CN107105052A - Heuristic web service composition method based on figure planning - Google Patents
Heuristic web service composition method based on figure planning Download PDFInfo
- Publication number
- CN107105052A CN107105052A CN201710343392.5A CN201710343392A CN107105052A CN 107105052 A CN107105052 A CN 107105052A CN 201710343392 A CN201710343392 A CN 201710343392A CN 107105052 A CN107105052 A CN 107105052A
- Authority
- CN
- China
- Prior art keywords
- state
- service
- mrow
- planning
- msub
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/51—Discovery or management thereof, e.g. service location protocol [SLP] or web services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention is to provide a kind of heuristic web service composition method based on figure planning.Services Composition problem is modeled first, its corresponding relation with intelligent planning problem is elaborated;In order to solve the shortcoming of the blind search based on the Service Combination Algorithm for scheming planning, it is proposed that the concept of state distance, its effect in approachability analysis is analyzed and demonstrated, the construction method of state distance matrix is given;According to state distance matrix, design heuristic function is estimated the accessibility of service, trims unnecessary service, reduces the scale of planning chart, improves the solution efficiency of algorithm.
Description
Technical field
The present invention relates to a kind of heuristic web service composition method.
Background technology
Nowadays, Services Composition problem obtains the concern of numerous researchers, wherein, it is directed generally to that group will be serviced in industrial quarters
It is main by the way of process modeling and the engine performed in terms of the method and theory of conjunction are applied to engineering practice, it there is spirit
The defect of activity and autgmentability difference.In academia, researcher focuses on the theoretical research of the automatic combination method of service, in order to
The Automatic Combined of service is realized, generally describes to service by semantic service, so that computer is capable of the semanteme of automatic understanding service
Information, and then the correlation theory of intelligent planning is combined, solve the problems, such as Services Composition.But the correlation theory of automatic combination method is simultaneously
It is immature, it tends to be difficult to be applied in Practical Project.Services Composition problem can obtain different points from different angles
Class result, is divided into static method and dynamic approach according to the difference of combination opportunity;Difference according to the technology of support is classified as figure and searched
Method, the method for business flow control and the method for intelligent planning theory control of rope control;Services Composition problem is divided into reality
When task solving scheme and the scheme of business-driven
The assembled scheme of Flow driving is realized by binding service node in flow definition figure, shows as participant
Taken action according to pre-defined rule and realize function.Because workflow has been widely used, therefore the assembled scheme of Flow driving
There are the theoretical foundation and technology application background of maturation.The combination of Flow driving is using workflow as technical foundation, with business
Template is core, and user is used it can be readily appreciated that improving the manageability during performing.Its shortcoming is that automaticity is poor,
Most of work must be adopted manually, and need domain expert to design service template, it is difficult to which reply is extensive, complicated
Operation flow.
The scheme that real-time task is solved is the assembled scheme based on AI planning theories, using the think of of classical artificial intelligence
Think, Services Composition problem is mapped as planning problem, solved by intelligent planning reasoning, i.e., in known initial state and target
The situation of state, searches an optimal Perform sequence to realize the progressively evolution between the two states from Web service candidate storehouse.
The assembled scheme planned based on AI, i.e., portray combinatorial problem, and automatic calculation according to the associated description method of intelligent planning.Therefore
Services Composition problem can be with abstract for intelligent planning problem, and numerous achievements in research in intelligent planning field are applied to service
Combination.
Figure planning algorithm is the process of an iteration, by iteration each time, and planning chart can extend new service and obtain
To new proposition.After the extension of continuous iteration, if proposition layer contains target complete proposition or planning chart is entered not
Dynamic layer then terminates expansion process.If planning chart enters motionless layer, illustrate the service solution for not meeting demand.Otherwise, planning chart
Into the solution extraction stage, the stage planning chart need to look for each dbjective state one group of service to meet the state, and form new mesh
Mark state, untill all dbjective states are both present in original state, then obtains the planning solution of the problem.
If so planning chart of the planning problem without solution will not return to solution;Otherwise, it will return to a solution
The action sequence set of problem scheme.In addition, each planning chart has a number of plies k, and it is complete to only need to polynomial time
The extension of paired planning chart.And scheming the algorithm of planning has reliability, integrality and terminability.
In the Service Combination Algorithm based on figure planning:
(1) service ws can be mapped as the action a in planning problem;
(2) the input parameter I of service can be mapped as precondition preconds (a);
(3) the action effect effects (a) that the output parameter O of service can be mapped as;
(4) r in user's requestinThe original state s that can be mapped as0;
(5) r in user's requestoutThe dbjective state g that can be mapped as;
(6) G=<SL1,AL1,SL2,AL2,...,SLi-1,ALi-1,SLi>:The planning being made up of action layer and state layer
Figure.
In classic map planning problem, there is good effect and point of negative effect in action, however, in Services Composition problem
In, Web service only exists good effect, therefore proposes simple planning chart in order to apply in Services Composition problem, in letter
It is separate in single planning chart, between action, in the absence of the mutex relation of action and action, proposition and proposition.For ease of
Description, the present invention will continue simple planning chart being referred to as planning chart.Service Combination Algorithm based on figure planning is divided into extension phase
Conciliate the extraction stage.
Extension phase:
(1) the primitive proposition set provided according to user builds first layer state layer
(2) the service ws for the condition that meets is extended according to service library and forms planning chart action layer ALi
(3) by the new state layer SL of effects (ws) formationi+1
(4) repeat step (2) and (3), until dbjective state is all appeared in state layer, or reach motionless layer.
(5) continuous iteration Expansion Planning figure, until then being obtained in planning chart comprising target complete state by solving extraction algorithm
To a Services Composition scheme, if planning chart reaches motionless layer, illustrate to meet user's request in the absence of one group of service.
Solve the extraction stage:
(1) to each dbjective state g, one group of service WS is found, its enable exports and obtain state g
(2) effects (WS) is constituted to new dbjective state set;
(3) above procedure is repeated, until reaching initial shape state layer, then one is obtained and reaches dbjective state from original state
Group
Conjunction solution ResultWS (<WS1,WS2,...,WSi>)。
Fig. 2 show service relation illustrated example, and wherein A, B and C is the original state of the combinatorial problem, and H and J are target-like
State, W1、W2、W3、W4And W5For existing service in service library.First proposition layer is constituted by original state A, B and C, in service
Searched in storehouse and choose all services that can be performed and constitute first element layer, the service W in layer will be acted1Output D and E
It is added in second layer proposition layer, forms new proposition layer, the like, until dbjective state G and H are integrally incorporated in proposition layer
In, now enter the solution extraction stage.In the stage, for sub-goal H and I, in previous action layer selecting one group can reach
Target H and I service W4, and the input D and F of the service are constituted into new dbjective state, the like, until reaching initial shape
State, then obtain one group of solution W1、W2、W4Enable and dbjective state H and J are reached by original state A, B and C.
Based on the service combining method of figure planning, in time complexity with having relative to other method on space complexity
Some superiority, but still there is deficiency:Breadth First is similar on the expansion process of Service Combination Algorithm based on figure planning
Algorithm, blind search, time complexity is high, easily occurs multiple shot array, it is difficult to applied to magnanimity services set.
The content of the invention
It is an object of the invention to propose a kind of opening based on figure planning that can be reduced planning map space, improve solution efficiency
Hairdo web service composition method.
The object of the present invention is achieved like this:
Step 1:Input the service that user is asked in summed data storehouse;
Step 2:Structure state distance matrix M (i, j);
Step 3:Calculate heuristic function value h (wsi), the accessibility to service is estimated, trims unnecessary service;
Step 4:Judge that planning chart enters fixed point layer, then perform step 5, otherwise perform step 6;
Step 5:Planning is without solution, and algorithm terminates;
Step 6:Judge that planning chart some state layer occurs and includes target complete state, then perform step 7, otherwise return to step
Rapid 3;
Step 7:Perform solution extraction algorithm;
Step 8:Combined result is exported, is terminated.
The present invention can also include:
The state distance matrix and heuristic function value are:
(1) service with triple ws (ID, a Sin,Sout) represent, wherein SinRepresenting one group is used to describe input state
Semantic concept set, SoutRepresent one group of semantic concept set for being used to describe output state;
(2) semantic state is apart from D (si,sj) represent semantic state siWith semantic state sjCorrelation degree:
(3) semantic state distance matrix M, is the matrix for representing the relation between semantic state, is specially:
M (i, j)=D (si,sj);
State distance describes the always too high estimation service solution of the weak accessibility between state and state, i.e. state distance
Accessibility, it is if the state distance between state A and state B is infinitely great, i.e., unreachable, then in the absence of one group of planning solution cause from
State A reaches state B;
(4) if original state is rin=(i1,i2...in), dbjective state rout=(o1,o2...on), service WS=(ws1,
ws2...wsn), N (S) represents the number of state contained by state set S, wherein gs(p) it is from state s to dbjective state p estimation
Value is state distance matrix M (s, p) value, then services wsiInspiration value h (wsi) be:
h(wsi) accessibility of current service and dbjective state set is have estimated using the value of the output state distance of service,
The value describes the average of the state distance between the output state of service and dbjective state, h (wsi) always in state in
Choose minimum value to be estimated, h (wsi) value for it is infinitely great when, service wsiDo not appear in solution.
The present invention analyzes the defect of the Service Combination Algorithm based on figure planning, and it is improved proposes figure planning framework
Under heuristic Web service combination algorithm (SDSC-GP).Services Composition problem is modeled first, itself and intelligence is elaborated
The corresponding relation of planning problem;In order to solve the shortcoming of the blind search based on the Service Combination Algorithm for scheming planning, it is proposed that shape
The concept of state distance, analyzes and demonstrates its effect in approachability analysis, give the construction method of state distance matrix;
According to state distance matrix, design heuristic function is estimated the accessibility of service, trims unnecessary service, reduces planning
The scale of figure, improves the solution efficiency of algorithm.
Due to being similar to breadth first algorithm on the expansion process of the Service Combination Algorithm based on figure planning, there is extension effect
The problem of rate is low, the present invention carries out beta pruning by evaluation services and the accessibility of dbjective state to planning chart.Propose state away from
From concept, to assess the correlation degree and accessibility of current service and dbjective state, and inaccessible service is cut
Branch, so as to reduce planning chart scale, improves the expansion efficiency of planning chart.Rule will not be destroyed by the Pruning strategy by finally demonstrating
Draw the solution of figure.This method possesses lower space consuming and Geng Gao solution efficiency on the basis of not influenceing to solve success rate.
In heuritic approach, in order to avoid the node that blind search is unnecessary, in order to reduce solution space, generally using base
In the heuristic strategies of loose problem.In the strategy, generally the constraints of former problem is reduced, simplifies problem scale,
By the cost under computational short cut problem, the target cost of the former problem of estimation.Its state distance proposed, passes through state distance
Computational algorithm, the State Reachability in former problem is estimated by the State Reachability in loose problem, and then to each action
(service) is estimated, and deletion does not have the service of facilitation to dbjective state accessibility.Under normal circumstances, with loosening degree
Reduction, simplify problem with former problem closer to loose estimation is also more accurate.
With reference to Fig. 4, according to the calculating of state range formula in figure, D (A, G)=2, D (B, H)=2, D (C, K)=2.It is right
For state A, state G state is reached apart from D (A, G)=2, due to service W4Lack input state C, and in the absence of one group
Service causes state A to reach state G, and state A to state G actual distance is 0;And for state B, its state is apart from D
(B, H)=2, while in the presence of one group of service (W2,W3,W5) such that state B reaches state H, i.e. state B reach state H it is true away from
From for 3.For state C, there is one group of service (W in its state apart from D (C, K)=26,W7) cause state C to reach state K,
That is the actual distance that state C reaches state K is 2.It is possible thereby to know, state is reachable between always too high estimated state
Property.
From the building process of state matrix it is seen that, state matrix describe it is stateful between correlation degree and can
Up to property.
The Pruning strategy of the present invention, can avoid the blindness of planning chart from extending, the effective size for reducing planning chart, lifting
Solution efficiency, and inspiration value h (wsi) computation complexity it is very low, if assume wsiOutput parameter maximum number be c, mesh
The number of parameters for marking state set is k, then h (wsi) computation complexity be O (k*c), c and k are one and are less than under normal circumstances
3 number, therefore h (wsi) computational efficiency height.
Brief description of the drawings
Fig. 1 is the service relation illustrated example of the Service Combination Algorithm based on figure planning.
Fig. 2 is the planning illustrated example of the Service Combination Algorithm based on figure planning.
Fig. 3 is the flow chart of the present invention.
Fig. 4 is the service relation figure of the present invention.
Fig. 5 is the service relation figure of the present invention.
Fig. 6 is state matrix example.
Fig. 7 is beta pruning effect exemplary plot.
Fig. 8 is the overview flow chart of the embodiment of the present invention.
Fig. 9 is the state distance matrix block diagram of the present invention.
Figure 10 be three kinds of methods in the case of different pieces of information scale, combine success rate situation of change.
Figure 11 be three kinds of methods in the case of different pieces of information scale, the situation of change of service request handling time.
Figure 12 is planning chart scale comparison diagram.
Embodiment
Illustrate below and the present invention is described in more detail.
With reference to Fig. 8, the heuristic Web service combination algorithm of the invention based on figure planning includes:
Step 1:Input the request (r of userin,rout) and service library in service;
Step 2:Structure state distance matrix M (i, j);
Step 3:Init state layer SL1;
Step 4:The service in traverse service storehouse;
Step 5:Judge that current state layer, whether comprising parameter is fully entered needed for service, performs step 6 if meeting,
Otherwise return to step 4;
Step 6:Inspiration value h (ws are calculated according to formula 2-3i);
Step 7:If h (wsi) be not maximum, i.e., critical value is not arrived, then step 8 is performed, otherwise it is assumed that current service is not
A part for planning solution, return to step 4 can be turned into;
Step 8:The service is added to first layer action layer AL1, the output state of service is added to new state layer
SL2;
Step 9:Judge whether to have traveled through the service in service library, then perform step 10, otherwise return to step 4;
Step 10:Produce complete new state layer SL2;
Step 11:Judge all dbjective state routState layer is appeared in, then performs step 12, step 14 is otherwise performed;
Step 12:Perform solution extraction algorithm;
Step 13:Combined result is exported, algorithm terminates;
Step 14:Judge that planning chart enters fixed point layer, then perform step 15, otherwise return to step 4;
Step 15:Planning is without solution, and algorithm terminates.
With reference to Fig. 9, the present invention builds state distance matrix in the expansion process of planning chart, builds state distance matrix
Method is as follows:
Step 1:Input user's request (rin,rout) and service library in service;
Step 2:Initialize matrix M, M (i, j)=∞;
Step 3:Each service in traverse service storehouse;
Step 4:Service M (i, j) comprising original state and dbjective state is entered as 1;
Step 5:Ergodic Matrices M each value;
Step 6:It is not 0 to judge service M (i, j), M (j, k), then performs step 7, otherwise return to step 5;
Step 7:Carry out calculating M (i, k);
Step 8:Judge k<J, then perform step 9, otherwise performs step 10;
Step 9:Labeled as true, return to step 5;
Step 10:Labeled as vacation;
Step 11:Complete state distance matrix.
It is terminable that planning chart, which is discussed below, and fixed point layer is provided first and is defined as follows:
The motionless layer for defining 1. planning chart G is kth layer, then rightLow i layers of G are constantly equal to kth layer.That is SLi=
SLk,ALi=ALk。
2. each planning chart G are defined in the presence of a minimum fixed point layer k so that SLk-1=SLk。
Defined from two above, due in the expansion process of planning chart, the state number in proposition layer must be
Monotonic increase, i.e. the scale of proposition layer constantly increases, and proposition number is limited in classical planning problem, therefore planning nomography
Certainly exist fixed point layer k.Therefore, if request is without planning solution, the algorithm one surely reaches fixed point layer, and terminates the calculation
Method.
Experimental verification:
To verify the validity of SDSC-GP methods, measurement is used as using combination success rate, planning chart scale, assembly time
Index, contrast test is carried out with PGSGA methods and ACSWS-PG methods.
The software environment and hardware environment of test are as follows:
CPU:Intel(R)Core(TM)2 Quad CPU
Physical memory:1.96GB
Dominant frequency:2.66GHz
Operating system:Windows XP Professional
Database:MySQL
Software tool:Prot é g é have the characteristics such as free download, graphic interface, abundant plug-in unit, therefore this problem is selected
Prot é g é are selected as ontology editor, Semantic Web Services description is carried out using the OWL-S Editor plug-in units under Prot é g é
The editor of OWL-S documents.OWL-S API are a set of Java API for being developed and being safeguarded by University of Maryland, for OWL-S files
Establishment, read-write and perform.This experiment obtains the input of OWL-S services using OWL-S API class libraries parsing OWL-S files
The result that intelligent planning is produced, is converted into the OWL-S files of composite services, performs the destination service by the information such as output afterwards.
Experimental data:OWLS-TC4 data sets are increased income, and can arbitrarily be used by personal and group, the data set by
Benedikt Fries, MahboobAlam Khalid, Martin Vasileski joint developments, include 1083 Semantic Webs
Service, is related to traffic, economy, education, food, geography, medicine, emulation, nine fields of tourism and weapon.
Test program is realized using java language.For the validity of measure algorithm, this problem in OWL S-TC4 with
Machine chooses 5 service libraries of different scales, respectively 200,400,600,800,1000.This problem and PGSCA algorithms and
ACSWS-PG algorithms are contrasted, and evaluation criterion is used as with request processing time using combination success rate.
Experimental result and analysis:
5 groups of service requests are generated for experiment, every group contains 100 service requests.Submitted respectively on 5 service libraries
This 5 groups of service requests, compare PGSCA, ACSWS-PG and SDSC-GP these three methods and are serviced on the service library of different scales
The success rate of combination.
Figure 10 be three kinds of methods in the case of different pieces of information scale, combine success rate situation of change.With data set
The increase of middle quantity of service, the combination success rate of PGSCA, ACSWS-PG and SDSC-GP method rises appreciably, this table
It is bright, with the increase of data set scale, the possibility increase that planning solution is present, this is consistent with actual conditions.ACSWS-
PG and SDSC-GP methods, in the case of service scale identical, it combines success rate and is above PGSCA methods, and this is due to
The matching way based on keyword that PGSCA is used when processing parameter is matched, it is impossible to match the body ginseng of equivalence class each other
Number, causes the reduction of Services Composition success rate.
Figure 11 be three kinds of methods in the case of different pieces of information scale, the situation of change of service request handling time.From figure
In can easily find out, the solution time of PGSCA, ACSWS-PG and SDSC-GP method, be in the increase of quantity of service
The combination request processing time of ascendant trend, wherein ACSWS-PG methods is less than PGSCA methods, and the combination of SDSC-GP methods please
Processing is asked to be significantly lower than ACSWS-PG methods, but with the increase of quantity of service, the combination request processing time of SDSC-GP methods
Start to level off to PGSCA methods.ACSWS-PG methods set up inverted index to the output parameter of service and optimize PGSCA methods,
Therefore the treatment effeciency of ACSWS-PG methods is rational higher than PGSCA methods.The solution time of SDSC-GP methods is less than
ACSWS-PG methods and PGSCA methods illustrate that SDSC-GP carries out beta pruning by setting up state distance matrix to planning chart
Method can effectively point out programming evaluation efficiency.But with the increase of quantity of service, the combination request of SDSC-GP methods is handled
Time starts to level off to PGSCA methods, and this explanation is with the increase of quantity of service, the correlation degree increase between semantic state,
The beta pruning effect of PGSCA methods is caused to die down.
In order to verify that SDSC-GP methods can reduce the scale of planning chart, experiment 100 requests of random generation, and advising
Carry out solution statistical rules figure on the service library that mould is 200,400,600,800,1000 with SDSC-GP and PGSCA methods respectively
The average value of the quantity of service of extension.Experimental configuration is as shown in figure 11.
As shown in Figure 12, the scale of the planning chart built with the increase of service scale, SDSC-GP and PGSCA methods increases
Greatly, the ratio of the difference station PGSCA methods of SDSC-GP and PGSCA methods scale is gradually reduced, the rule that SDSC-GP methods are built
Draw the planning chart scale that figure scale is consistently less than PGSCA methods structure.Two methods with service library increase planning chart scale
Increase, the planning chart scale that the planning chart scale that SDSC-GP methods are built is consistently less than PGSCA methods structure demonstrates SDSC-
The beta pruning effect of GP methods.The ratio of the difference station PGSCA methods of SDSC-GP and PGSCA method scales is gradually reduced, this explanation
With the increase of quantity of service, the correlation degree between state increases, and causes the beta pruning effect of PGSCA methods to die down.
Claims (3)
1. a kind of heuristic web service composition method based on figure planning, it is characterized in that:
Step 1:Input the service that user is asked in summed data storehouse;
Step 2:Structure state distance matrix M (i, j);
Step 3:Calculate heuristic function value h (wsi), the accessibility to service is estimated, trims unnecessary service;
Step 4:Judge that planning chart enters fixed point layer, then perform step 5, otherwise perform step 6;
Step 5:Planning is without solution, and algorithm terminates;
Step 6:Judge that planning chart some state layer occurs and includes target complete state, then perform step 7, otherwise return to step 3;
Step 7:Perform solution extraction algorithm;
Step 8:Combined result is exported, is terminated.
2. it is according to claim 1 based on figure planning heuristic web service composition method, it is characterized in that build state away from
Method from matrix is:
Step 1:Input user's request (rin,rout) and service library in service;
Step 2:Initialize matrix M, M (i, j)=∞;
Step 3:Each service in traverse service storehouse;
Step 4:Service M (i, j) comprising original state and dbjective state is entered as 1;
Step 5:Ergodic Matrices M each value;
Step 6:It is not 0 to judge service M (i, j), M (j, k), then performs step 7, otherwise return to step 5;
Step 7:Carry out calculating M (i, k);
Step 8:Judge k<J, then perform step 9, otherwise performs step 10;
Step 9:Labeled as true, return to step 5;
Step 10:Labeled as vacation;
Step 11:Complete state distance matrix.
3. it is according to claim 2 based on figure planning heuristic web service composition method, it is characterized in that the state away from
It is from matrix and heuristic function value:
(1) service with triple ws (ID, a Sin,Sout) represent, wherein SinRepresent one group of semanteme for being used to describe input state
Concept set, SoutRepresent one group of semantic concept set for being used to describe output state;
(2) semantic state is apart from D (si,sj) represent semantic state siWith semantic state sjCorrelation degree:
(3) semantic state distance matrix M, is the matrix for representing the relation between semantic state, is specially:
M (i, j)=D (si,sj);
State distance describes the weak accessibility between state and state, i.e., the always too high estimation service solution of state distance can
It is if the state distance between state A and state B is infinitely great, i.e., unreachable up to property, then cause in the absence of one group of planning solution from state
A reaches state B;
(4) if original state is rin=(i1,i2...in), dbjective state rout=(o1,o2...on), service WS=(ws1,
ws2...wsn), N (S) represents the number of state contained by state set S, wherein gs(p) it is from state s to dbjective state p estimation
Value is state distance matrix M (s, p) value, then services wsiInspiration value h (wsi) be:
<mrow>
<mi>h</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>ws</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mi>&Sigma;</mi>
<mrow>
<mi>s</mi>
<mo>&Element;</mo>
<mi>e</mi>
<mi>f</mi>
<mi>f</mi>
<mi>e</mi>
<mi>c</mi>
<mi>t</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>ws</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>,</mo>
<mi>p</mi>
<mo>&Element;</mo>
<msub>
<mi>r</mi>
<mrow>
<mi>o</mi>
<mi>u</mi>
<mi>t</mi>
</mrow>
</msub>
</mrow>
</munder>
<mi>min</mi>
<mo>(</mo>
<mrow>
<msub>
<mi>g</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>p</mi>
<mo>)</mo>
</mrow>
</mrow>
<mo>)</mo>
<mo>/</mo>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>e</mi>
<mi>f</mi>
<mi>f</mi>
<mi>e</mi>
<mi>c</mi>
<mi>t</mi>
<mo>(</mo>
<mrow>
<msub>
<mi>ws</mi>
<mi>i</mi>
</msub>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
h(wsi) accessibility of current service and dbjective state set, the value are have estimated using the value of the output state distance of service
Describe the average of the state distance between the output state of service and dbjective state, h (wsi) always chosen in state distance
Minimum value estimated, h (wsi) value for it is infinitely great when, service wsiDo not appear in solution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710343392.5A CN107105052B (en) | 2017-05-16 | 2017-05-16 | Heuristic Web service combination method based on graph planning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710343392.5A CN107105052B (en) | 2017-05-16 | 2017-05-16 | Heuristic Web service combination method based on graph planning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107105052A true CN107105052A (en) | 2017-08-29 |
CN107105052B CN107105052B (en) | 2020-02-14 |
Family
ID=59668931
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710343392.5A Active CN107105052B (en) | 2017-05-16 | 2017-05-16 | Heuristic Web service combination method based on graph planning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107105052B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108647787A (en) * | 2018-03-30 | 2018-10-12 | 中山大学 | A kind of multiple agent cognition planning algorithm based on heuristic search |
CN109408046A (en) * | 2018-09-05 | 2019-03-01 | 河海大学 | A kind of shortest path web service composition method based on figure |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060069995A1 (en) * | 2004-09-30 | 2006-03-30 | British Telecommunications Public Limited Company | Personalised process automation |
CN102571752A (en) * | 2011-12-03 | 2012-07-11 | 山东大学 | Service-associative-index-map-based quality of service (QoS) perception Top-k service combination system |
-
2017
- 2017-05-16 CN CN201710343392.5A patent/CN107105052B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060069995A1 (en) * | 2004-09-30 | 2006-03-30 | British Telecommunications Public Limited Company | Personalised process automation |
CN102571752A (en) * | 2011-12-03 | 2012-07-11 | 山东大学 | Service-associative-index-map-based quality of service (QoS) perception Top-k service combination system |
Non-Patent Citations (5)
Title |
---|
XIANRONG ZHENG等: "An Efficient Syntactic Web Service Composition Algorithm Based on the planning Graph Model", 《2008 IEEE INTERNATIONAL CONFERENCE ON WEB SERVICE》 * |
XINGGUO LI等: "A Semantic Web Service Composition Method based on an Enhanced Planning Graph", 《2010 INTERNATIONAL CONFERENCE ON E-BUSINESS AND E-GOVERNMENT》 * |
YAN YANG等: "Leveraging ontology-aided AI Planning For Automatic Composition of Semantic Web Services", 《2010 3RD INTERNATIONAL CONFERENCE ON INFORMATION MANAGEMENT,INNOVATION MANAGEMENT AND INDUSTRIAL ENGINEERING》 * |
付鹏斌等: "基于图规划的语义Web服务自动组合方法", 《计算机应用研究》 * |
王向辉等: "语义Web服务自动组合定义、方法及验证调查", 《计算机应用研究》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108647787A (en) * | 2018-03-30 | 2018-10-12 | 中山大学 | A kind of multiple agent cognition planning algorithm based on heuristic search |
CN109408046A (en) * | 2018-09-05 | 2019-03-01 | 河海大学 | A kind of shortest path web service composition method based on figure |
CN109408046B (en) * | 2018-09-05 | 2022-01-28 | 河海大学 | Shortest-path Web service combination method based on graph |
Also Published As
Publication number | Publication date |
---|---|
CN107105052B (en) | 2020-02-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Mosavi | Multiple criteria decision-making preprocessing using data mining tools | |
Miljković et al. | Application of modified multi-objective particle swarm optimisation algorithm for flexible process planning problem | |
Liu et al. | Balancing codification and personalization for knowledge reuse: a Markov decision process approach | |
CN112528591B (en) | Automatic PCB wiring method based on joint Monte Carlo tree search | |
CN111126668A (en) | Spark operation time prediction method and device based on graph convolution network | |
Zhang et al. | Simulation optimization using the particle swarm optimization with optimal computing budget allocation | |
Zhang | An immune genetic algorithm for simple assembly line balancing problem of type 1 | |
Shahnaghi et al. | A robust modelling and optimisation framework for a batch processing flow shop production system in the presence of uncertainties | |
Stevens et al. | Reducing run-time adaptation space via analysis of possible utility bounds | |
Lin et al. | A binary particle swarm optimization based on the surrogate information with proportional acceleration coefficients for the 0-1 multidimensional knapsack problem | |
CN107105052A (en) | Heuristic web service composition method based on figure planning | |
Lu et al. | Formal modelling of a sheet metal smart manufacturing system by using Petri nets and first-order predicate logic | |
Naranjo et al. | PRIMROSe: a graph-based approach for enterprise architecture analysis | |
Portoleau et al. | Robust decision trees for the multi-mode project scheduling problem with a resource investment objective and uncertain activity duration | |
Dowson et al. | Bi-objective multistage stochastic linear programming | |
Wongwiwat et al. | Production scheduling for injection molding manufacture using Petri Net model | |
Ignatyev et al. | Hybrid genetic‐paired‐permutation algorithm for improved VLSI placement | |
WO2022252694A1 (en) | Neural network optimization method and apparatus | |
Haddad et al. | Solving the unrelated parallel machine scheduling problem with setup times by efficient algorithms based on iterated local search | |
Clarkson et al. | Signposting for Design Process Improvement: A dynamic approach to design process planning | |
CN112817560B (en) | Computing task processing method, system and computer readable storage medium based on table function | |
Konicki et al. | Exploiting extensive-form structure in empirical game-theoretic analysis | |
Ma et al. | A novel artificial bee colony algorithm for robust permutation flowshop scheduling | |
Samanlioglu et al. | An interactive memetic algorithm for production and manufacturing problems modelled as a multi-objective travelling salesman problem | |
Peruzzini et al. | A dependencies satisfaction tool to support the cost oriented configuration of industrial products in the conceptual design stage |
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 |