CN104199820A - Cloud platform MapReduce workflow scheduling optimizing method - Google Patents
Cloud platform MapReduce workflow scheduling optimizing method Download PDFInfo
- Publication number
- CN104199820A CN104199820A CN201410315763.5A CN201410315763A CN104199820A CN 104199820 A CN104199820 A CN 104199820A CN 201410315763 A CN201410315763 A CN 201410315763A CN 104199820 A CN104199820 A CN 104199820A
- Authority
- CN
- China
- Prior art keywords
- workflow
- node
- cloud platform
- population
- new
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5066—Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Abstract
The invention relates to big data calculation and discloses a cloud platform MapReduce workflow scheduling optimizing method. The platform MapReduce workflow scheduling optimizing method approximately comprises the steps of conducting reconstruction, wherein existing workflow is reconstructed, so that new workflow is obtained; conducting optimization, wherein the workflow is optimized according to the genetic algorithm; obtaining historical data, wherein the historical data are reserved by recoding the historical data or by recording relevant data of a regression model after the regression model is established. In this way, different individuals can be generated through a part of the historical data during optimization. The platform MapReduce workflow scheduling optimizing method has the advantages that the operating time of the workflow is considered, cost for cluster renting required when calculation is conducted on a cloud platform is also considered, the optimization effect is good, and the problem that the efficiency is not high when workflow scheduling is conducted on a large cloud calculation platform can be solved fundamentally.
Description
Technical field
The present invention relates to large data and calculate field, particularly a kind of cloud platform MapReduce workflow schedule optimization method, has improved the optimization efficiency of the workflow schedule on cloud platform effectively.
Background technology
Along with take generation and the development of the novel information published method that Internet of Things, social network sites SNS, bioinformatics be representative, the data class of human society and quantity are with the speed increment of explosion type, and large data age arrives.At present, for large data, not yet there is a generally acknowledged definition, the difference of the concepts such as it and traditional " mass data ", " ultra-large data ", is mainly reflected in large data and need to possesses following three features: scale (volume), diversity (variety) and high speed (velocity).According to statistics, New York Stock Exchange produces the transaction data of about 1TB every day, and company of Baidu data to be processed every day reach 10~100PB.Large data are calculated and can be divided into single job single step calculating, single job iterative computation and the calculating of many job workflow etc., and each operation can be calculated to accelerate the speed of operation by a plurality of tasks in parallel, i.e. each operation can consist of the task of some data parallels.
" cloud computing provides computing platform for large data; the service providing to user by internet is provided for it, comprises that infrastructure serve (Infrastructure as a Service), platform serves (Platform as a Service) and software serve (Software as a Service)." cloud computing, by network, is used the mode of (Pay-as-you-go), for global user provides the information service based on effectiveness to pay.
Difference by tupe is divided, and the framework of processing large data can be divided into stream and process (stream processing) framework and batch processing (batch processing) framework.Batch processing is first processing (store-then-process) after data storage again, it is after data produce, directly to process (straight through processing) that stream is processed, in stream is processed, the value of data can die-off as time goes by.Large datamation stream can consist of batch processing job or stream processing operation, and existing large data processing optimization method, only for single homework, does not consider that the cluster rent when cloud platform moves is used.
In view of the above problems, in the present invention, we intend being optimized for performance and the expense of large batch work for the treatment of stream on cloud platform, to researching and developing a kind of novel method for optimizing scheduling that can more effectively maintain the optimization efficiency under former having ready conditions in scheduling process.
Summary of the invention
The present invention is directed in prior art, optimization method depends on starting condition, the shortcoming that its effect of optimization meeting time to time change even weakens, a kind of cloud platform MapReduce workflow schedule optimization method is provided, more stable effect of optimization can be provided, effectively improve the optimization efficiency of workflow schedule.
For achieving the above object, the present invention can take following technical proposals:
A cloud platform MapReduce workflow schedule optimization method, comprises following concrete steps:
Reconstruction step: the workflow W that at least comprises an operation that user is submitted to is reconstructed into a new workflow G, and described reconstruct comprises:
New jobs node in workflow forms set V, and the directed edge between the node of the directed acyclic graph that the new operation of take is node forms set E, and described new operation comprises beginning operation J
entry, synchronization job J
syn, finish operation J
exitand branch operation J
bran, described beginning operation J
entryrefer in workflow W without any the operation of father node described end operation J
exitrefer in workflow W without any the operation of child node described synchronization job J
synhave father node and child node simultaneously, and possess father node quantity and be more than or equal to two or child node quantity and be more than or equal to the character of two, described branch operation J
branrefer to complementary simple operation J
simset, described in interdepend and refer to different simple operation J
simdirected edge can connected component make all in the industry simple operations, described simple operation J
simrefer to the operation of only having a father node and a child node in workflow W;
The size of the input data set of All Jobs in calculation workflow G, and by the big or small composition of vector S of described input data set;
Optimization Steps: produce initial population, described initial population refers to by the operation in workflow G is composed to the group of individuals that random initial value obtains; The quantity that expands individuality in initial population by producing new individual mode obtains population of future generation, and described new individuality refers to the new individuality being produced by the mode that random point intersects and/or random point makes a variation; Calculate respectively the working time of all individualities in described population of future generation, at least one individuality of choosing in described population of future generation is exported as optimum results.
In embodiments of the invention, also comprise that historical data obtains step;
Described historical data obtains step and comprises: the workflow W that selects arbitrarily a user to submit to; With different operation configuration parameters and cluster virtual machine node number, move respectively the operation in described workflow W; The operation result of the operation in described workflow W is preserved.
In embodiments of the invention, described historical data obtains step and also comprises: to running on the operation result of the operation in the described workflow W under different operation configuration parameters and cluster virtual machine node number, carry out matching, obtain the parameter of curve after matching.
In embodiments of the invention, described operation configuration parameter comprises Mapper quantity N
m, Reducer quantity N
r, input data big or small S
inputand the number N of cluster virtual machine node
cluster, described input data are stored with the form of piecemeal, described N
m, N
ror N
clusterbe not more than N
block, described N
blockrefer to the piecemeal number of described input data.
In embodiments of the invention, described input data are carried out piecemeal with the size of every of 64MB.
In embodiments of the invention, the chromosome of described initial population or population of future generation is { N
m1, N
r1, N
cluster1, N
m2, N
r2, N
cluster2..., N
m (k+1), N
r (k+1), N
cluster (k+1).
In embodiments of the invention, in described chromosome, each element comprises two digits.
The present invention has following remarkable technique effect:
Effect of optimization is good, and stability is high, by the method for profile analysis of task, in the situation of different configuration informations and Virtual Cluster scale, moves same task, obtains working time.Use least square method to carry out multiple linear regression, use a model and predict the working time under new configuration parameter.Carrying out along with operation, the continuous accumulation of historical data, can effectively guarantee that dispatching method can adjust with different workflows, thereby reduce the variation of workflow for the impact of optimization efficiency, reach by this method the object that improves global optimization efficiency.
By improved genetic algorithm, obtain the approximate optimal solution of each task configuration information, fast convergence rate, computing velocity is fast, and the present invention not only considers the working time of workflow operation, can also consider the cost of renting in cloud computing epoch infrastructure.
Accompanying drawing explanation
Fig. 1 is the schematic flow sheet of cloud platform MapReduce workflow schedule optimization method.
Embodiment
Below in conjunction with embodiment, the present invention is described in further detail.
Embodiment 1
A cloud platform MapReduce workflow schedule optimization method, as shown in Figure 1, comprises following concrete steps:
Reconstruction step 100: the core of this step is that the workflow W that user is submitted to is reconstructed, thereby generate one and there is brand new, the new workflow that can adapt to better genetic Optimization Algorithm, particularly, the workflow W that at least comprises an operation that user is submitted to is reconstructed into a new workflow G, as a kind of optional scheme, workflow W can be expressed as W (Γ, Λ, s, d), Γ is task-set, represent the set of All Jobs in workflow W, herein, using an operation as a task, and be considered as the node of the directed acyclic graph of this workflow W, Λ is the set of oriented change, represent the connection between two nodes arbitrarily in directed acyclic graph, s represents the size of the original input data collection of workflow W, d represents the operation closing time of workflow W, it is the end time of workflow W operation.Wherein, described reconstruct comprises following concrete steps:
New jobs node in workflow forms set V, and the directed edge between the node of the directed acyclic graph that the new operation of take is node forms set E, and described new operation comprises beginning operation J
entry, synchronization job J
syn, finish operation J
exitand branch operation J
bran, described beginning operation J
entryrefer in workflow W without any the operation of father node described end operation J
exitrefer in workflow W without any the operation of child node described synchronization job J
synhave father node and child node simultaneously, and possess father node quantity and be more than or equal to two or child node quantity and be more than or equal to the character of two, described branch operation J
branrefer to complementary simple operation J
simset, described in interdepend and refer to different simple operation J
simdirected edge can connected component make all in the industry simple operations, described simple operation J
simrefer to the operation of only having a father node and a knot node in workflow W;
The size of the input data set of All Jobs in calculation workflow G, and by the big or small composition of vector S of described input data set; Above-mentioned reconstruct is completed by workflow schedule device, and this workflow schedule device also comprises and calculate approximate optimal solution scheduling, and uses this approximate optimal solution to dispatch corresponding parameter to be dispatched on cloud platform and to move.
Workflow G after above-mentioned reconstruct can be expressed as G (Γ, Λ, V, E, S, d), and in order to reduce the expense that between operation, data transmission produces, in workflow G, the operation that same branch operation comprises is used identical cluster virtual machine to move.
Optimization Steps 200: produce initial population, described initial population refers to by the operation in workflow G is composed to the group of individuals that random initial value obtains; The quantity that expands individuality in initial population by producing new individual mode obtains population of future generation, and described new individuality refers to the new individuality being produced by the mode that random point intersects and/or random point makes a variation; Calculate respectively the working time of all individualities in described population of future generation, at least one individuality of choosing in described population of future generation is exported as optimum results.In addition, as the selectable scheme of another kind, described optimum results also can obtain in the following manner: calculate each of the last population of future generation generating individual the needed rent use of work flow operation of corresponding scheduling strategy, choose in population of future generation cluster rent with minimum individuality as last scheduling strategy.
As an optional scheme, normally, the number of setting initial population is n, by random point, intersects and/or the number of the population of future generation that the mode of random point variation produces is 2n.Wherein,
Random point intersection refers to the new individuality that adopts following mode to generate, mode described here comprises, on individual chromosome, set the point of crossing of at least one, this point of crossing is normally set at random, then successively the part of both sides, point of crossing is exchanged, generate a new chromosome.
Random point variation refer to chromosome with binary-coded system in, it becomes chromosomal some genes into 0 by 1 randomly, or becomes 1 by 0.By mutation operation, can guarantee the diversity of genetic type in colony, so that search can be carried out in large as far as possible space, avoid being lost in hereditary information useful in search and be absorbed in local solution, obtain the optimization answer that quality is higher.
Further, as a kind of individual choice method that generates population of future generation, the method comprises: when the number of population of future generation is 2n, calculate the critical path of the weighting directed acyclic graph of each individual corresponding scheduling strategy in this next generation population as the work flow operation time, the n individuality of choosing minimum working time becomes the individuality that population of future generation comprises, and upgrading population of future generation, ExecTime working time that this place is recorded can be used following regression model to calculate.In addition, the individual choice method of the generation population of future generation of comparatively simplifying as another, can, simply by calculating each individual corresponding required ExecTime working time of dispatching algorithm, choose the several body of ExecTime minimum as the individuality of population of future generation.
Further, as a kind of prioritization scheme, described Optimization Steps 200 also comprises the iterative step 400 with set point number, and described iterative step 400 comprises:
Generate after population of future generation, calculate required ExecTime working time of each individual corresponding dispatching algorithm in this next generation population, retain less several of the numerical value of ExecTime, be preferably n individuality; N the individual individuality as new initial population of usining in this next generation population is to complete iterative process one time.
Typically, above-mentioned iterative process need to repeat 6-10 time, so that the operation of Optimization Work stream G.
Further, for the quasi-optimal workflow schedule of mould more exactly, described method also comprises that historical data obtains step 300;
Described historical data obtains step 300 and comprises: the workflow W that selects arbitrarily a user to submit to; With different operation configuration parameters and cluster virtual machine node number, move respectively the operation in described workflow W; The operation result of the operation in described workflow W is preserved.
Further, to running on the operation result of the operation in the described workflow W under different operation configuration parameters and cluster virtual machine node number, carry out matching, obtain the parameter of curve after matching.Further, in order to reflect more exactly workflow, described operation configuration parameter also comprises the input data set of described workflow W and the big or small ratio I ORate of output data set, similarly as operation result, preserves.
Described operation configuration parameter comprises Mapper quantity N
m, Reducer quantity N
r, input data big or small S
inputand the number N of cluster virtual machine node
cluster, described input data are stored with the form of piecemeal, described N
m, N
ror N
clusterbe not more than N
block, described N
blockrefer to the piecemeal number of described input data.This operation configuration parameter is associated with specific operation J.
As a kind of optional scheme, above-mentioned fit procedure need to be used multiple linear regression model, preferably, can adopt following regression model ExecTime=a * N
m+ b * N
r+ c * N
cluster+ d * S
input+ e, wherein, ExecTime is the working time of certain operation in described workflow W.In fit procedure, after having accumulated abundant historical data, can calculate value or the span of parameters in above-mentioned regression model; Setting the size of output data set and the size of input data set in described workflow W is relation in direct ratio, can be obtained exporting the big or small S of data by above-mentioned regression model
output, order
finally, parameter and the ratio I ORate of operation J and regression model associated with it are all preserved, normally deposit in the database table of a Job by name.
Described input data are carried out piecemeal with the size of every of 64MB, i.e. piecemeal number
The chromosome of described initial population or population of future generation is { N
m1, N
r1, N
cluster1, N
m2, N
r2, N
cluster2..., N
m (k+l), N
r (k+l), N
cluster (k+l), wherein, (k+l) refer to the summation that starts the quantity of operation, synchronization job, end operation and branch operation in workflow W.
In described chromosome, each element comprises two digits, and in other words, each element in chromosome only only includes a two digits, is convenient to the later stage to carry out chromosome operation.Single chromosomal length is 6 (k+l).
In a word, the foregoing is only preferred embodiment of the present invention, all equalizations of doing according to the present patent application the scope of the claims change and modify, and all should belong to the covering scope of patent of the present invention.
Claims (7)
1. a cloud platform MapReduce workflow schedule optimization method, is characterized in that, comprises following concrete steps:
Reconstruction step (100): the workflow W that at least comprises an operation that user is submitted to is reconstructed into a new workflow G, and described reconstruct comprises:
New jobs node in workflow forms set V, and the directed edge between the node of the directed acyclic graph that the new operation of take is node forms set E, and described new operation comprises beginning operation J
entry, synchronization job J
syn, finish operation J
exitand branch operation J
bran, described beginning operation J
entryrefer in workflow W without any the operation of father node described end operation J
exitrefer in workflow W without any the operation of child node described synchronization job J
synhave father node and child node simultaneously, and possess father node quantity and be more than or equal to two or child node quantity and be more than or equal to the character of two, described branch operation J
branrefer to complementary simple operation J
simset, described in interdepend and refer to different simple operation J
simbetween directed edge can connected component make all in the industry simple operations, described simple operation J
simrefer to the operation of only having a father node and a child node in workflow W;
The size of the input data set of All Jobs in calculation workflow G, and by the big or small composition of vector S of described input data set;
Optimization Steps (200): produce initial population, described initial population refers to by the operation in workflow G is composed to the group of individuals that random initial value obtains; The quantity that expands individuality in initial population by producing new individual mode obtains population of future generation, and described new individuality refers to the new individuality being produced by the mode that random point intersects and/or random point makes a variation; Calculate respectively the working time of all individualities in described population of future generation, at least one individuality of choosing in described population of future generation is exported as optimum results.
2. cloud platform MapReduce workflow schedule optimization method according to claim 1, is characterized in that, also comprises that historical data obtains step (300);
Described historical data obtains step (300) and comprising: the workflow W that selects arbitrarily a user to submit to;
With different operation configuration parameters and cluster virtual machine node number, move respectively the operation in described workflow W; The operation result of the operation in described workflow W is preserved.
3. cloud platform MapReduce workflow schedule optimization method according to claim 2, it is characterized in that, described historical data obtains step (300) and also comprises: to running on the operation result of the operation in the described workflow W under different operation configuration parameters and cluster virtual machine node number, carry out matching, obtain the parameter of curve after matching.
4. cloud platform MapReduce workflow schedule optimization method according to claim 2, is characterized in that, described operation configuration parameter comprises Mapper quantity N
m, Reducer quantity N
r, input data big or small S
inputand the number N of cluster virtual machine node
cluster, described input data are stored with the form of piecemeal, described N
m, N
ror N
clusterbe not more than N
block, described N
blockrefer to the piecemeal number of described input data.
5. cloud platform MapReduce workflow schedule optimization method according to claim 4, is characterized in that, described input data are carried out piecemeal with the size of every of 64MB.
6. cloud platform MapReduce workflow schedule optimization method according to claim 1, is characterized in that, the chromosome of described initial population or population of future generation is
{N
M1,N
R1,N
Cluster1,N
M2,N
R2,N
Cluster2,…,N
M(k+1),N
R(k+1),N
Cluster(k+1)}。
7. cloud platform MapReduce workflow schedule optimization method according to claim 6, is characterized in that, in described chromosome, each element comprises two digits.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410315763.5A CN104199820A (en) | 2014-07-03 | 2014-07-03 | Cloud platform MapReduce workflow scheduling optimizing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410315763.5A CN104199820A (en) | 2014-07-03 | 2014-07-03 | Cloud platform MapReduce workflow scheduling optimizing method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104199820A true CN104199820A (en) | 2014-12-10 |
Family
ID=52085113
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410315763.5A Pending CN104199820A (en) | 2014-07-03 | 2014-07-03 | Cloud platform MapReduce workflow scheduling optimizing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104199820A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104408176A (en) * | 2014-12-16 | 2015-03-11 | 湖南师范大学 | Data processing method based on cloud computing platform |
CN105117286A (en) * | 2015-09-22 | 2015-12-02 | 北京大学 | Task scheduling and pipelining executing method in MapReduce |
CN105404611A (en) * | 2015-11-09 | 2016-03-16 | 南京大学 | Matrix model based multi-calculation-engine automatic selection method |
CN105550825A (en) * | 2016-01-15 | 2016-05-04 | 中南民族大学 | Flexible factory work scheduling method based on MapReduce parallelization in cloud computing environment |
CN106383746A (en) * | 2016-08-30 | 2017-02-08 | 北京航空航天大学 | Configuration parameter determination method and apparatus of big data processing system |
CN108920206A (en) * | 2018-06-13 | 2018-11-30 | 北京交通大学 | A kind of plug-in unit dispatching method and device |
CN112906971A (en) * | 2021-03-09 | 2021-06-04 | 清华大学 | Method and device for predicting running time in batch processing operation and electronic equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090281818A1 (en) * | 2008-05-07 | 2009-11-12 | International Business Machines Corporation | Quality of service aware scheduling for composite web service workflows |
CN101763529A (en) * | 2010-01-14 | 2010-06-30 | 中山大学 | Rough set attribute reduction method based on genetic algorithm |
CN103226759A (en) * | 2013-04-25 | 2013-07-31 | 中山大学 | Dynamic cloud workflow scheduling method based on genetic algorithm |
-
2014
- 2014-07-03 CN CN201410315763.5A patent/CN104199820A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090281818A1 (en) * | 2008-05-07 | 2009-11-12 | International Business Machines Corporation | Quality of service aware scheduling for composite web service workflows |
CN101763529A (en) * | 2010-01-14 | 2010-06-30 | 中山大学 | Rough set attribute reduction method based on genetic algorithm |
CN103226759A (en) * | 2013-04-25 | 2013-07-31 | 中山大学 | Dynamic cloud workflow scheduling method based on genetic algorithm |
Non-Patent Citations (2)
Title |
---|
易小华等: "面向MapReduce的数据处理流程开发方法", 《计算机科学与探索》 * |
王忠华: "J2EE应用服务器资源优化机制的研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104408176A (en) * | 2014-12-16 | 2015-03-11 | 湖南师范大学 | Data processing method based on cloud computing platform |
CN105117286A (en) * | 2015-09-22 | 2015-12-02 | 北京大学 | Task scheduling and pipelining executing method in MapReduce |
CN105117286B (en) * | 2015-09-22 | 2018-06-12 | 北京大学 | The dispatching method of task and streamlined perform method in MapReduce |
CN105404611A (en) * | 2015-11-09 | 2016-03-16 | 南京大学 | Matrix model based multi-calculation-engine automatic selection method |
CN105404611B (en) * | 2015-11-09 | 2018-07-20 | 南京大学 | A kind of automatic selecting method of more computing engines based on matrix model |
CN105550825A (en) * | 2016-01-15 | 2016-05-04 | 中南民族大学 | Flexible factory work scheduling method based on MapReduce parallelization in cloud computing environment |
CN105550825B (en) * | 2016-01-15 | 2021-07-02 | 中南民族大学 | Flexible factory job scheduling method based on MapReduce parallelization in cloud computing environment |
CN106383746A (en) * | 2016-08-30 | 2017-02-08 | 北京航空航天大学 | Configuration parameter determination method and apparatus of big data processing system |
CN108920206A (en) * | 2018-06-13 | 2018-11-30 | 北京交通大学 | A kind of plug-in unit dispatching method and device |
CN108920206B (en) * | 2018-06-13 | 2022-01-18 | 北京交通大学 | Plug-in scheduling method and device |
CN112906971A (en) * | 2021-03-09 | 2021-06-04 | 清华大学 | Method and device for predicting running time in batch processing operation and electronic equipment |
CN112906971B (en) * | 2021-03-09 | 2022-02-18 | 清华大学 | Method and device for predicting running time in batch processing operation and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104199820A (en) | Cloud platform MapReduce workflow scheduling optimizing method | |
Wang et al. | A heuristic genetic algorithm for subcontractor selection in a global manufacturing environment | |
CN101916404A (en) | Multi-factory cooperative scheduling optimization method during equipment manufacturing | |
CN102254246A (en) | Workflow managing method and system | |
CN111461890A (en) | Distributed power generation transaction system and method based on multiple prediction and block chain technologies | |
Bremer et al. | Sampling the Search Space of Energy Resources for Self-organized, Agent-based Planning of Active Power Provision. | |
Shibin et al. | Energy sustainability in operations: an optimization study | |
CN104077634A (en) | Active-reactive type dynamic project scheduling method based on multi-objective optimization | |
Manupati et al. | Optimal process plan selection in networked based manufacturing using game-theoretic approach | |
Nguyen et al. | Sharing profit from joint offering of a group of wind power producers in day ahead markets | |
CN112837126A (en) | Industrial park distributed energy trading method and system based on block chain | |
CN109710372A (en) | A kind of computation-intensive cloud workflow schedule method based on cat owl searching algorithm | |
Colonetti et al. | A novel cooperative multi-search benders decomposition for solving the hydrothermal unit-commitment problem | |
Daifen | Evaluate the sustainable marketing strategy to optimal online leasing of new energy vehicles under the background big data economy | |
Djedović et al. | Optimization of business processes by automatic reallocation of resources using the genetic algorithm | |
Ren et al. | A flexible resource investment problem based on project splitting for aircraft moving assembly line | |
CN109067840A (en) | Method, system and the storage medium of artificial intelligence online service | |
Chan et al. | Combining genetic approach and integer programming to solve multi-facility economic lot-scheduling problem | |
CN105160403A (en) | Resource service sequence verification method of cloud manufacturing service | |
Zaman et al. | Resource constrained project scheduling with dynamic disruption recovery | |
Kucukkoc et al. | A new hybrid genetic algorithm to solve more realistic mixed-model assembly line balancing problem | |
Kumar et al. | Stochastic make-to-stock inventory deployment problem: an endosymbiotic psychoclonal algorithm based approach | |
CN113377656B (en) | Public testing recommendation method based on graph neural network | |
CN115545329A (en) | Integrated dynamic scheduling method suitable for clothing manufacturing production line | |
CN114417577A (en) | Cross-platform resource scheduling and optimization control method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20141210 |