CN109819032B - Cloud robot task allocation method considering base station selection and computing migration in combined manner - Google Patents

Cloud robot task allocation method considering base station selection and computing migration in combined manner Download PDF

Info

Publication number
CN109819032B
CN109819032B CN201910069397.2A CN201910069397A CN109819032B CN 109819032 B CN109819032 B CN 109819032B CN 201910069397 A CN201910069397 A CN 201910069397A CN 109819032 B CN109819032 B CN 109819032B
Authority
CN
China
Prior art keywords
task
robot
cloud
base station
time
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
CN201910069397.2A
Other languages
Chinese (zh)
Other versions
CN109819032A (en
Inventor
陈武辉
翟攀
郑子彬
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.)
Sun Yat Sen University
Original Assignee
Sun Yat Sen University
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 Sun Yat Sen University filed Critical Sun Yat Sen University
Priority to CN201910069397.2A priority Critical patent/CN109819032B/en
Publication of CN109819032A publication Critical patent/CN109819032A/en
Application granted granted Critical
Publication of CN109819032B publication Critical patent/CN109819032B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Manipulator (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

The invention discloses a joint consideration baseThe cloud robot task allocation method for station selection and computing migration comprises the steps of dividing a single robot task w into five parts, and completing the time t of the single robot task wwAccumulating and establishing a task distribution model by the five parts; inputting a task set W, a robot set R, a base station set B and each task W of a plurality of robots R to upload corresponding configuration parameters of a cloud to a task allocation model tw(ii) a Assigning tasks to models twAnd sequentially decomposing the time into two submodels to minimize the time for completing all tasks step by step, wherein the first submodel is a model based on a simulated annealing algorithm: the second submodel is a model based on the hungarian algorithm. The method comprises the steps of functionalizing a target task and obtaining the minimum time for completing all tasks. The method comprises the steps of task layout, calculation migration and base station selection, and then provides a heuristic algorithm which can obtain the distribution strategy of a plurality of tasks in a short time.

Description

Cloud robot task allocation method considering base station selection and computing migration in combined manner
Technical Field
The invention relates to the field of task scheduling of cloud robots, in particular to a cloud robot task allocation method considering base station selection and computing migration in a combined mode.
Background
In 2010, James Kuffner Ph first proposed the concept of "cloud robot" to combine a mobile robot with a large cloud computing infrastructure. Traditional robots are expensive in hardware, poor in performance and difficult to handle complex computational tasks such as learning and recognition of images and natural language. Therefore, the system is difficult to become intelligent and powerful, and is difficult to popularize to a general use scene. The advent of cloud computing has provided a new avenue for solving this problem. The traditional robot and the cloud with infinite resources are connected through the network, so that the robot can use the computational resources and storage resources with infinite cloud, the capability of the robot is greatly expanded, and the requirement of the robot on hardware is reduced.
Moreover, the cloud can be regarded as a unified brain of an independent robot to manage the whole cloud robot system, and the robot is like four limbs of a human and can respond to task instructions sent by the cloud. The cloud and the robot are organically combined, so that a plurality of things which cannot be finished by the independent robot before can be finished. For example, each robot in the cloud robot system can acquire data learned by other robots, so that a new robot can become an experienced robot, and the service quality of the whole system is improved. As early as 2009, european scientists initiated a program named RoboEarth, whose purpose was to build a huge robot database. The participants can upload some information learned by the robot to the database, so that other robots can directly retrieve and use the information from the database without relearning the knowledge.
Although cloud robots can provide a variety of benefits, technical challenges remain. Such as a computing task migration problem for a robot. In the hypothesis, the robot can use the infinite resources in the cloud, but in practice, the use is also at a cost. As shown in fig. 1, a robot task can be divided into five parts without loss of generality, the robot moves to a target position, then data acquisition and processing are performed, the acquired data are processed, a processing result is reported to a cloud brain, and finally a corresponding response is made according to the processing result. Where the data processing portion is typically computationally intensive, if the computation is too complex, we can consider migrating this portion of the computation to the cloud. However, as can be seen from the figure, the collected data also needs to be uploaded to the cloud end correspondingly. If the amount of data is too large and the network speed is slow, this may result in uploading time longer than the computing time saved by cloud computing, which is not worth paying.
At present, the whole cloud robot system is only considered from the perspective of one task, and in the system, the robot can be connected with a cloud server through different kinds of networking equipment. In the working area of the system, occasionally tasks appear in different positions, and the cloud robot system needs to assign proper robots to complete the tasks in the shortest time. This does not allow for the decision of computing migration from the perspective of a single task alone, requiring a compromise between the processing power of the robot and the network bandwidth resources that the networked devices can provide. For example, if a task is assigned to a robot with a weaker processing power but is assigned more bandwidth resources, then the computational migration may be considered. Therefore, how to find a balance among a plurality of variables, make a proper calculation migration strategy, and a base station selection strategy and a task allocation strategy makes the time required by the system to complete a group of tasks which occur randomly shortest, which is the problem to be solved by the patent.
The existing research on task allocation is mainly based on the research of a large-scale data center, and in the task process of the large-scale data center, communication requirements among tasks need to be considered, and the tasks with a large amount of data communication requirements are placed on the same physical machine or the same physical machine of a local area network as much as possible. The research on migration calculation is mainly based on edge calculation. And unloading the tasks of the nodes with heavier task loads to the nodes with lighter task loads on the edge nodes with the executed tasks.
Disclosure of Invention
The invention mainly aims to provide a cloud robot task allocation method considering base station selection and computing migration in a combined mode.
The invention provides a cloud robot task allocation method considering base station selection and computing migration in a combined manner, which comprises the following steps of:
s10, establishing a task distribution model: dividing a single robot r into five parts, wherein the first part represents the time from the robot r to the position of the task w; the second part represents the time required for the pre-processing part of the task to run on the robot r; the third part represents the time for migrating the computing task to the cloud and executing; the fourth part represents the time of executing the calculation task on the robot and uploading the calculation task to the cloud; the fifth part represents the time when the robot performs the calculation to obtain the result, and the completion time t of the single robot task wwAccumulating and establishing a task distribution model by the five parts;
s20 inputs task set W of a plurality of robots rThe robot set R, the base station set B and each task w upload corresponding configuration parameters of the cloud to the task allocation model tw；
S30 assigning task to model twThe completion time t for completing all tasks step by sequentially decomposing the two submodelswMinimization solving
Figure GDA0003058266690000041
The first sub-model is a model based on a simulated annealing algorithm: the second submodel is a model based on the hungarian algorithm.
Preferably, the function of the task allocation model is specifically as follows:
Figure GDA0003058266690000042
wherein W represents a task set, W represents a task, R represents a robot set, R represents a robot, B represents a base station set, B represents a base station, and x represents a task setwrWith 1, the task w is placed on the robot r for processing, Σr∈Rxwr1 denotes placing a task w on one robot in the robot cluster R to execute,
Figure GDA0003058266690000043
representing the distance between the robot r and the task w, d representing the distance, IrIndicating the position of the robot, IwIndicating task position, msrRepresents the moving speed of the robot r;
Figure GDA0003058266690000044
indicating the time required to upload the collected data to the cloud, cdwIndicating the size of the amount of data that task w needs to collect, bdwIs a variable of base station correlation selection, precwIndicating the size of the computation of the pre-processing collection data of task w, csrRepresents the calculated speed of the robot r to perform the task;
zwtaking 1 represents the state of migrating and executing the computing task to the cloud, zwTaking 0 represents the execution state when the task is placed locally;
procwindicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure GDA0003058266690000051
Representing the time of executing the computing task w in the cloud; when z iswWhen the content is equal to 0, the content,
Figure GDA0003058266690000052
representing the time at which the computational task w is performed on the robot r,
Figure GDA0003058266690000053
the time required for uploading part of data generated in task calculation to the cloud is represented; et alwIndicating the time required for the robot to perform the calculation.
Preferably, the variable selected by the base station b correlation degree is defined as
Figure GDA0003058266690000054
mbrbIs the bandwidth size physically reached by the ideal state, cbIndicates the number of connections of the base station, mbbWhich represents the total bandwidth of the base station,
Figure GDA0003058266690000055
and the minimum value of the bandwidth which can be allocated to each connected cloud robot is taken as the actual bandwidth of the cloud robot.
Preferably, the minimum solution of the completion time for completing all tasks step by the two submodels is specifically:
based on the assumption that the calculation capacity of the robot is insufficient due to calculation unloading, and the base station selection is to achieve balance between calculation migration and local calculation, the processing capacity of the robot set R is approximately solved, the task of calculation migration is obtained, and a proper base station b is selected for the task of calculation migration to guarantee data transmission, so that when the processing capacity of the robot set R is insufficient, the robot set R can perform processingThe force is very weak, the minimum value of the processing capacity of all the robots is taken, the tasks of selecting calculation migration become more, and the user can consider the situation that the number of tasks is more
Figure GDA0003058266690000056
Cs in this sectionrAverage speed of task execution for all robots
Figure GDA0003058266690000057
| R | represents the number of elements in the robot set R, and is obtained by:
Figure GDA0003058266690000061
then, then
Figure GDA0003058266690000062
Wherein the content of the first and second substances,
Figure GDA0003058266690000063
represents a pair twA temporary expression in the evaluation process has no practical significance; CD (compact disc)wRepresents the size of the amount of data that task w needs to collect, when ZWWhen the number is equal to 1, the alloy is put into a container,
Figure GDA0003058266690000064
represents the time, bd, required to upload the collected data to the cloudwIs a selection y with the base stationwbRelated variable, wherein ywbIs a decision variable with a value of 0 or 1, when y iswb1 indicates that task w selects base station b for communication, when ywb0 indicates that task w does not select base station b for communication, defined as
Figure GDA0003058266690000065
mbrbIs the physically achievable bandwidth size, cb＝∑w∈WywbIndicates the number of tasks connected to base station b, mbbWhich represents the total bandwidth of the base station,
Figure GDA0003058266690000066
the bandwidth which can be allocated to each connected cloud robot is represented, and the minimum value of the bandwidth and the bandwidth is taken as the actual bandwidth, proc, of the cloud robotwIndicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure GDA0003058266690000067
Representing the time of executing the computing task w in the cloud;
operate to minimize
Figure GDA0003058266690000068
Solving for the acquisition variable y for the simulated annealing algorithm of the objective functionwbAnd ZwA value of (d);
will solve ywbAnd ZwBrought into
Figure GDA0003058266690000069
Figure GDA00030582666900000610
In the middle, the Hungarian algorithm is constructed, and a variable x is solved and obtainedwr；
Output xwr、ywbAnd ZwTo the objective function
Figure GDA00030582666900000611
The minimum time for completion of the target task W is obtained.
Preferably, the simulated annealing algorithm can be replaced by other intelligent search algorithms.
Preferably, the other intelligent search algorithm includes a particle swarm algorithm and a genetic algorithm.
The robots in the cloud robot system have the characteristics of high heterogeneity and high mobility, the high heterogeneity determines that different robots process different delays generated by the same task, and in order to improve the system efficiency and guarantee the service quality, it is very important to allocate the task to a proper robot for execution. Meanwhile, due to the existence of calculation migration, robots with weak processing capacity can process complex tasks, task allocation complexity is further increased, high mobility generates extremely high requirements on network connection, coverage ranges of base stations of different types are different from available bandwidths, when a plurality of robots select the same base station to perform network peer, communication quality is reduced undoubtedly, and it is very important to select a proper base station in order to guarantee different communication requirements of different robots and a cloud.
The characteristics of the cloud robot system determine that more factors need to be considered when the task of the cloud robot system is distributed, and the distribution scheme is more complex. Only when considered more comprehensively, a better quality allocation scheme can be obtained. The patent provides a task allocation method for minimizing task completion time by comprehensively considering task allocation and computing migration and combining a base station selection problem caused by movement of a cloud robot. According to the method, the problem is divided into two sub-problems in a heuristic manner by analyzing the characteristics of the optimal solution, so that the problem search space is greatly reduced, and the algorithm execution efficiency is accelerated. The two subproblems are solved by a simulated annealing algorithm and a Hungarian algorithm respectively, due to the fact that the problem search space is reduced, the simulated annealing algorithm can search excellent solutions more easily, and the Hungarian algorithm is a deterministic algorithm, and therefore the calculation accuracy of the algorithm is greatly guaranteed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
FIG. 1 is a flow chart of a method according to an embodiment of the present invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that, if directional indications (such as up, down, left, right, front, and back … …) are involved in the embodiment of the present invention, the directional indications are only used to explain the relative positional relationship between the components, the movement situation, and the like in a specific posture (as shown in the drawing), and if the specific posture is changed, the directional indications are changed accordingly.
In addition, if there is a description of "first", "second", etc. in an embodiment of the present invention, the description of "first", "second", etc. is for descriptive purposes only and is not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In addition, technical solutions between various embodiments may be combined with each other, but must be realized by a person skilled in the art, and when the technical solutions are contradictory or cannot be realized, such a combination should not be considered to exist, and is not within the protection scope of the present invention.
The invention mainly aims to provide a cloud robot task allocation method considering base station selection and computing migration in a combined mode.
The scheme is briefly described as follows: the problem is first mathematically modeled as a non-convex optimization problem, with the objective function being to minimize the completion time of all tasks. The layout of the task, the calculation migration and the base station selection are three solving variables. Since such a problem is an np-hard problem, an accurate solution cannot be found in polynomial time. The patent provides a heuristic algorithm by analyzing the structure of the problem accurate solution. The algorithm can obtain the distribution strategy of 100 tasks within 1 second, and the minimized objective function value is obviously superior to meta-heuristic algorithms, such as simulated annealing algorithm, genetic algorithm and the like. In order to achieve the aim, the invention provides a cloud robot task allocation method combining base station selection and computing migration,
mathematical modeling
First, the problem is expressed in mathematical language, as shown in fig. 1, a robot task can be divided into five parts, and the completion time t of a single robot taskwCan be obtained by adding these five parts. The mathematical representation is as follows: w represents a task set, W represents a task, R represents a robot set, R represents a robot, B represents a base station set, B represents a base station:
Figure GDA0003058266690000101
two, approximate solution
The problem is a combined optimization problem, but because too many variables are combined, the number of states is too large to solve in an effective time, and therefore, the patent proposes a step-by-step solving heuristic algorithm to approximately solve the problem.
The method specifically comprises the following steps:
s10, establishing a task distribution model: dividing a single robot r into five parts, wherein the first part represents the time from the robot r to the position of the task w; the second part represents the time required for the pre-processing part of the task to run on the robot r; the third part represents the time for migrating the computing task to the cloud and executing; the fourth part represents the time of executing the calculation task on the robot and uploading the calculation task to the cloud; the fifth part represents the robot performing the calculationsTime to result, time to completion t of a single robot task wwAccumulating and establishing a task distribution model by the five parts;
s20, inputting task sets W, R and B of multiple robots R and uploading corresponding configuration parameters of cloud ends to a task allocation model t by each task Ww；
S30 assigning task to model twThe completion time t for completing all tasks step by sequentially decomposing the two submodelswMinimization solving
Figure GDA0003058266690000102
The first sub-model is a model based on a simulated annealing algorithm: the second submodel is a model based on the hungarian algorithm.
Preferably, the function of the task allocation model is specifically as follows:
Figure GDA0003058266690000111
wherein W represents a task set, W represents a task, R represents a robot set, R represents a robot, B represents a base station set, B represents a base station, and x represents a task setwrWith 1, the task w is placed on the robot r for processing, Σr∈Rxwr1 denotes placing a task W on one robot in the robot cluster R for execution,
Figure GDA0003058266690000112
representing the distance between the robot r and the task w, d representing the distance, IrIndicating the position of the robot, IwIndicating task position, msrRepresents the moving speed of the robot r;
Figure GDA0003058266690000113
indicating the time required to upload the collected data to the cloud, cdwIndicating the size of the amount of data that task w needs to collect, bdwIs a variable of base station correlation selection, precwIndicating the size of the computation of the pre-processing collection data of task w, csrRepresents the calculated speed of the robot r to perform the task;
Zwtaking 1 represents the state of migrating and executing the computing task to the cloud, zwTaking 0 represents the execution state when the task is placed locally;
procwindicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure GDA0003058266690000114
Representing the time of executing the computing task w in the cloud; when z iswWhen the content is equal to 0, the content,
Figure GDA0003058266690000115
representing the time at which the computational task w is performed on the robot r,
Figure GDA0003058266690000116
the time required for uploading part of data generated in task calculation to the cloud is represented; et alwIndicating the time required for the robot to perform the calculation.
Preferably, the variable selected by the base station b correlation degree is defined as
Figure GDA0003058266690000117
mbrbIs the bandwidth size physically reached by the ideal state, cbIndicates the number of connections of the base station, mbbWhich represents the total bandwidth of the base station,
Figure GDA0003058266690000118
and the minimum value of the bandwidth which can be allocated to each connected cloud robot is taken as the actual bandwidth of the cloud robot.
Preferably, the minimum solution of the completion time for completing all tasks step by the two submodels is specifically:
the calculation unloading is based on the purpose of making up the insufficient computing capability of the robot, and the base station selection is based on the purpose of calculation migration andlocal calculation reaches a balanced assumption to approximately solve the processing capacity of the robot set R, obtain the task of calculation migration, and select a proper base station b for the task of calculation migration to guarantee data transmission
Figure GDA0003058266690000121
Cs in this sectionrAverage speed of task execution for all robots
Figure GDA0003058266690000122
| R | represents the number of elements in the robot set R, and is obtained by:
Figure GDA0003058266690000123
Figure GDA0003058266690000124
then
Figure GDA0003058266690000125
Wherein the content of the first and second substances,
Figure GDA0003058266690000126
represents a pair twA temporary expression in the evaluation process has no practical significance; CD (compact disc)wRepresents the size of the amount of data that task w needs to collect, when ZWWhen the number is equal to 1, the alloy is put into a container,
Figure GDA0003058266690000127
represents the time, bd, required to upload the collected data to the cloudwIs a selection y with the base stationwbRelated variable, wherein ywbIs a decision variable with a value of 0 or 1, when y iswb1 indicates that task w selects base station b for communication, when ywb0 indicates that task w does not select base station b for communication, defined as
Figure GDA0003058266690000128
mbrbIs the physically achievable bandwidth size, cb＝∑w∈WywbIndicates the number of tasks connected to base station b, mbbWhich represents the total bandwidth of the base station,
Figure GDA0003058266690000129
the bandwidth which can be allocated to each connected cloud robot is represented, and the minimum value of the bandwidth and the bandwidth is taken as the actual bandwidth, proc, of the cloud robotwIndicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure GDA00030582666900001210
Representing the time of executing the computing task w in the cloud;
operate to minimize
Figure GDA0003058266690000131
Solving for the acquisition variable y for the simulated annealing algorithm of the objective functionwbAnd ZwA value of (d);
will solve ywbAnd ZwBrought into
Figure GDA0003058266690000132
Figure GDA0003058266690000133
In the middle, the Hungarian algorithm is constructed, and a variable x is solved and obtainedwr；
Output xwr、ywbAnd ZwTo the objective function
Figure GDA0003058266690000134
The minimum time for completion of the target task W is obtained.
Preferably, the simulated annealing algorithm can be replaced by other intelligent search algorithms.
Preferably, the other intelligent search algorithm includes a particle swarm algorithm and a genetic algorithm.
The simulated annealing algorithm in the invention can be replaced by other intelligent search algorithms, such as particle swarm algorithm, genetic algorithm and the like, and can also obtain the same effect.
The robots in the cloud robot system have the characteristics of high heterogeneity and high mobility, the high heterogeneity determines that different robots process different delays generated by the same task, and in order to improve the system efficiency and guarantee the service quality, it is very important to allocate the task to a proper robot for execution. Meanwhile, due to the existence of calculation migration, robots with weak processing capacity can process complex tasks, task allocation complexity is further increased, high mobility generates extremely high requirements on network connection, coverage ranges of base stations of different types are different from available bandwidths, when a plurality of robots select the same base station to perform network peer, communication quality is reduced undoubtedly, and it is very important to select a proper base station in order to guarantee different communication requirements of different robots and a cloud.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and all modifications and equivalents of the present invention, which are made by the contents of the present specification and the accompanying drawings, or directly/indirectly applied to other related technical fields, are included in the scope of the present invention.

Claims (3)

1. A cloud robot task allocation method jointly considering base station selection and computing migration is characterized by comprising the following steps:
s10, establishing a task distribution model: dividing a single robot r into five parts, wherein the first part represents the time from the robot r to the position of the task w; the second part represents the time required for the pre-processing part of the task to run on the robot r; the third part represents the time for migrating the computing task to the cloud and executing; the fourth part represents the time of executing the calculation task on the robot and uploading the calculation task to the cloud; the fifth part represents the time when the robot performs the calculation to obtain the result, and the completion time t of the single robot task wwAccumulating and establishing a task distribution model by the five parts;
the function of the task allocation model is specifically as follows:
Figure FDA0003058266680000011
wherein W represents a task set, W represents a task, R represents a robot set, R represents a robot, B represents a base station set, B represents a base station, and x represents a task setwrWith 1, the task w is placed on the robot r for processing, Σr∈RxwrI denotes placing the task w on one robot in the robot cluster R for execution,
Figure FDA0003058266680000012
representing the distance between the robot r and the task w, d representing the distance, lrIndicating the position of the robot,/wIndicating task position, msrRepresents the moving speed of the robot r;
Figure FDA0003058266680000013
indicating the time required to upload the collected data to the cloud, cdwIndicating the size of the amount of data that task w needs to collect, bdwIs a variable of base station correlation selection, precwIndicating the size of the computation of the pre-processing collection data of task w, csrRepresents the calculated speed of the robot r to perform the task;
the variable selected by the base station b correlation degree is defined as
Figure FDA0003058266680000014
mbrbIs the bandwidth size physically reached by the ideal state, cbIndicates the number of connections of the base station, mbbWhich represents the total bandwidth of the base station,
Figure FDA0003058266680000015
the bandwidth which can be allocated to each connected cloud robot is represented, and the minimum value of the bandwidth and the bandwidth is taken as the actual bandwidth of the cloud robot;
zwtaking 1 represents the state of migrating and executing the computing task to the cloud, zwTaking 0 represents the execution state when the task is placed locally;
procwindicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure FDA0003058266680000021
Representing the time of executing the computing task w in the cloud; when Z iswWhen the content is equal to 0, the content,
Figure FDA0003058266680000022
representing the time at which the computational task w is performed on the robot r,
Figure FDA0003058266680000023
the time required for uploading part of data generated in task calculation to the cloud is represented; et alwRepresenting the time required for the robot to perform the calculation;
s20, inputting task sets W, R and B of multiple robots R and uploading corresponding configuration parameters of cloud ends to a task allocation model t by each task Ww；
S30 assigning task to model twThe completion time t for completing all tasks step by sequentially decomposing the two submodelswMinimization solving
Figure FDA0003058266680000024
The first sub-model is a model based on a simulated annealing algorithm: the second sub-model is a model based on the Hungarian algorithm;
the minimum solution of the completion time for completing all tasks step by step of the two submodels is specifically as follows:
the calculation unloading is based on the purpose of making up the insufficient computing capability of the robot, and the base station selection is based on the purpose of calculation migration and self-calculationAnd (3) calculating an assumption of reaching a balance to approximately solve the processing capacity of the robot set R, solving the task of calculating migration, and selecting a proper base station b for the task of calculating migration to ensure data transmission, wherein when the processing capacity of the robot set R is very weak and the processing capacity of all the robots is the minimum value, the number of tasks of selecting calculating migration is increased, and the number of tasks of selecting calculating migration is increased according to the condition that the processing capacity of all the robots is the minimum value
Figure FDA0003058266680000025
Cs in this sectionrAverage speed of task execution for all robots
Figure FDA0003058266680000026
| R | represents the number of elements in the robot set R, and is obtained by:
Figure FDA0003058266680000027
Figure FDA00030582666800000311
then
Figure FDA0003058266680000031
Wherein the content of the first and second substances,
Figure FDA0003058266680000032
represents a pair twA temporary expression in the evaluation process has no practical significance; CD (compact disc)wRepresents the size of the amount of data that task w needs to collect, when Zw＝1When the temperature of the water is higher than the set temperature,
Figure FDA0003058266680000033
represents the time, bd, required to upload the collected data to the cloudwIs a selection y with the base stationwbRelated variable, wherein ywbIs a decision variable with a value of 0 or 1, when y iswb1 indicates that task w selects base station b for communication, when ywb0 indicates that task w does not select base station b for communication, defined as
Figure FDA0003058266680000034
mbrbIs the physically achievable bandwidth size, cb＝∑w∈WywbIndicates the number of tasks connected to base station b, mbbWhich represents the total bandwidth of the base station,
Figure FDA0003058266680000035
the bandwidth which can be allocated to each connected cloud robot is represented, and the minimum value of the bandwidth and the bandwidth is taken as the actual bandwidth, proc, of the cloud robotwIndicating the amount of computation of the processing task w, cscRepresenting the computing speed of the cloud end, then
Figure FDA0003058266680000036
Representing the time of executing the computing task w in the cloud;
operate to minimize
Figure FDA0003058266680000037
Solving for the acquisition variable y for the simulated annealing algorithm of the objective functionwbAnd ZwA value of (d);
will solve ywbAnd ZwBrought into
Figure FDA0003058266680000038
Figure FDA0003058266680000039
In the middle, the Hungarian algorithm is constructed, and a variable x is solved and obtainedwr；
Output xwr、ywbAnd ZwTo the objective function
Figure FDA00030582666800000310
The minimum time for completion of the target task W is obtained.
2. The cloud robot task allocation method combining base station selection and computing migration according to claim 1, wherein the simulated annealing algorithm may be replaced with other intelligent search algorithms.
3. The cloud robot task allocation method combining base station selection and computing migration according to claim 2, wherein the other intelligent search algorithms include a particle swarm algorithm and a genetic algorithm.
CN201910069397.2A 2019-01-24 2019-01-24 Cloud robot task allocation method considering base station selection and computing migration in combined manner Active CN109819032B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910069397.2A CN109819032B (en) 2019-01-24 2019-01-24 Cloud robot task allocation method considering base station selection and computing migration in combined manner

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910069397.2A CN109819032B (en) 2019-01-24 2019-01-24 Cloud robot task allocation method considering base station selection and computing migration in combined manner

Publications (2)

Publication Number Publication Date
CN109819032A CN109819032A (en) 2019-05-28
CN109819032B true CN109819032B (en) 2021-06-29

Family

ID=66603209

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910069397.2A Active CN109819032B (en) 2019-01-24 2019-01-24 Cloud robot task allocation method considering base station selection and computing migration in combined manner

Country Status (1)

Country Link
CN (1) CN109819032B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110850789A (en) * 2019-12-09 2020-02-28 常州市小先信息技术有限公司 Robot energy efficiency optimization method based on edge calculation
CN111539863B (en) * 2020-03-26 2021-03-19 光控特斯联(重庆)信息技术有限公司 Intelligent city operation method and system based on multi-source task line
CN114385362A (en) * 2022-01-12 2022-04-22 东南大学 Multi-transfer robot scheduling method based on cloud-edge computing
CN114554529B (en) * 2022-04-21 2022-07-19 三峡智控科技有限公司 Intelligent optimization method for cell selection of 5G user terminal
CN114880857B (en) * 2022-05-13 2024-04-30 中国人民解放军国防科技大学 Weapon resource multi-stage optimization distribution method based on hybrid intelligent search

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103970609A (en) * 2014-04-24 2014-08-06 南京信息工程大学 Cloud data center task scheduling method based on improved ant colony algorithm
CN106936892A (en) * 2017-01-09 2017-07-07 北京邮电大学 A kind of self-organizing cloud multi-to-multi computation migration method and system
CN107273184A (en) * 2017-06-14 2017-10-20 沈阳师范大学 A kind of optimized algorithm migrated based on high in the clouds big data with processing cost

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10824475B2 (en) * 2017-03-24 2020-11-03 Tata Consultancy Services Limited Systems and methods for estimating computation times a-priori in fog computing robotics

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103970609A (en) * 2014-04-24 2014-08-06 南京信息工程大学 Cloud data center task scheduling method based on improved ant colony algorithm
CN106936892A (en) * 2017-01-09 2017-07-07 北京邮电大学 A kind of self-organizing cloud multi-to-multi computation migration method and system
CN107273184A (en) * 2017-06-14 2017-10-20 沈阳师范大学 A kind of optimized algorithm migrated based on high in the clouds big data with processing cost

Also Published As

Publication number Publication date
CN109819032A (en) 2019-05-28

Similar Documents

Publication Publication Date Title
CN109819032B (en) Cloud robot task allocation method considering base station selection and computing migration in combined manner
Lin et al. Smart manufacturing scheduling with edge computing using multiclass deep Q network
CN109492774B (en) Deep learning-based cloud resource scheduling method
CN109783224A (en) Method for allocating tasks, device and terminal device based on load allotment
CN113435472A (en) Vehicle-mounted computing power network user demand prediction method, system, device and medium
CN109993299A (en) Data training method and device, storage medium, electronic device
CN111586146B (en) Wireless internet of things resource allocation method based on probability transfer deep reinforcement learning
CN113543176A (en) Unloading decision method of mobile edge computing system based on assistance of intelligent reflecting surface
CN112784362B (en) Hybrid optimization method and system for unmanned aerial vehicle auxiliary edge calculation
CN113064671A (en) Multi-agent-based edge cloud extensible task unloading method
CN111988787B (en) Task network access and service placement position selection method and system
KR20190054449A (en) Method for placing compute node for deep neural network acceleration in heterogeneous cluster
CN111880911A (en) Task load scheduling method, device and equipment and readable storage medium
CN114585006B (en) Edge computing task unloading and resource allocation method based on deep learning
CN114595049A (en) Cloud-edge cooperative task scheduling method and device
CN112732444A (en) Distributed machine learning-oriented data partitioning method
CN114970834A (en) Task allocation method and device and electronic equipment
CN113592077B (en) Cloud edge DNN collaborative reasoning acceleration method for edge intelligence
Pérez et al. Parallel/distributed implementation of cellular training for generative adversarial neural networks
Henna et al. Distributed and collaborative high-speed inference deep learning for mobile edge with topological dependencies
CN112445617A (en) Load strategy selection method and system based on mobile edge calculation
CN116193395A (en) Unmanned cluster organization reconstruction role distribution method, system, equipment and storage medium
WO2018168695A1 (en) Distributed machine learning device, distributed machine learning method, and distributed machine learning recording medium
CN116367190A (en) Digital twin function virtualization method for 6G mobile network
CN115914237A (en) Deep learning task scheduling method, device and medium in edge environment

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