CN111812982A - Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing - Google Patents

Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing Download PDF

Info

Publication number
CN111812982A
CN111812982A CN202010669005.9A CN202010669005A CN111812982A CN 111812982 A CN111812982 A CN 111812982A CN 202010669005 A CN202010669005 A CN 202010669005A CN 111812982 A CN111812982 A CN 111812982A
Authority
CN
China
Prior art keywords
resource
task
tasks
model
quality
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
CN202010669005.9A
Other languages
Chinese (zh)
Other versions
CN111812982B (en
Inventor
高惠琳
龚军
温秀秀
刘佩云
王凡
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.)
CETC Information Science Research Institute
Original Assignee
CETC Information Science Research Institute
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 CETC Information Science Research Institute filed Critical CETC Information Science Research Institute
Priority to CN202010669005.9A priority Critical patent/CN111812982B/en
Publication of CN111812982A publication Critical patent/CN111812982A/en
Application granted granted Critical
Publication of CN111812982B publication Critical patent/CN111812982B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a dynamic sharing and intelligent allocation method for cloud manufacturing resources for intelligent manufacturing, which comprises the steps of obtaining a plurality of tasks, wherein the types of the tasks are the same; modeling the resource quality correlation of a plurality of tasks to obtain a resource quality correlation model; establishing a problem model of a plurality of tasks, and determining a multi-target multi-constraint formula according to the problem model; determining a virtual resource combination and an optimized scheduling model of the tasks according to a multi-objective multi-constraint formula; and solving the virtual combination and optimized scheduling model, and realizing dynamic sharing and intelligent allocation of resources according to a solving result. Manufacturing raw materials, equipment, products and the like virtualized and digitized through a unified description modeling and simulation technology of manufacturing resources are connected and managed through an information framework, the capability of dynamically sharing and intelligently allocating various resources in the information framework is provided, resource scheduling and allocation according to needs are achieved, and cloud scheduling of the resources and cloud of the manufacturing process are really achieved.

Description

Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing
Technical Field
The invention belongs to the technical field of intelligent manufacturing, and particularly relates to a dynamic sharing and intelligent allocation method for cloud manufacturing resources for intelligent manufacturing.
Background
Cloud manufacturing is a new knowledge-based, service-oriented, networked intelligent manufacturing model. Cloud computing is a new computing mode based on the internet, a large number of highly virtualized computing resources are managed based on a cloud computing platform to form a large resource pool for uniformly providing services, and computing services acquired at any time as required are provided for individuals and enterprise users in an internet heterogeneous and autonomous service form. If the computing resources are replaced by manufacturing resources, the computing mode and the operation mode of cloud computing can be used for manufacturing informatization, and a feasible new thought is provided for the manufacturing informatization, the service, the high efficiency and the low consumption. Manufacturing resources include manufacturing equipment of all kinds in a full lifecycle activity and models, data, software, domain knowledge, etc. in the manufacturing process. In order to realize virtualization, optimized scheduling and cooperative interconnection of manufacturing resources, technologies such as semantic Web, an embedded system technology, an Internet of things and high-efficiency calculation need to be fused. It is in this context that the concept of cloud manufacturing arises. At present, the cloud manufacturing is rarely researched internationally, so that a huge development space exists for research and application of the cloud manufacturing. Cloud manufacturing and key technologies thereof are also important issues to be solved for breakthrough development of the Chinese manufacturing industry in the next decade.
Since cloud manufacturing emphasizes greenization and intelligence, in addition to time, cost, and quality as optimization objectives for cloud manufacturing resource allocation and scheduling issues, the scheduling optimization objective of system energy consumption should be increased. The task scheduling problem and the optimization method taking energy consumption as a target become a new research direction for cloud manufacturing scheduling. According to the hierarchical division, the cloud manufacturing resource scheduling problem can be divided into four levels, namely a cloud manufacturing platform layer virtual resource scheduling problem, an inter-enterprise task and resource scheduling problem, an intra-enterprise scheduling problem and an inter-vehicle layer scheduling problem. Cloud manufacturing resource scheduling oriented to intelligent manufacturing is mainly focused on virtual resource scheduling of a cloud manufacturing platform layer and cross-enterprise task resource scheduling.
The virtual resource scheduling in the cloud manufacturing platform is divided into computing resource scheduling and manufacturing resource scheduling. For the cloud manufacturing platform computing resource scheduling, in the mapping of tasks and computing resources, a situation that a plurality of tasks run on the same computing resource at the same time may occur, that is, a computing capacity allocation situation of the computing resources must be considered. These problems all play a crucial role in the operation of the cloud manufacturing platform, but are not considered in the existing research. From the perspective of a task model, multiprocessor task scheduling can be divided into scheduling tasks that are independent of each other and scheduling tasks that are dependent on each other. The traditional scheduling algorithm is not suitable for the multi-objective optimization problem of cloud manufacturing platform computing resource scheduling, and a novel computing resource scheduling method based on an intelligent heuristic task scheduling algorithm becomes the mainstream, such as a simulated annealing algorithm, a particle swarm optimization algorithm and the like. For the scheduling of the cloud manufacturing platform manufacturing resources, the main optimization scheduling method comprises an intelligent heuristic algorithm, a task scheduling algorithm based on a game theory, a task scheduling method based on workload and the like. At present, research on tasks of manufacturing resources and service scheduling problems in a cloud manufacturing environment is still few, and the used method is mainly based on an intelligent optimization algorithm, such as a genetic algorithm, a bee colony algorithm, a particle swarm optimization algorithm, an ant colony algorithm and the like.
Disclosure of Invention
The invention aims to at least solve one of the technical problems in the prior art and provides a cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing.
The invention provides a dynamic sharing and intelligent allocation method for cloud manufacturing resources for intelligent manufacturing, which comprises the following steps:
acquiring a plurality of tasks, wherein the tasks are the same in type;
modeling the resource quality correlation of the tasks to obtain a resource quality correlation model;
establishing a problem model of the tasks, and determining a multi-target multi-constraint formula according to the problem model;
determining a virtual resource combination and an optimized scheduling model of the tasks according to the multi-objective multi-constraint formula;
and solving the virtual combination and optimized scheduling model, and realizing dynamic sharing and intelligent allocation of resources according to a solving result.
Optionally, the modeling the resource quality correlations of the plurality of tasks to obtain a resource quality correlation model includes:
respectively modeling the intra-task resource quality correlation of each task to obtain an intra-task resource quality correlation model; and the number of the first and second groups,
and modeling the inter-task resource quality correlation of the tasks to obtain an inter-task resource quality correlation model.
Optionally, the modeling the intra-task resource quality correlation of each task respectively to obtain an intra-task resource quality correlation model includes:
dividing the resource attributes of the task into two types of attribute sets of QRAS and OAS, wherein QRAS represents the attribute set related to quality, and OAS represents other attribute sets unrelated to quality;
establishing a resource quality correlation model in the task according to the attribute set of the resource, wherein the resource quality correlation model is shown in the following formula (1):
Figure BDA0002581541260000031
wherein q (VR) is a value of a quality attribute of the resource;
Value0a default quality attribute value of the resource VR when the resource quality correlation in the task is not considered;
Figure BDA0002581541260000032
other candidate resource sets which do not contain virtual resource VR in the task are obtained;
qRA belongs to QRAS as a certain quality related attribute;
the Boolean function g is a resource correlation function;
Figure BDA0002581541260000033
representing the selected resources in the candidate resource set corresponding to other subtasks in the task
Figure BDA0002581541260000034
qRA so that when the Boolean function g is true, then q (VR) takes the Value1；
The negation and operation of the recursive operation of the boolean function g satisfies the following formula (2):
Figure BDA0002581541260000035
gi＝gj()&&gk()，i≠j≠k (2)
when a plurality of resource correlation functions g are simultaneously satisfied, determining to take the maximum value or the minimum value according to the positive and negative directions of the quality attribute, and establishing the following model as shown in formula (3):
Figure BDA0002581541260000041
the time and cost calculation model considering the intra-task resource quality dependency is shown in equation (4) below:
T(VR)＝MIN{T_T0，...,T_Ti,...},gi＝true；
C(VR)＝MIN{T_C0，...,T_Ci,...},gi＝true； (4)
wherein, T _ TiAnd (C) a cost calculation model representing the intra-task resource quality correlation.
Optionally, the modeling the inter-task resource quality correlation of the plurality of tasks to obtain an inter-task resource quality correlation model includes:
according to the attribute set of the resource, establishing a resource quality correlation model between tasks, as shown in the following formula (5):
q(VR)＝{default：DefaultValue；
correlationFunc1():Value1；
correlationFunc2():Value2；
...} (5)
wherein the DefaultValue is obtained by calculation according to a formula (3);
correlationFunc1():Value1indicates that the resource selected in the 1 st scheduling process satisfies the correlation function1() When q (VR) takes Value of Value1；
When a plurality of resource correlation functions are simultaneously satisfied, determining a maximum value or a minimum value according to the positive and negative directions of the quality attribute, and establishing the following model as shown in formula (6):
Figure BDA0002581541260000042
the time and cost calculation model considering the resource dependency between tasks is shown in the following equation (7):
T(VR)＝MIN{P_T0，...,P_Ti,...},gi＝true；
C(VR)＝MIN{P_C0，...,P_Ci,...},gi＝true； (7)
wherein, P _ TiAnd C (VR) a cost calculation model for representing the resource quality correlation between tasks.
Optionally, the establishing a problem model of the plurality of tasks and determining a multi-target multi-constraint formula according to the problem model includes:
n task sets Ta ═ Ta of the same type are received within a period of time1，Ta2，...,TanAnd Price corresponding to each task is { Price }1，Price2，...,PricenThe delivery date of each task is the number of days from the current date { T }1，T2，...,TnDividing each task into J subtasks;
let execution path set P be { P ═ P1，P2，...,PnThe task set Ta is a feasible solution, and each subtask has a corresponding candidate resource set
Figure BDA0002581541260000051
Is performed with J in each candidate setKA virtual resource
Figure BDA0002581541260000052
The feasible solution satisfies the following equations (9) to (11):
m≤n (9)
T(Pi)≤T(Ta(Pi)) (10)
Figure BDA0002581541260000053
wherein, formula (9) indicates that the task is allowed to be unsatisfied, Ta (P)i) Represents a feasible solution PiCorresponding tasks, the whole formula (10) shows that each feasible solution meets the delivery date requirements of the tasks; equation (11) represents the time period [ t ]1,t2]Internal, virtual resources
Figure BDA0002581541260000054
If is by PiOccupied, then it cannot be occupied by any other feasible solution, while at [ t1,t2]Out of any other feasible solution;
determining a feasible solution P ═ { P) according to different targets1，P2，...,PmWhether the solution is the optimal solution or not is judged, and a multi-target multi-constraint formula is formed, wherein the formula (12) and the formula (13) are as follows:
the higher the profit, the better, i.e.
Figure BDA0002581541260000055
Wherein, Profit (P)i)＝Price(Ta(Pi))-C(Pi)；
The greater the number of tasks that can be satisfied, the better, namely:
MAX(Num(P)) (13)
where num (p) is the number of tasks that a feasible solution can accomplish.
Optionally, the determining, according to the multi-objective multi-constraint formula, a virtual resource combination and an optimized scheduling model of the multiple tasks includes:
changing the multi-target multi-constraint formula into a single target function in a linear weighting mode, wherein the following formula (14) is shown:
Figure BDA0002581541260000061
wherein α, β are weights, and α + β ═ 1;
the profit is normalized as shown in the following equation (15):
Figure BDA0002581541260000062
given the variables: x ═ X1,x2,...,xn) Wherein x isiIs a Boolean value, if xi1, then represents the task T αiIs executed, otherwise, it means not executed, the objective function is transformed as shown in the following equation (16):
Figure BDA0002581541260000063
constraint formula T (P)i)≤T(Ta(Pi) ) into the following formula (17):
T(Pi)*xi∑Ti*xi(17)
to constrain the formula
Figure BDA0002581541260000064
Simplifying, and dividing T into a plurality of equal parts under the assumption of scheduling in a future T time period, wherein scheduling is usually performed in a day unit in the manufacturing process, so that in any time period T e T, a variable y (i, j, k, T) is given, and the variable shows that if in the time period T, a candidate resource set is obtained
Figure BDA0002581541260000065
Is applied to task TaiIf y (i, j, k, t) is 1, otherwise it is 0, and the constraint equation is transformed to equation (18):
Figure BDA0002581541260000066
meanwhile, y (i, j, k, t) should also satisfy the following constraint, which indicates that a subtask can be completed by only one resource at most in the time period t, as shown in the following formula (19):
Figure BDA0002581541260000067
the above equations (16) to (19) form the virtual resource combination and optimization scheduling model.
Optionally, the solving the virtual combination and optimized scheduling model includes:
and solving the virtual resource combination and the optimized scheduling model by adopting a resource combination and optimized scheduling genetic algorithm which is based on real matrix coding and takes time sharing into consideration.
Optionally, the solving the virtual resource combination and optimal scheduling model by using a resource combination and optimal scheduling genetic algorithm considering time sharing based on real matrix coding includes:
given the following real matrix, as shown in equation (20) below:
Figure BDA0002581541260000071
element a in matrix AijtSubscripts have the following meanings:
i represents the identification of a task i;
j represents the jth subtask or jth virtual resource candidate set of the task;
t represents the t-th time period;
according to resource time-sharing and resource correlation, randomly generating an initial population according to a time period t;
designing a cross mutation operator according to a real number matrix adopting real number coding;
and obtaining the optimal solution of the virtual resource combination and the optimized scheduling model according to a preset adaptive value function and a genetic strategy.
Optionally, the obtaining an optimal solution of the virtual resource combination and the optimal scheduling model according to a preset adaptive value function and a genetic strategy includes:
based on the objective function, an adaptive value function is given as shown in the following equation (21):
Figure BDA0002581541260000072
for a task, there is a first penalty function as shown in equation (22):
Figure BDA0002581541260000073
mu is a first penalty factor;
defining a second penalty function for the entire feasible solution according to the first penalty function and the constraint formula, as shown in the following formula (23):
Figure BDA0002581541260000081
based on the second penalty function of the feasible solution, a fitness function of the final feasible solution is obtained, as shown in the following formula (24):
FF＝F*Pu (24)
the population genetic strategy adopts a proportional selection strategy, and adds pressure calibration to the fitness function, namely:
for an individual in Size, if the adaptive value is FFsThen the probability of its selection is, as shown in the following equation (25):
Figure BDA0002581541260000082
wherein χ is a pressure coefficient, the numeric area is [0,1 ], and the larger the numeric area is, the larger the selection pressure is.
Optionally, the cross mutation operator comprises at least one of a row cross, a column cross, a block cross, a row mutation, a column mutation, and a block mutation.
The invention relates to a method for preparing a high-performance composite material.
Drawings
Fig. 1 is a flowchart of a method for dynamically sharing and intelligently allocating cloud manufacturing resources for intelligent manufacturing according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
The background to the invention is first described below.
In a cloud manufacturing environment, the use of resources is usually accomplished by invoking corresponding services after the resources are virtualized. Aiming at the ubiquitous multi-task resource combination and optimized scheduling conditions of cloud enterprises, current research mainly focuses on resource service combination and optimized scheduling of single tasks and multi-task optimized scheduling based on the single tasks, so that the scheduling result is easy to fall into local optimum, and the result of global optimum cannot be achieved. Although some research has been considered from a global optimization perspective, there are not enough considerations in terms of the associations between resources and tasks and the time sharing shareability of resources that is unique in a cloud manufacturing environment. Therefore, it is an important problem to set up an optimized service flow aiming at the intra-task and inter-task correlation of cloud manufacturing resources and the time-sharing characteristic of resources among multiple tasks, so that the value of both the client and the cloud enterprise can be added.
When the cloud enterprise receives the customer requirement, the cloud enterprise can complete the following steps: 1) classifying the tasks; 2) decomposing a task; 3) resource allocation; 4) optimizing and scheduling resources; 5) task execution and service feedback. As described above, if the multi-task problem solution is performed based on a single task, the scheduling result is often optimal in the single task, and the global angle is not necessarily optimal.
However, when multi-task needs are met, if the tasks need to be sequenced one by simply adopting a scheduling mode facing to a single task, typical strategies include first serving, task priority marking and the like, and then resource configuration and optimal scheduling are performed on the tasks in sequence. This makes the configuration and the optimized scheduling of each resource meet the current task as the optimization target, so that when the key resource can meet multiple tasks, only the task with the front priority is met, and the optimization cannot be realized from the global perspective of the multiple tasks.
In addition, it is not only possible to consider exclusivity of resources, but also shareability of resources. Exclusivity of a resource refers to the resource being occupied or used while not being usable by other tasks. In the research results related to service computing and cloud computing, the monopoly of resources is realized in the way that the resources are used only once in one scheduling process, and no matter the time is that multiple tasks are completed simultaneously or only one task is completed. This is consistent with the scheduling characteristics of service computing and cloud computing, because in service computing and cloud computing, the time unit of one scheduling is usually in the order of seconds, even milliseconds and microseconds. However, unlike in a cloud manufacturing environment, scheduling of resources is typically performed on a daily or weekly basis, and thus, in the case of considering improvement of resource utilization rate or reuse of key resources, it should also be considered that resources can be reused by other tasks after completion of one task in the case of multitasking.
Furthermore, the relevance of resources is also taken into account. Most of the existing resource scheduling, service selection, discovery and optimization combinations are optimized and invoked based on the characteristics of the resource or service itself. However, in practice there is a correlation between resources, particularly resources associated with tasks. For example, it is often more efficient for two consecutive subtasks to use the resources of the same provider than to use the resources of different providers. Here, the same provider is the correlation manifestation of two resources.
Finally, in a cloud manufacturing environment, customer tasks may be constant, forming a stream of tasks over time, however, as with the planning cycle of a manufacturing enterprise, multiple tasks may be undertaken in a staged batch process, i.e., tasks within a period of time are processed in batches at the same time. That is, virtual resource combination and optimized scheduling need to be performed on multiple tasks at the same time, characteristics of resource sharing, resource correlation and the like are fully considered, the limitation of local optimization is broken through, and overall optimization of the multiple tasks is achieved.
Based on the reasons, the inventor of the invention designs a dynamic sharing and intelligent allocation method for cloud manufacturing resources facing intelligent manufacturing.
As shown in fig. 1, a method S100 for dynamic sharing and intelligent allocation of cloud manufacturing resources for intelligent manufacturing includes:
and S110, acquiring a plurality of tasks, wherein the tasks are the same in type.
And S120, modeling the resource quality correlation of the tasks to obtain a resource quality correlation model.
S130, establishing a problem model of the tasks, and determining a multi-target multi-constraint formula according to the problem model.
And S140, determining a virtual resource combination and an optimized scheduling model of the tasks according to the multi-objective and multi-constraint formula.
And S150, solving the virtual combination and optimized scheduling model, and realizing dynamic sharing and intelligent allocation of resources according to a solving result.
The cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing of the embodiment breaks through the defect that resources, environments and equipment in a traditional manufacturing system are limited by actual physical space, connects and manages manufacturing raw materials, equipment, products and the like virtualized and digitized through a unified description modeling and simulation technology of manufacturing resources through an information frame, provides the capability of dynamically sharing and intelligently allocating each resource in the information frame, realizes resource scheduling and allocation according to needs, and really realizes cloud scheduling of resources and cloud allocation of a manufacturing process.
Like the traditional single-task service combination and optimization scheduling problem, each task in the multi-task virtual resource combination and optimization scheduling problem in the cloud manufacturing environment also needs to be subjected to the steps of task decomposition, generation of candidate resource sets, resource combination optimization, execution path selection and the like.
Next, we will focus on the multi-tasking virtual resource composition and optimal scheduling problem for the same type of task. For the combination and optimized scheduling of the multi-task virtual resources of different types of tasks, the execution path and the candidate resource set are completely different, and the multi-task virtual resource combination and optimized scheduling can be completely executed according to the traditional single-task method. And for the mixed type tasks, the tasks can be divided into different types of tasks and the tasks of the same type in a staged manner, so that the combined scheduling of the tasks of different types and the tasks of the same type is converted.
Before the problem model is built, the quality dependence of the resource is modeled first. Corresponding models are respectively established for the intra-task resource quality correlation and the inter-task resource quality correlation.
(a) An intra-task resource quality dependency model.
Intra-task resource quality dependency refers to when a subtask STa is in a task or a class of tasks TaiA certain resource VR ofiQuality attribute QRA (VR)i) With another subtask STajA certain resource VR ofjQuality attribute QRA (VR)j) When there is some dependency and some attribute quality of at least one resource changes, it is called intra-task resource quality dependency.
It makes sense to discuss an intra-task resource quality dependency model only if intra-task resources are relevant. The ontology attributes are divided into two classes, QRAS and OAS, wherein QRAS represents the attribute set related to quality, and OAS represents the other attribute set not related to quality.
According to the attribute set of the resource, an intra-task resource quality correlation model can be established, which is described in detail as follows:
Figure BDA0002581541260000111
wherein q (VR) is a value of a quality attribute of the resource;
Value0a default quality attribute value of the resource VR when the resource quality correlation in the task is not considered;
Figure BDA0002581541260000121
other candidate resource sets which do not contain virtual resource VR in the task are obtained;
qRA belongs to QRAS as a certain quality related attribute;
the Boolean function g is a resource correlation function;
Figure BDA0002581541260000122
Value1representing the selected resources in the candidate resource set corresponding to other subtasks in the task
Figure BDA0002581541260000123
qRA so that when the Boolean function g is true, then q (VR) takes the Value1；
Obviously, there is a recursive operation in the boolean function g, and the recursive operation of the present embodiment only considers the non-sum operation, that is:
Figure BDA0002581541260000124
gi＝gj()&&gk()，i≠j≠k (2)
for boolean functions or and exclusive-or operations, it is sufficient to split up into several different correlation values or to express them by a not or and operation. From an algebraic point of view, the non-operation of the function is better than the AND operation.
When a plurality of resource correlation functions g are simultaneously satisfied, determining a maximum value or a minimum value according to the positive and negative directions of the quality attribute, and establishing the following model:
Figure BDA0002581541260000125
thus, for the present embodiment, the time and cost calculation model that considers intra-task resource quality dependencies is as follows:
T(VR)＝MIN{T_T0，...,T_Ti,...},gi＝true；
C(VR)＝MIN{T_C0，...,T_Ci,...},gi＝true； (4)
wherein, T _ TiAnd (C) a cost calculation model representing the intra-task resource quality correlation.
(b) Inter-task resource quality correlation model
The inter-task resource quality correlation refers to that in a certain task scheduling, a certain task TaiIn a certain resource VRiQuality attribute QRA (VR)i) With another subtask TajA certain resource VR ofjQuality attribute QRA (VR)j) There is some kind of dependency and when some quality attribute of at least one resource changes, it becomes the resource quality dependency between tasks, otherwise it is called as irrelevant.
It makes sense to discuss the inter-task resource quality correlation model only if the inter-task resources are correlated. According to the attribute set of the resource, an inter-task resource quality correlation model can be established, which is described in detail as follows:
q(VR)＝{default：DefaultValue；
correlationFunc1():Value1；
correlationFunc2():Value2；
...} (5)
wherein the DefaultValue is obtained by calculation according to a formula (3);
correlationFunc1():Value1indicates that the resource selected in the 1 st scheduling process satisfies the correlation function1() When q (VR) takes Value of Value1；
Similarly, when a plurality of resource correlation functions are simultaneously satisfied, determining a maximum value or a minimum value according to the positive direction and the negative direction of the quality attribute, and establishing the following model, as shown in formula (6):
Figure BDA0002581541260000131
the time and cost calculation model considering the resource dependency between tasks is shown in the following equation (7):
T(VR)＝MIN{P_T0，...,P_Ti,...},gi＝true；
C(VR)＝MIN{P_C0，...,P_Ci,...},gi＝true； (7)
wherein, P _ TiAnd C (VR) a cost calculation model for representing the resource quality correlation between tasks.
In the following, a simple example is given to illustrate the inter-task resource quality dependency model. If the number of times of one-time scheduling use of the resource is more than 1, the cost of each use is 9.5. Then the cost-quality correlation model c (vr) for the resource can be simply defined as follows:
C(VR)＝{default：DefaultValue；
UsedNum(VR)＞1:default*0.95；
...} (8)
wherein, usednum (vr) is the number of times used in one multitask scheduling process.
Description of the problemThe following were used: n task sets Ta ═ Ta of the same type are received within a period of time1，Ta2，...,TanAnd Price corresponding to each task is { Price }1，Price2，...,PricenThe delivery date of each task is the number of days from the current date { T }1，T2，...,TnDividing each task into J subtasks;
let execution path set P be { P ═ P1，P2，...,PnThe task set Ta is a feasible solution, and each subtask has a corresponding candidate resource set
Figure BDA0002581541260000141
Is performed with J in each candidate setKA virtual resource
Figure BDA0002581541260000142
The feasible solution satisfies the following equations (9) to (11):
m≤n (9)
T(Pi)≤T(Ta(Pi)) (10)
Figure BDA0002581541260000143
wherein, formula (9) indicates that the task is allowed to be unsatisfied, Ta (P)i) Represents a feasible solution PiCorresponding tasks, the whole formula (10) shows that each feasible solution meets the delivery date requirements of the tasks; equation (11) represents the time period [ t ]1,t2]Internal, virtual resources
Figure BDA0002581541260000144
If is by PiOccupied, then it cannot be occupied by any other feasible solution, while at [ t1,t2]Out of any other feasible solution;
determining a feasible solution P ═ { P) according to different targets1，P2，...,PmWhether the solution is the optimal solution or not is judged, and a multi-target multi-constraint formula is formed, wherein the formula (12) and the formula (13) are as follows:
the higher the profit, the better, i.e.
Figure BDA0002581541260000145
Wherein, Profit (P)i)＝Price(Ta(Pi))-C(Pi)；
The greater the number of tasks that can be satisfied, the better, namely:
MAX(Num(P)) (13)
where num (p) is the number of tasks that a feasible solution can accomplish.
The two objectives, not necessarily linear, may be that a profit for one task of an enterprise, in an extreme example, may be more than the sum of profits obtained for other tasks with less profit.
Under the cloud manufacturing environment, the problem of multitask-oriented virtual resource combination and optimized scheduling belongs to the problem of multi-objective multi-constraint planning. The problem can be solved by a pareto optimal method, and multiple targets can be converted into a single target by methods such as weighting and the like. If the linear weighting mode is adopted to change the multiple objectives of the objective function into a single objective function, the form is as follows:
Figure BDA0002581541260000151
wherein α, β are weights, and α + β ═ 1;
to reduce the number of variables in the problem model and to facilitate easier solution, the profits are first normalized:
Figure BDA0002581541260000152
given the variables: x ═ X1,x2,...,xn) Wherein x isiIs a Boolean value, if xi1, then represents the task T αiIs executed, otherwise, it means not executed, the objective function is transformed as shown in the following equation (16):
Figure BDA0002581541260000153
constraint formula T (P)i)≤T(Ta(Pi) ) into the following formula (17):
T(Pi)*xi≤Ti*xi(17)
to constrain the formula
Figure BDA0002581541260000154
Simplifying, and dividing T into a plurality of equal parts under the assumption of scheduling in a future T time period, wherein scheduling is usually performed in a day unit in the manufacturing process, so that in any time period T e T, a variable y (i, j, k, T) is given, and the variable shows that if in the time period T, a candidate resource set is obtained
Figure BDA0002581541260000155
Is applied to task TaiIf y (i, j, k, t) is 1, otherwise it is 0, and the constraint equation is transformed to equation (18):
Figure BDA0002581541260000156
meanwhile, y (i, j, k, t) should also satisfy the following constraint, which indicates that a subtask can be completed by only one resource at most in the time period t, as shown in the following formula (19):
Figure BDA0002581541260000161
so far, a model for multitask-oriented virtual resource combination and optimized scheduling problem in a cloud manufacturing environment is as follows:
Figure BDA0002581541260000162
s.t.T(Pi)*xi≤Ti*xi(21)
Figure BDA0002581541260000163
Figure BDA0002581541260000164
xi∈{0,1}，1≤i≤n (24)
obviously, this problem is of NP-Hard.
And then, solving the multitask-oriented virtual resource combination and optimized scheduling problem in the cloud manufacturing environment by adopting a real matrix coding-based resource combination and optimized scheduling genetic algorithm considering time sharing.
(a) Chromosomal coding
Given the following real matrix, as shown in equation (25) below:
Figure BDA0002581541260000165
the matrix A is n rows T multiplied by J columns, the rows represent n tasks in the demand, the columns are divided into T sections, each section represents a time period T, each section of time T has J columns, and the J sub-tasks or the J virtual resource candidate sets of the tasks are represented. Thus, element a in matrix AijtSubscripts have the following meanings:
i represents the identification of a task i;
j represents the jth subtask or jth virtual resource candidate set of the task;
t denotes the t-th time period.
A in the matrixijtIs taken as value of [0, Jk]If a is an integerijtNot equal to 0, it means that within the time period t, the jth virtual resource candidate set is in the jth virtual resource candidate setijtOne resource for completing task TaiElse task TaiIs not executed during the time period t.
(b) Initial population generation
The individuals in the initial population are randomly generated according to the time period t under the premise of considering resource time sharing and resource correlation, namely the individuals are randomly generated according to the time period t and the subtask j. Therefore, the randomness of individuals is guaranteed, the time-sharing of the resources considered herein and the correlation of the quality of the resources can be further considered, the situation of falling into the local optimal condition is avoided, and meanwhile, the optimal solution can be quickly approached.
Description of the algorithm:
1) the core idea of the algorithm is that the resource is generated randomly according to the line and the row step by step according to the time period t, and simultaneously the resource correlation and the time sharing property are fully considered, so that the resources after the task is completed are released for other tasks to use, and an initial population is formed.
2) The algorithm considers the cases in rows 4 and 5, respectively, how to randomly generate columns when the number of tasks is smaller and larger than the number of alternative resources in the candidate set.
3) After the first column is generated, row 7 of the algorithm is a rule that all tasks can be converted into a serial flow, and the subsequent J-1 columns of the period where t is 0 are all 0.
4) When the 12 th and 13 th rows of the algorithm indicate that the jth sub-task is completed in the t-1 period, aijtNot equal to 0, while ai(j+1)t≠0，ai(j+1)tShould be taken asi(j+1)t＝ai(j+1)(t-1)And selecting a numerical value corresponding to the optimal resource from the resources related to the quality from the residual numbers after the value taking is finished (the resources after the task is finished at the moment are released), and randomly selecting if no quality is related.
Initial population generation algorithm GenInitPop (VRS, Rel (VRS)) considering resource relevance and time sharing
Input cloud Enterprise resource set VRS, resource dependency Rel (VRS)
Output initial Total group A
Figure BDA0002581541260000171
Figure BDA0002581541260000181
(c) Cross and variance
Because the matrix adopts real number coding, the traditional Boolean value cross mutation method is not enough to effectively solve the coding cross and mutation, so the following centralized cross mutation operators are designed:
I. line crossing
Randomly generating an n-dimensional 0-1 column vector VC if VCiAnd if not, the two ith rows of the two feasible solutions of the parent item are exchanged, and if not, the two ith rows are unchanged, and two new child items are finally generated. The constraint specification is broken because the same non-zero number may appear on the same column after the row interleaving (i.e., one resource is applied on both tasks in the same time period). Therefore, the newly generated sub-items need to be subjected to constraint satisfaction repair, and the repair algorithm is as follows:
constraint satisfaction repair algorithm Constrefix (A, Constraint)
Input is population A, Constraint
Output: population A satisfying constraint conditions
Figure BDA0002581541260000182
Figure BDA0002581541260000191
Description of the algorithm:
1) the core idea of the algorithm is that the resource exclusivity violated after line crossing and line variation in the population A and the constraint completed by one resource only in one subtask are subjected to constraint satisfaction, and the corresponding value is adjusted to be an unused resource value, if the unused resource value does not exist, namely when all resources are occupied, the unused resource value is directly set to be 0, so that the constraint satisfaction is realized.
2) VC in line 5 of the Algorithmi0 means that the line has no variation and therefore does not need to be adjusted.
Column crossing
Similar to the line crossing, a 0-1 line vector VC with dimension of T multiplied by J is randomly generated if VCjtExchanging the jth columns of the two feasible solutions of the parent item if the two feasible solutions are 1, otherwise, not changing the two feasible solutions, and finally generating two new childrenAn item. It can be seen that column crossing does not violate the constraint specification.
Block crossing
The block crossing crosses by one time period t as a block. I.e. the matrix is divided into T blocks, so that a 0-1 row vector VC of one T column is randomly generated if VCtAnd exchanging the tth blocks of the two feasible solutions of the parent item, otherwise, generating two new child items. It can be seen that block intersections are similar to column intersections and do not violate the constraint rules.
Line variation of IV
Because the matrix adopts real number coding, the traditional inversion variation of boolean values cannot effectively solve the variation rule in this text, and therefore, the following variation rule needs to be designed. For non-zero values, the following variation is made, and for zero values the variation is not:
aijt＝Jk-aijt+1
thus, for row variance, an n-dimensional 0-1 column vector VC is first randomly generated if VCiAnd if not, keeping unchanged, and finally generating a new child item. Since the same non-zero number may appear on the same column after mutation (i.e., one resource is applied to two tasks at the same time), the constraint specification is broken, and therefore, the above constraint on the generated new sub-item is satisfied with the repairing procedure.
Variation of V.column
Similar to the row variation, a 0-1 row vector VC with dimension of T × J is randomly generated if VCjtAnd if not, keeping unchanged, and finally generating a new child item. It can be seen that column variation does not violate constraint rules
Block variation of VI
Block mutation is performed by taking a time period t as a block. I.e. the matrix is divided into T blocks, so that a 0-1 row vector VC of one T column is randomly generated if VCtAnd (4) carrying out mutation on a non-zero value in the tth block of the parent item, otherwise, not changing, and finally generating a new child item. It can be seen that the block variation andcolumn variations are similar and do not break the constraint rules.
(d) Adaptive value function and genetic strategy
Based on the objective function, the following fitness function is given:
Figure BDA0002581541260000211
meanwhile, in order to better select a new generation of population, for a certain task, a penalty function is as follows:
Figure BDA0002581541260000212
mu is the first penalty factor when TaiWhen the constraint formula (21) is satisfied, punishment is not carried out, otherwise punishment is carried out, and when x is satisfiediWhen 1, | Ti-T(Pi) The larger the | is, the larger μ is. So that if the constraint equation (13) is to be satisfied, the corresponding number of tasks needs to be reduced or the execution time or sequence of the tasks needs to be adjusted.
Thus, a penalty function for the entire feasible solution can be defined as:
Figure BDA0002581541260000213
based on the penalty function of the feasible solution, the fitness of the final feasible solution is obtained as follows:
FF＝F*Pu (29)
the population genetic strategy adopts a proportional selection strategy, and adds pressure calibration to a fitness function, so that the algorithm can approach to an optimal solution more quickly.
For an individual in Size, if the adaptive value is FFsThen the probability of its selection is:
Figure BDA0002581541260000214
wherein χ is a pressure coefficient, the numeric area is [0,1 ], and the larger the numeric area is, the larger the selection pressure is.
It will be understood that the above embodiments are merely exemplary embodiments taken to illustrate the principles of the present invention, which is not limited thereto. It will be apparent to those skilled in the art that various modifications and improvements can be made without departing from the spirit and substance of the invention, and these modifications and improvements are also considered to be within the scope of the invention.

Claims (10)

1. A cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing is characterized by comprising the following steps:
acquiring a plurality of tasks, wherein the tasks are the same in type;
modeling the resource quality correlation of the tasks to obtain a resource quality correlation model;
establishing a problem model of the tasks, and determining a multi-target multi-constraint formula according to the problem model;
determining a virtual resource combination and an optimized scheduling model of the tasks according to the multi-objective multi-constraint formula;
and solving the virtual combination and optimized scheduling model, and realizing dynamic sharing and intelligent allocation of resources according to a solving result.
2. The method of claim 1, wherein the modeling the resource quality dependencies of the plurality of tasks to obtain a resource quality dependency model comprises:
respectively modeling the intra-task resource quality correlation of each task to obtain an intra-task resource quality correlation model; and the number of the first and second groups,
and modeling the inter-task resource quality correlation of the tasks to obtain an inter-task resource quality correlation model.
3. The method of claim 2, wherein the separately modeling intra-task resource quality dependencies for each of the tasks to obtain an intra-task resource quality dependency model comprises:
dividing the resource attributes of the task into two types of attribute sets of QRAS and OAS, wherein QRAS represents the attribute set related to quality, and OAS represents other attribute sets unrelated to quality;
establishing a resource quality correlation model in the task according to the attribute set of the resource, wherein the resource quality correlation model is shown in the following formula (1):
Figure FDA0002581541250000021
wherein q (VR) is a value of a quality attribute of the resource;
Value0a default quality attribute value of the resource VR when the resource quality correlation in the task is not considered;
Figure FDA0002581541250000022
other candidate resource sets which do not contain virtual resource VR in the task are obtained;
qRA belongs to QRAS as a certain quality related attribute;
the Boolean function g is a resource correlation function;
Figure FDA0002581541250000023
Value1representing the selected resources in the candidate resource set corresponding to other subtasks in the task
Figure FDA0002581541250000024
qRA so that when the Boolean function g is true, then q (VR) takes the Value1；
The negation and operation of the recursive operation of the boolean function g satisfies the following formula (2):
Figure FDA0002581541250000025
gi＝gj()&&gk()，i≠j≠k (2)
when a plurality of resource correlation functions g are simultaneously satisfied, determining to take the maximum value or the minimum value according to the positive and negative directions of the quality attribute, and establishing the following model as shown in formula (3):
Figure FDA0002581541250000026
the time and cost calculation model considering the intra-task resource quality dependency is shown in equation (4) below:
T(VR)＝MIN{T_T0，...,T_Ti,...},gi＝true；
C(VR)＝MIN{T_C0，...,T_Ci,...},gi＝true； (4)
wherein, T _ TiAnd (C) a cost calculation model representing the intra-task resource quality correlation.
4. The method of claim 2, wherein the modeling the inter-task resource quality dependencies of the plurality of tasks to obtain an inter-task resource quality dependency model comprises:
according to the attribute set of the resource, establishing a resource quality correlation model between tasks, as shown in the following formula (5):
q(VR)＝{default：DefaultValue；
correlationFunc1():Value1；
correlationFunc2():Value2；
...} (5)
wherein the DefaultValue is obtained by calculation according to a formula (3);
correlationFunc1():Value1indicates that the resource selected in the 1 st scheduling process satisfies the correlation function1() When q (VR) takes Value of Value1；
When a plurality of resource correlation functions are simultaneously satisfied, determining a maximum value or a minimum value according to the positive and negative directions of the quality attribute, and establishing the following model as shown in formula (6):
Figure FDA0002581541250000031
the time and cost calculation model considering the resource dependency between tasks is shown in the following equation (7):
T(VR)＝MIN{P_T0，...,P_Ti,...},gi＝true；
C(VR)＝MIN{P_C0，...,P_Ci,...},gi＝true； (7)
wherein, P _ TiAnd C (VR) a cost calculation model for representing the resource quality correlation between tasks.
5. The method of any of claims 1 to 4, wherein said establishing a problem model for said plurality of tasks and determining a multi-objective multi-constraint equation from said problem model comprises:
n task sets Ta ═ Ta of the same type are received within a period of time1，Ta2，...,TanAnd Price corresponding to each task is { Price }1，Price2，...,PricenThe delivery date of each task is the number of days from the current date { T }1，T2，...,TnDividing each task into J subtasks;
let execution path set P be { P ═ P1，P2，...,PnThe task set Ta is a feasible solution, and each subtask has a corresponding candidate resource set
Figure FDA0002581541250000041
Is performed with J in each candidate setKA virtual resource
Figure FDA0002581541250000042
The feasible solution satisfies the following equations (9) to (11):
m≤n (9)
T(Pi)≤T(Ta(Pi)) (10)
Figure FDA0002581541250000043
wherein, formula (9) indicates that the task is allowed to be unsatisfied, Ta (P)i) Represents a feasible solution PiCorresponding tasks, the whole formula (10) shows that each feasible solution meets the delivery date requirements of the tasks; equation (11) represents the time period [ t ]1,t2]Internal, virtual resources
Figure FDA0002581541250000044
If is by PiOccupied, then it cannot be occupied by any other feasible solution, while at [ t1,t2]Out of any other feasible solution;
determining a feasible solution P ═ { P) according to different targets1，P2，...,PmWhether the solution is the optimal solution or not is judged, and a multi-target multi-constraint formula is formed, wherein the formula (12) and the formula (13) are as follows:
the higher the profit, the better, i.e.
Figure FDA0002581541250000045
Wherein, Profit (P)i)＝Price(Ta(Pi))-C(Pi)；
The greater the number of tasks that can be satisfied, the better, namely:
MAX(Num(P)) (13)
where num (p) is the number of tasks that a feasible solution can accomplish.
6. The method of claim 5, wherein determining the virtual resource combination and the optimal scheduling model for the plurality of tasks according to the multi-objective multi-constraint formula comprises:
changing the multi-target multi-constraint formula into a single target function in a linear weighting mode, wherein the following formula (14) is shown:
Figure FDA0002581541250000046
wherein α, β are weights, and α + β ═ 1;
the profit is normalized as shown in the following equation (15):
Figure FDA0002581541250000051
given the variables: x ═ X1,x2,...,xn) Wherein x isiIs a Boolean value, if xi1, then represents the task T αiIs executed, otherwise, it means not executed, the objective function is transformed as shown in the following equation (16):
Figure FDA0002581541250000052
constraint formula T (P)i)≤T(Ta(Pi) ) into the following formula (17):
T(Pi)*xi≤Ti*xi(17)
to constrain the formula
Figure FDA0002581541250000053
Simplifying, and dividing T into a plurality of equal parts under the assumption of scheduling in a future T time period, wherein scheduling is usually performed in a day unit in the manufacturing process, so that in any time period T e T, a variable y (i, j, k, T) is given, and the variable shows that if in the time period T, a candidate resource set is obtained
Figure FDA0002581541250000054
Is applied to task TaiIf y (i, j, k, t) is 1, otherwise it is 0, and the constraint equation is transformed to equation (18):
Figure FDA0002581541250000055
meanwhile, y (i, j, k, t) should also satisfy the following constraint, which indicates that a subtask can be completed by only one resource at most in the time period t, as shown in the following formula (19):
Figure FDA0002581541250000056
the above equations (16) to (19) form the virtual resource combination and optimization scheduling model.
7. The method of claim 6, wherein solving the virtual portfolio and optimization scheduling model comprises:
and solving the virtual resource combination and the optimized scheduling model by adopting a resource combination and optimized scheduling genetic algorithm which is based on real matrix coding and takes time sharing into consideration.
8. The method of claim 7, wherein solving the virtual resource combination and optimal scheduling model using a real matrix coding-based time-sharing-considered resource combination and optimal scheduling genetic algorithm comprises:
given the following real matrix, as shown in equation (20) below:
Figure FDA0002581541250000061
element a in matrix AijtSubscripts have the following meanings:
i represents the identification of a task i;
j represents the jth subtask or jth virtual resource candidate set of the task;
t represents the t-th time period;
according to resource time-sharing and resource correlation, randomly generating an initial population according to a time period t;
designing a cross mutation operator according to a real number matrix adopting real number coding;
and obtaining the optimal solution of the virtual resource combination and the optimized scheduling model according to a preset adaptive value function and a genetic strategy.
9. The method of claim 8, wherein obtaining the optimal solution of the virtual resource combination and the optimal scheduling model according to a preset fitness function and a genetic strategy comprises:
based on the objective function, an adaptive value function is given as shown in the following equation (21):
Figure FDA0002581541250000062
for a task, there is a first penalty function as shown in equation (22):
Figure FDA0002581541250000063
mu is a first penalty factor;
defining a second penalty function for the entire feasible solution according to the first penalty function and the constraint formula, as shown in the following formula (23):
Figure FDA0002581541250000071
based on the second penalty function of the feasible solution, a fitness function of the final feasible solution is obtained, as shown in the following formula (24):
FF＝F*Pu (24)
the population genetic strategy adopts a proportional selection strategy, and adds pressure calibration to the fitness function, namely:
for an individual in Size, if the adaptive value is FFsThen it selectsThe probability of (d) is as shown in the following equation (25):
Figure FDA0002581541250000072
wherein χ is a pressure coefficient, the numeric area is [0,1 ], and the larger the numeric area is, the larger the selection pressure is.
10. The method of claim 8, wherein the cross mutation operator comprises at least one of a row cross, a column cross, a block cross, a row mutation, a column mutation, and a block mutation.
CN202010669005.9A 2020-07-13 2020-07-13 Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing Active CN111812982B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010669005.9A CN111812982B (en) 2020-07-13 2020-07-13 Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010669005.9A CN111812982B (en) 2020-07-13 2020-07-13 Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing

Publications (2)

Publication Number Publication Date
CN111812982A true CN111812982A (en) 2020-10-23
CN111812982B CN111812982B (en) 2022-08-26

Family

ID=72842443

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010669005.9A Active CN111812982B (en) 2020-07-13 2020-07-13 Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing

Country Status (1)

Country Link
CN (1) CN111812982B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743646A (en) * 2021-07-19 2021-12-03 北京工业大学 Enterprise production resource reasonable configuration cloud platform scheduling method based on processing and manufacturing granularity
CN114327839A (en) * 2022-03-11 2022-04-12 阿里巴巴达摩院(杭州)科技有限公司 Task optimization method and system
WO2023185971A1 (en) * 2022-03-31 2023-10-05 阿里巴巴（中国）有限公司 Task optimization method and apparatus
CN118445031A (en) * 2024-07-05 2024-08-06 知鱼智联科技股份有限公司 Calculation force analysis method and system based on big data

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102360462A (en) * 2011-09-30 2012-02-22 北京理工大学 Virtual resource model based on cloud manufacturing and modeling method thereof
CN107133705A (en) * 2017-07-04 2017-09-05 西北工业大学 The intelligent packaging service modeling of Oriented Green logistics and dispatching task dynamic optimization method
CN110059942A (en) * 2019-04-02 2019-07-26 南京邮电大学 A kind of cloud manufacturing recourses service Optimization Scheduling based on fuzzy multiobjective optimization
CN110298589A (en) * 2019-07-01 2019-10-01 河海大学常州校区 Based on heredity-ant colony blending algorithm dynamic Service resource regulating method
CN110471762A (en) * 2019-07-26 2019-11-19 南京工程学院 A kind of cloud resource distribution method and system based on multiple-objection optimization
CN110866591A (en) * 2019-10-28 2020-03-06 浙江大学 Method for carrying out prospective cloud manufacturing service lease configuration based on demand prediction

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102360462A (en) * 2011-09-30 2012-02-22 北京理工大学 Virtual resource model based on cloud manufacturing and modeling method thereof
CN107133705A (en) * 2017-07-04 2017-09-05 西北工业大学 The intelligent packaging service modeling of Oriented Green logistics and dispatching task dynamic optimization method
CN110059942A (en) * 2019-04-02 2019-07-26 南京邮电大学 A kind of cloud manufacturing recourses service Optimization Scheduling based on fuzzy multiobjective optimization
CN110298589A (en) * 2019-07-01 2019-10-01 河海大学常州校区 Based on heredity-ant colony blending algorithm dynamic Service resource regulating method
CN110471762A (en) * 2019-07-26 2019-11-19 南京工程学院 A kind of cloud resource distribution method and system based on multiple-objection optimization
CN110866591A (en) * 2019-10-28 2020-03-06 浙江大学 Method for carrying out prospective cloud manufacturing service lease configuration based on demand prediction

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
胡艳娟等: "云制造环境下的资源调度研究综述", 《制造技术与机床》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743646A (en) * 2021-07-19 2021-12-03 北京工业大学 Enterprise production resource reasonable configuration cloud platform scheduling method based on processing and manufacturing granularity
CN114327839A (en) * 2022-03-11 2022-04-12 阿里巴巴达摩院(杭州)科技有限公司 Task optimization method and system
CN114327839B (en) * 2022-03-11 2022-08-16 阿里巴巴达摩院(杭州)科技有限公司 Task optimization method and system
WO2023185971A1 (en) * 2022-03-31 2023-10-05 阿里巴巴（中国）有限公司 Task optimization method and apparatus
CN118445031A (en) * 2024-07-05 2024-08-06 知鱼智联科技股份有限公司 Calculation force analysis method and system based on big data

Also Published As

Publication number Publication date
CN111812982B (en) 2022-08-26

Similar Documents

Publication Publication Date Title
CN111812982B (en) Cloud manufacturing resource dynamic sharing and intelligent allocation method for intelligent manufacturing
Chen et al. Self-adaptive resource allocation for cloud-based software services based on iterative QoS prediction model
Pourghebleh et al. The importance of nature-inspired meta-heuristic algorithms for solving virtual machine consolidation problem in cloud environments
Hamad et al. Genetic-based task scheduling algorithm in cloud computing environment
Mahmoud et al. Multiobjective task scheduling in cloud environment using decision tree algorithm
Panwar et al. TOPSIS–PSO inspired non-preemptive tasks scheduling algorithm in cloud environment
Malik et al. Comparison of task scheduling algorithms in cloud environment
Cui et al. Multiple DAGs workflow scheduling algorithm based on reinforcement learning in cloud computing
Tong et al. DDQN-TS: A novel bi-objective intelligent scheduling algorithm in the cloud environment
Zhang et al. Cost-aware scheduling systems for real-time workflows in cloud: An approach based on Genetic Algorithm and Deep Reinforcement Learning
Belgacem et al. Dynamic resource allocation method based on symbiotic organism search algorithm in cloud computing
Chakravarthi et al. TOPSIS inspired cost-efficient concurrent workflow scheduling algorithm in cloud
Li et al. Chaotic-nondominated-sorting owl search algorithm for energy-aware multi-workflow scheduling in hybrid clouds
Asghari et al. Combined use of coral reefs optimization and reinforcement learning for improving resource utilization and load balancing in cloud environments
Shahane Optimizing Cloud Resource Allocation: A Comparative Analysis of AI-Driven Techniques
Nazeri et al. Energy aware resource provisioning for multi-criteria scheduling in cloud computing
Ben Alla et al. A priority based task scheduling in cloud computing using a hybrid MCDM model
Chhabra et al. QoS-aware energy-efficient task scheduling on HPC cloud infrastructures using swarm-intelligence meta-heuristics
Sundar Rajan Design and implementation of fuzzy priority deadline job scheduling algorithm in heterogeneous grid computing
Singhal et al. A job scheduling algorithm based on rock hyrax optimization in cloud computing
Rizvi et al. Intelligent salp swarm scheduler with fitness based quasi-reflection method for scientific workflows in hybrid cloud-fog environment
Zavvar et al. Fuzzy logic-based algorithm resource scheduling for improving the reliability of cloud computing
Antonova et al. Development of a method and a software for decision-making, system modeling and planning of business processes
Hashemi et al. Multi-objective Optimization for Computer Security and Privacy.
Zhang et al. A self-adaptive exception adjustment approach of multi-core value nets in industry alliance

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