CN110119268A - Workflow optimization method based on artificial intelligence - Google Patents
Workflow optimization method based on artificial intelligence Download PDFInfo
- Publication number
- CN110119268A CN110119268A CN201910424800.9A CN201910424800A CN110119268A CN 110119268 A CN110119268 A CN 110119268A CN 201910424800 A CN201910424800 A CN 201910424800A CN 110119268 A CN110119268 A CN 110119268A
- Authority
- CN
- China
- Prior art keywords
- service
- web service
- workflow
- manufacturing process
- service request
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/20—Software design
- G06F8/24—Object-oriented
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/103—Workflow collaboration or project management
-
- 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
Abstract
The Workflow optimization method based on artificial intelligence that the present invention provides a kind of, this method comprises: receiving user's Web service request;User service request is converted into the description of semantic class manufacturing process;The Web service instance index for meeting Web service request is searched by indexing, using the specific Web service instance construction work stream assembled scheme found, and executes workflow.The Workflow optimization method based on artificial intelligence that the invention proposes a kind of, it is easy to determine globally optimal solution, save calculation amount, help to provide richer workflow service assembled scheme simultaneously, and can be trained from limited knowledge data and obtain manufacturing process model, so that new data are predicted and be estimated, obtain more preferably workflow service result, with good generalization ability, it is suitable for the changeable Web service combination environment of dynamic.
Description
Technical field
The present invention relates to workflow, in particular to a kind of Workflow optimization method based on artificial intelligence.
Background technique
Multiple services according to certain rules, are found and are assembled into a value-added integrity service, be by Web service combination
Industry constructs the operation flow in Networked Manufacturing, realizes the effective way of manufacturing recourses distribution according to need.By to enterprise
Industry application system carries out Web service encapsulation, establishes service-oriented frame system, the application system between enterprise is taken with Web
The mode of business integrates, realize Services Composition across enterprise with cooperate, and Business Stream is realized by cross-enterprise workflow system
The automation of journey.But form the Service Instance of composite service in the implementation procedure of composite service may dynamic change, this makes
It obtains Service Instance to be difficult to decide in design phase or compilation phase, and its QoS attribute may also change at any time.It is based on
This, the prior art proposes the workflow service combined method based on MDP.But these conventional methods lack generalization ability, study
Result it is also not accurate enough, it is affected by noise larger.And it needs threshold status transition probability and returns the environment mould of value function
Type.And this is usually not achievable in the actual environment.
Summary of the invention
To solve the problems of above-mentioned prior art, the invention proposes a kind of workflow based on artificial intelligence is excellent
Change method, comprising:
Receive user's Web service request;
User service request is converted into the description of semantic class manufacturing process;
The Web service instance index for meeting Web service request is searched by indexing,
Using the specific Web service instance construction work stream assembled scheme found, and execute workflow.
Preferably, it is described by user service request be converted to semantic class manufacturing process description after, further includes:
Abstract workflow assembled scheme is generated based on intelligent algorithm.
Preferably, the specific Web service instance construction work stream assembled scheme that the utilization is found further comprises:
The abstract workflow assembled scheme is replaced using the specific Web service instance found.
Preferably, described that user service request is converted into the description of semantic class manufacturing process, further comprise:
Service agent receives the service request description of user, and describing to search in library in operation flow can satisfy service and ask
The manufacturing process description asked carries out service binding according to QoS and user preference constraint to the service template in operation flow description.
Preferably, the Web service for meeting Web service request, specifically:
It is adapted to the manufacturing process state parameter set of manufacturing service request, and the manufacturing process with manufacturing service request
The Web service to be selected of implementing result parameter sets adaptation.
The present invention compared with prior art, has the advantage that
The Workflow optimization method based on artificial intelligence that the invention proposes a kind of is easy to determine globally optimal solution, save
Calculation amount, while helping to provide richer workflow service assembled scheme, and can train from limited knowledge data
Manufacturing process model is obtained, so that new data are predicted and be estimated, obtains more preferably workflow service as a result, having good
Good generalization ability is suitable for the changeable Web service combination environment of dynamic.
Detailed description of the invention
Fig. 1 is the flow chart of the Workflow optimization method according to an embodiment of the present invention based on artificial intelligence.
Specific embodiment
Retouching in detail to one or more embodiment of the invention is hereafter provided together with the attached drawing of the diagram principle of the invention
It states.The present invention is described in conjunction with such embodiment, but the present invention is not limited to any embodiments.The scope of the present invention is only by right
Claim limits, and the present invention covers many substitutions, modification and equivalent.Illustrate in the following description many details with
Just it provides a thorough understanding of the present invention.These details are provided for exemplary purposes, and without in these details
Some or all details can also realize the present invention according to claims.
An aspect of of the present present invention provides a kind of Workflow optimization method based on artificial intelligence.Fig. 1 is according to the present invention
The Workflow optimization method flow diagram based on artificial intelligence of embodiment.
Workflow optimization system of the invention include service request reception unit, Services Composition generate unit, execution unit,
Five adaptation unit, service storage unit components: service storage unit stores specific manufacturing operations Web service description and reality
Example;Service request reception unit receives user's Web service request, and user service request is converted to the semanteme that machine can identify
Grade manufacturing process description, is then forwarded to Services Composition and generates unit;Services Composition generates in unit combination service storage unit
Web service description, abstract workflow assembled scheme is generated based on artificial intelligence (or machine learning/nerve net) algorithm, and
And submit to execution unit;Service adapter unit is searched suitable Web service instance index in service storage unit and is returned to
Execution unit, execution unit is using acquisition Web service instance replacement abstract Web Service in service adapter unit and executes work
Stream.
When receiving user's Web service request req, Web service s energy to be selected has been searched whether in service storage unit
Meet its input and output requirement, if it is found, user, which can be used, has the function that service completes oneself.Only when manufacturing service is asked
It asks and is adapted to the manufacturing process state parameter set of Web service to be selected and manufacturing process implementing result parameter sets, can just be recognized
It is adaptation for service request and Web service to be selected.For service request req and composite services (s1, s2, s3... sn) adaptation need
Meet two constraint conditions:
①
②
Wherein, rinIndicate the manufacturing process state parameter set of service request req, routIndicate the manufacture of service request req
Process implementing result parameter sets, si inIndicate Web service s to be selectediManufacturing process state parameter set, si outIndicate siSystem
Make process implementing result parameter sets.
In another optional embodiment, the Web service request of user is handled by service agent, and service agent receives
The service request of user describes, and describes to search the manufacturing process description that can satisfy service request in library in operation flow.This is looked into
Looking for process is to request description to describe the input/output argument that each operation flow describes in library with operation flow respectively user
Set is matched according to the semantic similarity of I/O concept.Each of operation flow description if successful match, to obtaining
Service template carries out service binding according to constraints such as QoS and user preferences, obtains executable workflow composite services result;It is no
Then, service agent is in the dynamic construction for locally carrying out operation flow description.If specific business objective needed for realizing can be generated
Operation flow description is registered to operation flow and described in library, re-executes above-mentioned semantic similarity by operation flow description
Match.If do not generate realize needed for specific business objective operation flow description, service agent by with global workflow
The cooperation of Agent carries out global operation flow and describes dynamic construction;If successfully constructing the industry of specific business objective needed for realizing
Operation flow description is then registered to operation flow and described in library, it is similar then to re-execute above-mentioned semanteme by process of being engaged in description
The step of degree matching, otherwise the result by combination failure returns to user.
Due to usually there are a large amount of manufacturing services in service storage unit, therefrom finds suitable service and need cost longer
Time.The present invention accelerates search process by the way of establishing index to service parameter.Index structure is realized by Hash table.It is right
Index is established respectively in the manufacturing process state parameter and manufacturing process implementing result parameter of service, uses manufacturing process shape respectively
State parameter reference and manufacturing process implementing result parameter reference carry out adaptation lookup, and after adaptation, generation two is to be selected
The set of Web service.One is that the manufacturing process state parameter that meets found after manufacturing process state parameter adaptation is wanted
The set asked, the other is meeting manufacturing process implementing result ginseng by what is found after manufacturing process implementing result parameter adaptation
The set that number requires.Finally acquire two parameter intersections, it is available while meeting manufacturing process state parameter and manufacturing process
The Web service to be selected of implementing result parameter request.
Wherein, index is established to manufacturing process state parameter, manufacturing process implementing result parameter respectively, further includes using s1,
s2... indicate the Web service to be selected in service storage unit.i1, i2, i3...inIndicate manufacturing process state parameter, o1, o2,
o3...onIndicate manufacturing process implementing result parameter.For service request req, manufacturing process state parameter collection is combined into rin=
{r1 in, r2 in, r3 in..., manufacturing process implementing result parameter sets are rout={ r1 out, r2 out, r3 out....It is indexed establishing
It needs to divide the subset of parameter sets before, obtains the key assignments of multiple indexes.
Wherein ri p_inIndicate i-th of dividing subset of manufacturing process state parameter, ri p_outIndicate manufacturing process implementing result
I-th of dividing subset of parameter, N=2n- 1, n are the quantity of manufacturing process state parameter.In service searching process, using tree
Ergodic algorithm, when establishing node to next layer every time, all concentrate one new service of addition in composite services, and deposited from service
A manufacturing service is deleted in storage unit, the output of to the last whole tree can satisfy rout.Which single manufacture is selected later
Service can export rout, the manufacturing process state of these services, all the way up rollback can be generated by which manufacturing service.Finally look for
To service be exactly the required service of entire operation flow.
Step 1: initialization one manufacturing process implementing result set Q and manufacturing process Web service set S, and S is put into
The tail portion Q.A composite services s is chosen from team's head of Q, the manufacturing process state parameter and manufacturing process of the service in s are executed
Result parameter and state parameter set rinIt is put into set R;
Step 2: taking the subset R of R1, R2, R3...Rn.Wherein R1For the complete or collected works of R, RnParameter is 1, RiMiddle parameter is with i value
Increase and reduces (1≤i≤n);Successively choose Ri, manufacturing process shape is found from manufacturing process Web service set by indexing
State parameter and RiGather identical Web service to be selected, forms set Rsad;
Step 3: successively taking RsadIn manufacturing service sad, use sadManufacturing process state parameter and RiParameter carry out language
Justice matching;If sadSuccessful match and sadThe parameter produced different from R then enters step 4;
Step 4: by sadA Combination nova services set S is saved into all elements of SnewAnd delete sad;If SnewMiddle clothes
Manufacturing process implementing result parameter of being engaged in and rinThe parameter set R of compositionnewIt is equal to or comprising rout, then 5 are entered step, otherwise will
SnewIt is put into the tail of the queue of Q, enters step 3;
Step 5: one empty parameter sets R of initializationpara, mark SnewMiddle service manufacturing process implementing result parameter and
With routThere is the manufacturing service of intersection and these service manufacturing process state parameters are put into Rpara;
Step 6: label SnewMiddle manufacturing process implementing result parameter and RparaThere is the service of intersection, empties RparaAnd it will newly mark
The service manufacturing process state parameter of note is put into Rpara;Repeating label step is until SnewMiddle manufacturing process implementing result parameter with
RparaThere is no intersection, exports SnewIn be labeled service;
Step 7: step 2-6 is repeated, until set Rsad, element is sky in R and Q.
The above method by choose service request state parameter set subset, find be able to use division after parameter simultaneously
And service different from R parameter is produced, to expand the type of composite services workflow execution result parameter, until that can expire
Sufficient routDemand, extra service is excluded using the method for rollback from bottom to top.
The above method in order to better understand is illustrated by following example.Assuming that the manufacturing process shape of service request
State parameter sets rin={ a, b, c }, manufacturing process implementing result parameter rout={ x, y, z }.It is first when being adapted to by algorithm
It first needs to divide parameter sets.Element in parameter sets R is rin.R is divided, the subset of R include a, b,
C }, { a, b }, { b, c }, { a, c }, { a }, { b }, { c }.When search index table, according to the element number in parameter set from more to few
Selection subset is inquired.Therefore, { a, b, c } is most begun to use to be inquired.If not finding Web to be selected by concordance list
Service, then select next subset to be inquired, such as { a, b }.If having found a service and the service producing new ginseng
The service, then be added in services set by number.For example, finding Web service S to be selected by { a, b }1。S1Manufacturing process execute knot
Fruit parameter sets S1 out={ d, e }, element is not in parameter sets R={ a, b, c }, at this time by S1It is added in result set,
R={ a, b, c, d, e } simultaneously.Since there is no include service request manufacturing process implementing result parameter sets r by parameter sets Rout
Whole parameters then need to continue to inquire and re-start division to R.R={ a, b, c, d, e } at this time.The subset that R is divided
For { a, b, c, d, e }, { a, b, c, d }, { a, b, c, e } ..., when R contains rinIn whole elements when, forward lookup terminates,
At this point, result set is { S1, S2, S3, S4, S5}.Certain services are likely to become extra service in these services, and extra service is needed
Bottom-up rollback is carried out to remove.When carrying out rollback to result set, label is held comprising service request manufacturing process first
The Web service to be selected of element in row result parameter set.But Web service manufacturing process state parameter to be selected all not is from
Service request manufacturing process state parameter, some also need other service manufacturing process implementing result parameter sets as manufacture stream
Journey state parameter.Therefore continue rollback, mark can generate S2、S4、S5The Web service S to be selected of manufacturing process state parameter1.This
When, S1、S2、S4、S5Manufacturing process state parameter all be from the state parameter set of service request, and clothes can be generated
Whole manufacturing process implementing result parameters of business request.Without the Web service S to be selected of label3Then can by as it is extra to
Web service is selected to be deleted.
After completing being adapted to of service request and Web service to be selected, it is preferable that the Services Composition generation unit is into one
Step is directed to the subset of Web service to be selected, most preferably turns to target with QoS attribute, and manufacturing service is carried out objective optimization combination.If
There is I manufacture subservice in workflow, each subservice is corresponded to J services to be selected, expressed using following objective function:
In formula: T, P, A respectively indicate time, cost and the performance of Web service.kijIt is decision variable, value 0 or 1, kij
J-th of manufacturing service s to be selected of i-th of subservice is indicated when being 1ijIt is selected, kijThe jth of i-th of subservice is indicated when being 0
A manufacturing service s to be selectedijIt is not selected.
The QoS attribute of Web service combination is taken by the composite structure of subservice in manufacturing operations process and the manufacture of selection
The QoS index of business determines.In operation flow, the logical relation of subservice determines that the composite structure of Web service, logical relation include
Sequentially, parallel and selection.When d subservice is that sequence or parallel organization, each subservice are bound respectively from c-th of subservice
One specific service is constrained with the workflow of following formula order of representation and parallel organization;
When f subservice is selection structure from e-th of subservice, only only one business is allowed to bind specific clothes
Business, is indicated with following formula.
c+d≤I;E+f≤I:
In practical manufacturing operations, business demand usually has particular demands to workflow global service, i.e., takes about combination
The global restriction of business, it is also possible to have particular demands to some specific business, i.e., about the local restriction individually serviced, therefore simultaneously
There are local restrictions and global restriction:
In formula: cons (Ti)、cons(Pi)、cons(Ai) respectively indicate to service request reqiLocal time constraint, generation
Valence constraint and performance constraints, cons (T), cons (P), cons (A) are respectively indicated to length of a game's constraint of operation flow, generation
Valence constraint and performance constraints.Global restriction and local restriction are used as to the constraint condition of optimization problem simultaneously, can effectively be kept away
Exempt from composite services weight Plan Problem, improves convergence.
There are incidence relations between multiple Web services provided by manufacturing industry service provider, it is therefore desirable to establish matrix
riji’j’To indicate different service sijWith si’j’Between incidence relation, i.e. two manufacturing services with incidence relation need same
When selected:
The optimization of above-mentioned Web service combination is the selection of manufacturing service example in each services set to be selected, by each clothes to be selected
One group of vector of the Service Instance composition that business collection is selected is one group of solution;The present invention is based on incremental learning algorithms, use
Service Instance is selected for the probability distribution of Web service combination Optimized model, steps are as follows:
Step 1: randomly selecting 1 example respectively by being uniformly distributed from the Service Instance of each services set to be selected, repeat
N times are chosen, N number of initial solid is obtained, N number of initial solid are constituted into initial cluster, note iteration coefficient g is 1.
Step 2: by selected Service Instance vector, substitute into fitness function:
MaxF (s)=ω1·minT+ω2·minP+ω3·maxA
Calculate the fitness of each point in g wheel cluster.
Step 3: choosing the highest h Service Instance vector compositional optimization cluster of fitness in g wheel cluster, wherein h
≤N。
Step 4: the discrete probability distribution function of Estimation Optimization cluster.The probability of i.e. each Service Instance only by where it to
The number of Service Instance in services set is selected to determine.Using the optimization cluster selected in step 3 as sample, to specific probability distribution P
Estimated:
Assuming that the number that the service of serial number s occurs in h optimization entity isThenAs correspond to the clothes of ID
Business example probabilistic.Then probability-distribution function is smoothed.
Assuming that it is respectively maxQoS that all service instance response times are maximum in each services set to be selectedrt i, i=1,
2 ..., n then estimate the value of the independent variable of each Service Instance in each services set to be selected are as follows:
Augments=(ω1·minT+ω2·minP+ω3·maxA)/maxQoSrt i。
To the Service Instance of serial number 0, probability P0=0, it is smoothed after services set i to be selected (i=1,2 ..., n), respectively
The probability of example are as follows:
Probability-distribution function is updated according to following rule:
P " (s | i)=λ P ' (s | i)+(1- λ) Ps, wherein λ indicates study ratio.
Step 5: being sampled according to the probability distribution of estimation, generate novel entities.It is i.e. real in the service of each services set to be selected
In the matrix of example composition, h Service Instance is generated using roulette algorithm by probability distribution P ', it is real as the part in next round
Body.
Step 6: the h entity that will be generated in the highest h entity of fitness in g wheel cluster and step 5, composition
New round cluster, and iteration coefficient is increased 1.
Step 7: if g reaches set the number of iterations, algorithm terminates, and g takes turns the optimal entity in cluster as optimization
Result;Otherwise algorithm goes to step 2 and continues to execute.
By above-mentioned multi-objective Algorithm, the Web service combination problem in workflow is optimized, the complexity of calculating is reduced
And improve the accuracy of optimum results.
For the probability distribution P of each manufacturing service ss, also can record the availability q of corresponding web servicesav(Ps), generation
Valence qpr(Ps), reliability qre(Ps) and time delay qde(Ps) multidimensional property, as the index for measuring operation flow performance superiority and inferiority, thus
To construct the QoS parameter evaluations matrix M=[q of operation flowI, j] (1≤i≤4,1≤j≤r), expression formula are as follows:
Wherein, every a line in matrix indicates the same QoS attribute of different business process, and each column in matrix are corresponding
Available service process set PrIn a service execution process.
The service execution process for being unsatisfactory for the limitation of user's condition is filtered out, such as delay is no more than user demand duration;
Then second level screening is carried out, was carried out by comparing the relative distance of service providing node in workflow and service requesting node
Filter finally determines that composite services execute the performance superiority and inferiority of process by score size, optimal operation flow is requested as user
Response returns to user.
Wherein, Workflow optimization system of the invention is supervised preferably during manufacturing operations Web service combination executes
It controls and analyzes its change information for running context, change information based on context is adjusted Web service combination.Accordingly
Ground, it further comprises three modules, respectively combiner that the Services Composition, which generates unit, context-aware device and amendment
Device.
Combiner completes the parsing that file is described to Web service combination, obtains corresponding abstract service composite sequence, finds
The Web service collection for meeting functional requirement carries out the combination and selection of Web service, is sent in the execution unit and executes.Context
Perceptron monitors the variation of corresponding Service context information and user context information, analyzes the variation pair of contextual information
The influence of Web service combination implementation procedure and trigger workflow adjustment event.Web service combination when corrector is for operation
Different situations be adjusted, to adapt to the dynamic change of workflow context.
After receiving the Web service collection from registration center, these Web service collection are sequentially written in alternative library first
In.Context monitor reads previous contextual information from contextual information bank, it is collected into current monitor upper
Context information compares, will be in new contextual information write-in contextual information bank if contextual information changes.
Judge whether the change information still meets user demand and related constraint.If the variation of contextual information is to Web service group
Closing to execute does not influence, then the Web service being currently executing still can satisfy user demand and related constraint.Otherwise, to
Execution unit issues the message that pause executes, and the corresponding strategy that analysis result and user provide is passed to corrector, transports
Manufacturing service combination and selection are re-started with above-mentioned web service composition method, obtains the Web service of Services Composition optimal quality
Composite sequence.
For the service quality of manufacturing service provider, may dynamically become at any time for QoS attribute in combination of network service
The characteristics of change, the present invention are further analyzed and are described to the QoS attribute of Web service, are modeled as comprising reliability, the degree of coupling
Attribute set.
The reliability is guaranteed reliability's ability that identified single Service Instance itself is capable of providing, and is denoted as σi,
It combines to obtain influence of the single-point to entire workflow in the dimension of operation flow are as follows:
N is the quantity of service in composite services.
When the degree of coupling is defined on the Service Instance for participating in operation flow building, based on determined by Components Development information
Service Instance i and j degree of compatibility, is denoted as μij∈ [0,1], the combination in reliable service process are as follows:
Wherein: w >=1 is control parameter.
Next, establish utility function to measure the reliability level of reliable service process, the effectiveness based on each attribute of QoS
Function is as follows:
FP (s1, s2 ..., sn)=α σP (s1, s2 ..., sn)+βμP(s1, s2 ..., sn)
Wherein: alpha+beta=1, α, β are respectively the weighted value of reliability, the degree of coupling.
In view of the high dynamic of cloud manufacturing service resource, after following procedure is selected based on service ability and a route to be valid phase
After the workflow generation strategy of Service Instance, extracts service node information and service routing current state carrys out Shared Decision Making service in fact
Example instancing processes, select that a route to be valid phase is longer and the biggish Service Instance of service ability completes corresponding function, are increasing
The success rate of manufacturing service combination is improved while strong operation flow stability to the maximum extent.
A route to be valid phase RE minimum value of a route to be valid phase as constructed by adjacent node determines, i.e. RE=min { LET
(θI, j)};Wherein: θI, jIndicate adjacent two nodes i, the routing between j;LET(θI, j) it is routing θI, jPrediction validity period.When one
When a node is removed out of another node communication range, routing occurs and interrupts.
Service ability is to user node NiObtain service node NjOn Service Instance sjStability expression, be denoted as:
RbNi(sj)=α × Qj/rly(Ni, Nj)
Wherein: QjIt is service node NjService s is providedjAbility;rly(Ni, Nj) indicate from user node NiTo service Nj
Handover number;α is a constant.
Service Instance selection algorithm specific steps are carried out such as based on routing prediction validity period and service ability optimal policy
Under:
A) assume to combine based on user service in the workflow topological diagram for the Services Composition scheme that request generates containing m clothes
Pragmatic example: S={ S1, S2..., Sm, it is predefined to generate services set to be selected belonging to each Service Instance with the same function: SA
(Sj)={ s1 j, s2 j..., smj j, wherein 1≤j≤m, mj> 0, smj jFor specific Service Instance.Its node indicates finishing service
Services set to be selected.
B) in operation flow topological diagram, present node V is judged based on following stepsiWith predecessor node VjQoS whether protect
It holds consistent:
(1) assume present node ViWith predecessor node VjAffiliated services set to be selected is respectively SA (Si)、SA(Sj);
(2) to SA (Sj) in each service sn j(1≤n≤mj), with SA (Si) in all service sn i(1≤n≤mi) two
Two compare, if sn jService quality F (the s of outputn j) meet sn iThe business demand of submission, by sn jLabeled as effective service to be selected.
C) node serve information and network routing state are extracted, calculates the Service Instance for realizing next service in present node
ViThe service ability RbN at placei(sj) and Service Instance routing predict validity period ti,j, sort after weighted average.Selection result
Optimal and suboptimum Service Instance generates final service process as services selection return value, constitutes subgraph G ' (V ', E '), figure
Vertex V ' be Service Instance s after being instantiated in corresponding services set to be selectedn j, two Service Instance V of side E ' expressioniIt is saved with forerunner
Point VjBetween exist input, output relation, while solve obtain manufacturing service ability and routing Lifetime weighted average it is optimal
Value.
In conclusion the invention proposes a kind of Workflow optimization method based on artificial intelligence, be easy to determine it is global most
Excellent solution saves calculation amount, while helping to provide richer workflow service assembled scheme, and can be from limited knowledge number
Manufacturing process model is obtained according to middle training, so that new data are predicted and be estimated, obtains more preferably workflow service knot
Fruit has good generalization ability, is suitable for the changeable Web service combination environment of dynamic.
Obviously, it should be appreciated by those skilled in the art, each module of the above invention or each steps can be with general
Computing system realize that they can be concentrated in single computing system, or be distributed in multiple computing systems and formed
Network on, optionally, they can be realized with the program code that computing system can be performed, it is thus possible to they are stored
It is executed within the storage system by computing system.In this way, the present invention is not limited to any specific hardware and softwares to combine.
It should be understood that above-mentioned specific embodiment of the invention is used only for exemplary illustration or explains of the invention
Principle, but not to limit the present invention.Therefore, that is done without departing from the spirit and scope of the present invention is any
Modification, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.In addition, appended claims purport of the present invention
Covering the whole variations fallen into attached claim scope and boundary or this range and the equivalent form on boundary and is repairing
Change example.
Claims (5)
1. a kind of Workflow optimization method based on artificial intelligence characterized by comprising
Receive user's Web service request;
User service request is converted into the description of semantic class manufacturing process;
The Web service instance index for meeting Web service request is searched by indexing,
Using the specific Web service instance construction work stream assembled scheme found, and execute workflow.
2. the method according to claim 1, wherein described be converted to semantic class manufacture stream for user service request
After journey description, further includes:
Abstract workflow assembled scheme is generated based on intelligent algorithm.
3. according to the method described in claim 2, it is characterized in that, the specific Web service instance building that the utilization is found
Workflow assembled scheme further comprises:
The abstract workflow assembled scheme is replaced using the specific Web service instance found.
4. the method according to claim 1, wherein described be converted to semantic class manufacture stream for user service request
Journey description further comprises:
Service agent receives the service request description of user, and describing lookup in library in operation flow can satisfy service request
Manufacturing process description carries out service binding according to QoS and user preference constraint to the service template in operation flow description.
5. the method according to claim 1, wherein it is described meet Web service request Web service, specifically:
It is adapted to the manufacturing process state parameter set of manufacturing service request, and is executed with the manufacturing process of manufacturing service request
The Web service to be selected of result parameter set adaptation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910424800.9A CN110119268B (en) | 2019-05-21 | 2019-05-21 | Workflow optimization method based on artificial intelligence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910424800.9A CN110119268B (en) | 2019-05-21 | 2019-05-21 | Workflow optimization method based on artificial intelligence |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110119268A true CN110119268A (en) | 2019-08-13 |
CN110119268B CN110119268B (en) | 2023-05-02 |
Family
ID=67522992
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910424800.9A Active CN110119268B (en) | 2019-05-21 | 2019-05-21 | Workflow optimization method based on artificial intelligence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110119268B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110457490A (en) * | 2019-08-15 | 2019-11-15 | 桂林电子科技大学 | A kind of semantic work stream index construction and search method based on domain body |
CN111629053A (en) * | 2020-05-27 | 2020-09-04 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Credible geographic information service self-adaptive combination method and system |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101655943A (en) * | 2009-09-14 | 2010-02-24 | 中兴通讯股份有限公司 | Management method and system of enterprise application integrated working flow |
CN101820428A (en) * | 2010-04-22 | 2010-09-01 | 北京航空航天大学 | Composite service optimizing method and device based on protocol composition mechanism |
CN102780580A (en) * | 2012-06-21 | 2012-11-14 | 东南大学 | Trust-based composite service optimization method |
CN103336763A (en) * | 2013-06-05 | 2013-10-02 | 华南理工大学 | Complex similarity measurement method of semantic Web service composition results |
CN105046351A (en) * | 2015-07-01 | 2015-11-11 | 内蒙古大学 | Reinforcement learning-based service combination method and system in uncertain environment |
CN105184403A (en) * | 2015-09-01 | 2015-12-23 | 华东师范大学 | Workflow optimal allocation optimizing method based on machine learning and statistical model checking |
CN105488166A (en) * | 2015-11-30 | 2016-04-13 | 北京金山安全软件有限公司 | Index establishing method and device |
US20180218276A1 (en) * | 2017-01-30 | 2018-08-02 | Bank Of America Corporation | Optimizing Application Performance Using Finite State Machine Model and Machine Learning |
-
2019
- 2019-05-21 CN CN201910424800.9A patent/CN110119268B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101655943A (en) * | 2009-09-14 | 2010-02-24 | 中兴通讯股份有限公司 | Management method and system of enterprise application integrated working flow |
CN101820428A (en) * | 2010-04-22 | 2010-09-01 | 北京航空航天大学 | Composite service optimizing method and device based on protocol composition mechanism |
CN102780580A (en) * | 2012-06-21 | 2012-11-14 | 东南大学 | Trust-based composite service optimization method |
CN103336763A (en) * | 2013-06-05 | 2013-10-02 | 华南理工大学 | Complex similarity measurement method of semantic Web service composition results |
CN105046351A (en) * | 2015-07-01 | 2015-11-11 | 内蒙古大学 | Reinforcement learning-based service combination method and system in uncertain environment |
CN105184403A (en) * | 2015-09-01 | 2015-12-23 | 华东师范大学 | Workflow optimal allocation optimizing method based on machine learning and statistical model checking |
CN105488166A (en) * | 2015-11-30 | 2016-04-13 | 北京金山安全软件有限公司 | Index establishing method and device |
US20180218276A1 (en) * | 2017-01-30 | 2018-08-02 | Bank Of America Corporation | Optimizing Application Performance Using Finite State Machine Model and Machine Learning |
Non-Patent Citations (2)
Title |
---|
NARAYAN DEBNATH 等: ""A method to evaluate QoS of web services required by a workflow"" * |
黄秋波 等: ""业务流程管理中基于规格属性及索引机制的服务匹配算法"" * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110457490A (en) * | 2019-08-15 | 2019-11-15 | 桂林电子科技大学 | A kind of semantic work stream index construction and search method based on domain body |
CN110457490B (en) * | 2019-08-15 | 2021-06-18 | 桂林电子科技大学 | Semantic workflow index construction and retrieval method based on domain ontology |
CN111629053A (en) * | 2020-05-27 | 2020-09-04 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Credible geographic information service self-adaptive combination method and system |
CN111629053B (en) * | 2020-05-27 | 2023-10-20 | 深圳市规划国土房产信息中心(深圳市空间地理信息中心) | Trusted geographic information service self-adaptive combination method and system |
Also Published As
Publication number | Publication date |
---|---|
CN110119268B (en) | 2023-05-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112016812B (en) | Multi-unmanned aerial vehicle task scheduling method, system and storage medium | |
CN107404523A (en) | Cloud platform adaptive resource dispatches system and method | |
CN110119399A (en) | Work Flow Optimizing method based on machine learning | |
CN106055395A (en) | Method for constraining workflow scheduling in cloud environment based on ant colony optimization algorithm through deadline | |
CN109818786A (en) | A kind of cloud data center applies the more optimal choosing methods in combination of resources path of appreciable distribution | |
CN103345514A (en) | Streamed data processing method in big data environment | |
CN102158417A (en) | Method and device for optimizing multi-constraint quality of service (QoS) routing selection | |
CN110297699A (en) | Dispatching method, scheduler, storage medium and system | |
CN113037877B (en) | Optimization method for time-space data and resource scheduling under cloud edge architecture | |
CN109491761A (en) | Cloud computing multiple target method for scheduling task based on EDA-GA hybrid algorithm | |
Coles et al. | Cost-sensitive concurrent planning under duration uncertainty for service-level agreements | |
CN113742089B (en) | Method, device and equipment for distributing neural network computing tasks in heterogeneous resources | |
CN105205052B (en) | A kind of data digging method and device | |
CN104834751A (en) | Data analysis method based on Internet of things | |
CN111966495B (en) | Data processing method and device | |
CN112685153A (en) | Micro-service scheduling method and device and electronic equipment | |
CN111752678A (en) | Low-power-consumption container placement method for distributed collaborative learning in edge computing | |
CN106371924A (en) | Task scheduling method for maximizing MapReduce cluster energy consumption | |
CN110119268A (en) | Workflow optimization method based on artificial intelligence | |
CN113010296B (en) | Formalized model based task analysis and resource allocation method and system | |
CN114461368A (en) | Multi-target cloud workflow scheduling method based on cooperative fruit fly algorithm | |
Lin et al. | A bottom-up tree based storage approach for efficient iot data analytics in cloud systems | |
Li et al. | Efficient adaptive matching for real-time city express delivery | |
Mostafa et al. | An intelligent dynamic replica selection model within grid systems | |
CN110135747A (en) | Process customizing method neural network based |
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 |