation method of cognitive radio network based on binary dragonfly optimization algorithm
CN111382896B (en) WTA target optimization method of self-adaptive chaotic parallel clone selection algorithm
Alkhashai et al. BF-PSO-TS: Hybrid heuristic algorithms for optimizing task schedulingon cloud computing environment
CN107094284B (en) Opportunistic network group moving method based on spectral clustering algorithm
Jiang et al. An adaptive immune‐following algorithm for intelligent optimal schedule of multiregional agricultural machinery
Zhou et al. Waste Not, Want Not: Service Migration-Assisted Federated Intelligence for Multi-Modality Mobile Edge Computing
CN111695667A (en) MapReduce-based distributed particle swarm clustering algorithm
Zhang Multiple particle swarm optimizers with inertia weight for multi-objective optimization
Guo et al. Multiple Methods Based Resource Allocation Optimization of Jamming System for UAV Clusters
CN107609138B (en) Cloud model data layout method and system
Kang et al. Application of the evolutionary algorithms for task allocation in uncertain environments with stochastic tuning
CN113253989B (en) Software and hardware cooperative integration architecture method based on embedded system

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