CN116471281A - Decentralised service combination method considering node selfiness - Google Patents
Decentralised service combination method considering node selfiness Download PDFInfo
- Publication number
- CN116471281A CN116471281A CN202310248327.XA CN202310248327A CN116471281A CN 116471281 A CN116471281 A CN 116471281A CN 202310248327 A CN202310248327 A CN 202310248327A CN 116471281 A CN116471281 A CN 116471281A
- Authority
- CN
- China
- Prior art keywords
- service
- demand
- node
- services
- selfiness
- 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
- 238000000034 method Methods 0.000 title claims abstract description 62
- 239000011159 matrix material Substances 0.000 claims abstract description 21
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 16
- 238000013145 classification model Methods 0.000 claims abstract description 11
- 238000003860 storage Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 8
- 230000008901 benefit Effects 0.000 claims description 7
- 230000004044 response Effects 0.000 claims description 7
- 230000002068 genetic effect Effects 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000012886 linear function Methods 0.000 claims description 3
- 239000000203 mixture Substances 0.000 abstract description 8
- 238000013461 design Methods 0.000 abstract description 5
- 238000000605 extraction Methods 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 230000010354 integration Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000037213 diet Effects 0.000 description 1
- 235000005911 diet Nutrition 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/104—Peer-to-peer [P2P] networks
- H04L67/1059—Inter-group management mechanisms, e.g. splitting, merging or interconnection of groups
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a decentralization service combination method considering node selfiness. The method first models the selfiness of the service provider. Then, in order to improve the combination efficiency and the quality of the solution, a rule of thumb is utilized in the history combination, a demand mode, a service mode and an association matrix are proposed to mine commonalities in the history combination, and a multi-label classification model extraction characteristic is designed to reduce the scale of candidate services. Based on these services, a service composition algorithm and service coordination protocol are designed to coordinate the service providers and build service solutions. To protect privacy between service providers, models are trained based on a mature federal learning framework and privacy disclosure is avoided in algorithm and protocol design.
Description
Technical Field
The invention is suitable for the technical field of service computing, and particularly relates to a decentralization service combination method considering node selfiness.
Background
With the rapid development of theory and technology such as big data and cloud computing, a large number of software services appear on the internet. Various physical resources and manual services are also accessed on the Internet through virtualization and Internet of things technologies, and cooperation and integration are established with software services. With the accompanying increasing complex user demands, single services or preset service schemes are difficult to meet. Therefore, it is necessary to provide a composite service solution for users by dynamically aggregating various services, not only to meet their functional needs, but also to optimize their non-functional needs for quality, value, etc.
As a countermeasure, researchers have proposed service combinations and devised a series of methods. These methods optimize for a centralized service library, and can aggregate multiple services into one service solution. However, these methods often have some distraction when applied, mainly because of the different distribution of services in reality, which is manifested by the following features: (1) Services are typically distributed across different third party service platforms or providers (nodes). The platforms are distributed in a decentralized service network, and a centralized service resource library is not provided for service combination; (2) In a decentralized service network, a single platform often does not have the ability to provide service solutions alone, requiring collaboration between platforms; (3) Complete data, such as user requirements, complete service libraries, etc., is not shared between platforms due to business confidentiality or competition. Meanwhile, the platform is selfish, such as not providing services to other platforms due to insufficient resources, and the like. The present invention refers to such a combination of services that both cooperate and compete between platforms in a decentralized environment as a decentralized service combination.
Currently, some researchers try to solve the problem of decentralizing service composition from different angles, such as expanding the traditional service composition method to a decentralizing environment, designing middleware to ensure data security in the interaction process or designing a decentralizing algorithm to directly solve the problem. However, these methods have the following general problems: (1) They generally perform in terms of execution efficiency and solution quality, and it is difficult to achieve performance comparable to conventional methods; (2) They rarely take into account incomplete data sharing between service providers, most research focusing on privacy between users and service providers; (3) They do not take into account the selfiness of the service provider, but in reality the service provider may not be completely private. Thus, existing methods have difficulty supporting rapid and accurate construction of service plans in a decentralized environment.
Disclosure of Invention
The invention aims to solve the problems that the current decentralization service combination technology is low in efficiency, the quality of a construction scheme is insufficient, data privacy and selfiness among service providers are less considered, and the like, and provides a decentralization service combination method considering node selfiness.
The invention is realized by the following technical scheme, and provides a decentralization service combination method considering node selfiness, which specifically comprises the following steps:
step S1, modeling the selfish of a service provider; in a decentralized service network, when other service providers, i.e. service nodes, make a request to a certain node to use a certain service, this node is for some reason not willing to provide the service, referred to as node is selfish with the service; the selfiness of the node for each service is different, assessed by selfiness; calculating the selfish degree of each service by using a selfish degree definition formula, and updating in time when running;
s2, selecting a proper candidate service mode; extracting commonalities in the historical combination by using the demand mode, the service mode and the incidence matrix; the node first uses a demand pattern to match demands, called demand matching, which breaks the demands into a set of matched demand patterns and a set of unmatched sub-demands; then, based on the current environment, the demand modes are mapped to service modes through an association matrix, so that the association probability of each service mode is given, and the service modes with high association probability are used for subsequent combination;
s3, reducing the search space of the candidate service; training a multi-label classification model under the federal learning framework, and finding a proper service for the unmatched sub-requirements in the step S2;
s4, coordinating all nodes to construct a service scheme; utilizing a centralized service combination method to optimize the candidate service modes and services in the step S2 and the step S3, constructing an initial service solution, and providing service use requests to other nodes, wherein the other nodes selectively provide services according to the selfiness of the other nodes, and the responses of the other nodes are divided into three types: agreeing to provide, disagreeing to provide but providing other services and disagreeing to provide; the node receiving the demand reconstructs the initial service solution from the response, and when all services in the initial service solution are available, a final service solution is obtained.
Further, the selfish degree comprises objective selfish degree and subjective selfish degree, wherein the objective selfish degree is influenced by objective conditions of the node, and the subjective selfish degree follows subjective willingness of the node; the selfish S is defined as follows:
S＝f(S R ，S P ，ω)
wherein S is R E (0, 1) is the available service resource factor, S P ∈[0，1]Is a benefit factor of service execution, ω is a subjective willingness factor of the node, and f is a linear function; the first two factors are defined as follows:
where R is the number of available service resources, ω R And κ is a parameter that measures the influence of R on the selfiness, P is the service execution benefit, ω P And P 0 Is a parameter for measuring the influence of P on the selfish degree.
Further, each node sets a threshold for the selfiness of a service, and when the selfiness of the service is less than the threshold, the node is willing to provide the service.
Further, when services in a service network cooperate to meet the demand, there are often fixed collocations called service patterns, and there are two types of service patterns, one is designed by a domain expert, and the other is a potential service pattern, which needs to be mined through historical combination or service knowledge patterns.
Further, defining pairs to represent the association relationship between the demand mode and the service mode, and converging all pairs together to obtain a three-dimensional matrix comprising the demand mode, the service mode and the situation, which is called an association matrix.
Further, the multi-label classification model is divided into three parts:
the demand part: firstly, extracting functional and non-functional requirements in user requirements, modeling the functional and non-functional requirements as a constraint graph, and then, using a graph convolution network to learn a requirement representation;
service part: firstly, extracting service characteristics, modeling the service as a service vector, then using a multi-layer perceptron to represent the service, and using a transducer encoder to find the dependency relationship between the services to obtain service embedding;
the transducer section: using a cross-attention module in the transformer decoder, services are embedded as queries, features associated with the services are detected from the demand representation, and the probability that each service is used to satisfy the demand is predicted.
Further, the result of demand matching consists of two parts: a matched set of demand patterns and an unmatched set of sub-demands; for each demand pattern, finding a service pattern with high association probability from the association matrix based on the current user context; for the sub-requirement set, the search space will be further reduced by step S3.
Further, the centralized service combination method is a centralized service combination method based on a genetic algorithm.
The invention provides an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps of the decentralizing service combination method considering node selfiness when executing the computer program.
The present invention proposes a computer readable storage medium storing computer instructions which, when executed by a processor, implement the steps of the method for decentralizing service composition that takes account of node selfiness.
The invention has the beneficial effects that:
the invention provides a method for combining the decentralized services by considering node selfiness, which aims to solve the problems that the current technology for combining the decentralized services is low in efficiency, the quality of a construction scheme is insufficient, the data privacy and selfiness among service providers are less considered, and the like. The method first models the selfiness of the service provider. Then, in order to improve the combination efficiency and the quality of the solution, experience rules are utilized in the history combination, a demand mode, a service mode and an association matrix are provided for mining commonalities in the history combination, and a multi-label classification model extraction characteristic is designed, so that the scale of candidate services is reduced. Based on these services, a service composition algorithm and service coordination protocol are designed to coordinate the service providers and build service solutions. To protect privacy between service providers, models are trained based on a mature federal learning framework and privacy disclosure is avoided in algorithm and protocol design.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for combining decentralized services in consideration of node selfiness according to the present invention;
FIG. 2 is a block diagram of the overall idea of a method for combining decentralized services, which considers node selfiness;
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
1-2, the invention provides a method for combining decentralizing services by considering node selfiness, which specifically comprises the following steps:
step S1, modeling the selfish of a service provider; in a decentralized service network, when other service providers, i.e. service nodes, make a request to a certain node to use a certain service, this node is for some reason not willing to provide the service, called node is selfish to the service, and the selfiness of the node to each service is different, as the willingness of the node to provide each service is different, assessed by the selfiness; calculating the selfish degree of each service by using a selfish degree definition formula, and updating in time when running;
selfish includes objective and subjective selfish, where objective selfish is affected by node objective conditions, such as available service resources and service execution revenues. The fewer service resources are available, the less revenue is performed by the service, and the higher the privacy of the node to the service. Subjective selfish follows the subjective willingness of a node, such as being more private when signing a cooperation protocol with other nodes. Referring to the epidemic model and the hyperbolic selfish behavior, the selfish S is defined as follows:
S＝f(S R ，S P ，ω)
wherein S is R E (0, 1) is the available service resource factor, S P ∈[0，1]Is a service execution benefit factor, ω is a node subjective willingness factor, f is a linear function, and the weight is determined by the importance of the three influencing factors; the first two factors are defined as follows:
where R is the number of available service resources, ω R And κ is a parameter that measures the influence of R on the selfiness, larger κ and smaller ω R Representing a more non-private node, i.e., a less selfish. P is the service execution benefit, ω P And P 0 Is to measure P pair selfishParameters of degree influence. Smaller omega P And smaller P 0 Representing a more private node.
In use, each node sets a threshold for selfiness. When the selfiness of a service is less than this threshold, the node is willing to offer the service. Furthermore, when one node does not want to cooperate with another node, it will increase its subjective selfiness, ω.
S2, selecting a proper candidate service mode; to improve the combining efficiency and the quality of the solution, the commonalities in the historical combination are extracted by using the demand mode, the service mode and the incidence matrix; the node first uses a demand pattern to match demands, called demand matching, which breaks the demands into a set of matched demand patterns and a set of unmatched sub-demands; then, based on the current environment, the demand modes are mapped to the service modes through the association matrix, so that the association probability of each service mode is given, and the service modes with high association probability are used for subsequent combination, so that the problem scale can be effectively reduced, and the combination efficiency is improved;
it can be seen from the historical user demands that there is a correlation between sub-demands in some demands. For example, the goal of travel will most often be a attraction, and the needs of the attraction are often associated with surrounding traffic, diet, housing, etc. These sub-demand segments always appear together in some user demands, which are mined and characterized as demand patterns.
When services in a service network cooperate to meet a demand, there is often a fixed collocation, called a service mode. The service patterns essentially express "a priori knowledge" accumulated during the execution of the service network and represent implicit rules for the combination of services within a particular domain or between domains. There are two types of service modes, one designed by a domain expert, such as travel agencies issuing travel plans. The other is a potential service model, which needs to be mined by historical combinations or service knowledge patterns.
In order to use patterns in a combination to reduce the search scale, it is necessary to further explore the correlation between service patterns and demand patterns. The correlation between patterns is many-to-many, that is, there are multiple service patterns that can match one demand pattern and one service pattern that can match multiple demand patterns. Pairs are defined to represent the association of demand patterns and service patterns. The probability of association of the relationship varies in different contexts. Common contextual elements include user context, such as age, gender, relationship, etc., and environmental context, such as time, date, place, season, etc. All pairs are clustered together to obtain a three-dimensional matrix comprising demand patterns, service patterns and contexts, called an incidence matrix. The calculation of the associated probabilities is mainly based on the number of successful matches between demand patterns and service patterns, the time correlation of historical combinations, the constraint satisfaction of service patterns, and the provided probability predictions of services of other nodes.
S3, reducing the search space of the candidate service; training a multi-label classification model under the federal learning framework, and finding a proper service for the unmatched sub-requirements in the step S2;
the multi-label classification model is divided into three parts:
the demand part: firstly, extracting functional and non-functional requirements in user requirements, modeling the functional and non-functional requirements as a constraint graph, and then, using a graph rolling network (GCN) to learn a requirement representation;
service part: firstly, extracting service characteristics and modeling the service as a service vector, then, using a multi-layer perceptron (MLP) to represent the service, and using a transducer encoder to find the dependency relationship between the services to obtain service embedding;
the transducer section: using a cross-attention module in the transformer decoder, services are embedded as queries, features associated with the services are detected from the demand representation, and the probability that each service is used to satisfy the demand is predicted.
And the service with high probability is selected for subsequent combination, so that the candidate service scale can be reduced, and the combination efficiency is further improved. The multi-label classification model is trained under the federal learning framework to effectively protect node data privacy.
S4, coordinating all nodes to construct a service scheme; utilizing a centralized service combination method to optimize the candidate service modes and services in the step S2 and the step S3, constructing an initial service solution, and providing service use requests to other nodes, wherein the other nodes selectively provide services according to the selfiness of the other nodes, and the responses of the other nodes are divided into three types: agreeing to provide, disagreeing to provide but providing other services and disagreeing to provide; the node receiving the demand (initial node) reconstructs the initial service solution from the response, and when all services in the initial service solution are available, a final service solution is obtained.
Examples
The invention provides a method for combining decentralizing services by considering node selfiness, wherein the flow of the method is shown in fig. 1, the whole thought is shown in fig. 2, and the method comprises the following steps:
step S1, modeling the selfiness of the service provider
And calculating the selfish degree of each service by using a selfish degree definition formula, and updating in time when running by using the service provider.
Step S2, selecting a proper candidate service mode
And constructing a demand mode, a service mode and an association matrix according to the historical service combination record of the service provider.
In order to use the bilateral pattern and the correlation matrix for the de-centralized service combination, it is necessary to use the demand pattern to cover the user demand, i.e. demand matching. In demand matching, "one demand pattern can cover a part of demand" requires that the following conditions are satisfied: (1) The sub-requirements and flows in the requirements model are the same as the partial requirements; (2) The constraints in the demand pattern may contain all of the constraints for the portion of demand. There are many results of demand matching, and the following two rules are introduced to evaluate these results: (1) Solutions that use fewer demand patterns and have higher sub-demand coverage are more optimal; (2) A solution where the average frequency of usage of demand patterns is higher is preferable.
This problem is similar to the aggregate coverage problem, and this step is optionally solved using a greedy policy-based demand matching method, as shown in algorithm 1. The greedy strategy of this approach prioritizes demand patterns with more sub-demands and more frequent use.
The result of demand matching consists of two parts: a matched set of demand patterns and an unmatched set of sub-demands. For each demand pattern, a service pattern with a high probability of association is found from the association matrix based on the current user context. For the sub-requirement set, the search space will be further reduced by step S3.
Step S3, reducing search space of candidate service
This step designs a multi-label classification model to predict the probability that each service can meet the unmatched sub-requirement set of step S2. The entire model can be divided into three parts: a requirements section, a services section, and a transducer decoder section.
For the demand portion, the demand is modeled with a constraint map in view of the advantage that it can relate functional and non-functional demands together. The constraint graph is composed of nodes and edges, wherein the nodes comprise child demand nodes and constraint nodes, and when a child demand limited by a constraint comprises a child demand, the constraint nodes and the child demand nodes have one edge. Considering that the constraint graph is not too large, a graph rolling network (GCN) is selected to extract the demand features therein. Let a certain node in the graph be n i The GCN updates the node with the following formula:
wherein, the liquid crystal display device comprises a liquid crystal display device,Θ is a trainable parameter matrix and D is the dimension of the GCN hidden layer. After obtaining the output of the last layer of GCN, sub-demand nodes in the graph are extracted to be groupedForming the demand feature F. This part will be trained together with the service part in the transducer decoder part.
For the services section, note that some studies design a service relationship graph and use a Graph Neural Network (GNN) to extract service embeddings. Considering that the unmatched sub-demand sets are not typically satisfied by commonly used services, the relationships between them are less in the historical portfolio. Thus, the present invention does not use this method, but directly encodes services as vectors, and uses a multi-layer perceptron (MLP) to obtain service embedding, and a transducer encoder to discover dependencies between services.
A service vector is first defined. The service vector contains two parts, the first part being the service function, occupying one dimension. The second part is quality of service (QoS), with each normalized QoS occupying one dimension. The coded service vector is then projected to a D-dimensional vector using MLP, matching the required dimension. Finally, to obtain dependencies between services, S' ∈R is embedded for all services using a transducer encoder defined as follows k×D Self-attention was given:
where k is the number of services,is an intermediate variable, and MultiHead (value) and FFN (x) are functions defined by a standard transducer encoder.
After obtaining the demand and service features, the service features are used as queries and cross-attention is paid, and service related features are assembled from the demand features using a multi-layer transducer decoder. the transducer decoder consists of a self-attention mechanism, a cross-attention mechanism and a position feed forward network. Each decoder layer i is from its previous layer Q i-1 The update is defined as follows:
wherein the method comprises the steps ofAnd->Is two intermediate variables. Notably, in the cross-attention module, key and value are demand features. Each service feature examines the desired feature where attention is needed and selects features of interest to combine. Thus Q i-1 Will be updated layer by layer and gradually inject the required information by cross-attention.
Assuming that the encoder has L layers in total, the output of the L-layer encoder is Q L ∈R k×D . For the multi-label classification problem, each service feature Q is classified using a linear layer and a sigmoid function L，j Probability p projected to prediction j . To solve the problem of sample imbalance, an asymmetric loss is used in training as a loss function:
wherein y is j =1 indicates that the j-th service is in the best service scheme corresponding to the demand, y j The opposite is true for =0. Since a mature federal learning framework is available, the method of the present invention selects FATE (Federated AI Technology Enabler) to train the model. This framework can enable data collaboration while protecting data security and privacy.
The predicted high probability service will be used in step S4 with the high probability of association service pattern in step S2 to build a service solution.
Step S4, coordinating all nodes to construct service scheme
The steps are divided into two parts: centralized service composition and node coordination.
For the first partThe results of the first two steps are processed first. For the demand pattern obtained after the demand matching in step S2, the first k service patterns SP with the highest association probability are selected from the association matrix match . For each unmatched sub-demand obtained after demand matching, selecting the top k services S with highest prediction probability from the multi-label classification model match . Then, a centralized service composition method is designed to construct an initial service scheme. There are many optimization methods available, including meta-heuristics, reinforcement learning, etc. Considering that the constraint condition is complex, the action space is not fixed, and the meta-heuristic algorithm is selected for optimization. The Genetic Algorithm (GA) was found to perform well in question, compared to a variety of meta-heuristic algorithms. The centralized service combination method based on the genetic algorithm is shown in the algorithm 2.
The algorithm firstly initializes population scale pop, and the maximum iteration times t max And randomly generates pop service schemes P. Then, for each service scenario sol i Crossover and mutation are performed according to probabilities. The interdigitation step refers to the algorithm randomly selecting a service scheme sol j At randomly selected locations, it is combined with sol i Split and swap right. The mutation step means sol i Each of the services is replaced by another service according to the probability of variation. Cross and mutated sol i Added to the new population P' and the best service plan sol updated according to P * . P' will also go through a selection step to generate the population for the next iteration. This step is sequentially selected from P' using roulette selection. When the total iteration number exceeds t max When the algorithm stops. At this time, sol * Is the initial service plan obtained.
For the second part, due to the sol obtained in the last part * Possibly containing services from other nodes, node n receiving the demand i It is necessary to provide these services to the node n where they are located j A use request is initiated. n is n i Initiate request and request from n j The step of receiving the response is referred to as a node coordination protocol. The step designs a node coordination protocol according to the following reputation principle: when one node makes a service usage request to another node, the service must appear in the final service solution unless the node rejects.
When n is j When a service use request is received, it responds in three ways.
Consent to provide: according to the selfish degree defined in step S1, n j Agreeing to provide services.
Reject but provide other services: based on the selfish degree defined in step S1, n j The service is refused to be provided, but QoS-like services may be provided.
Refusing to provide: based on the selfish degree defined in step S1, n j Service provision is denied.
Wherein the similarity of "QoS-like services" is evaluated with euclidean distance. At n i After receiving all replies, if there are unusable services, it will re-find the high probability services corresponding to these sub-demands (excluding unusable services) from the candidate service reduction results. If n j Other services are provided that will be added to the high probability service set. This high probability service set will repeat step S4 and optimize as the next input.
When all nodes agree to use these services, sol * Which is the final service solution.
The invention provides an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps of the decentralizing service combination method considering node selfiness when executing the computer program.
The present invention proposes a computer readable storage medium storing computer instructions which, when executed by a processor, implement the steps of the method for decentralizing service composition that takes account of node selfiness.
The memory in embodiments of the present application may be either volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory. The nonvolatile memory may be a Read Only Memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable EPROM (EEPROM), or a flash memory. The volatile memory may be random access memory (random access memory, RAM) which acts as an external cache. By way of example, and not limitation, many forms of RAM are available, such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous DRAM (SLDRAM), and direct memory bus RAM (DRRAM). It should be noted that the memory of the methods described herein is intended to comprise, without being limited to, these and any other suitable types of memory.
In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, the processes or functions described in accordance with embodiments of the present application are produced in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a high-density digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a Solid State Disk (SSD)), or the like.
In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The steps of a method disclosed in connection with the embodiments of the present application may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in the processor for execution. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method. To avoid repetition, a detailed description is not provided herein.
It should be noted that the processor in the embodiments of the present application may be an integrated circuit chip with signal processing capability. In implementation, the steps of the above method embodiments may be implemented by integrated logic circuits of hardware in a processor or instructions in software form. The processor may be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, or discrete hardware components. The disclosed methods, steps, and logic blocks in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software modules in a decoded processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method.
The invention has been described in detail with respect to a method for combining decentralized services and considering node selfiness, and specific examples are applied to illustrate the principles and embodiments of the invention, and the description of the above examples is only for helping to understand the method and core ideas of the invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.
Claims (10)
1. A decentralizing service combination method considering node selfiness is characterized in that: the method specifically comprises the following steps:
step S1, modeling the selfish of a service provider; in a decentralized service network, when other service providers, i.e. service nodes, make a request to a certain node to use a certain service, this node is for some reason not willing to provide the service, referred to as node is selfish with the service; the selfiness of the node for each service is different, assessed by selfiness; calculating the selfish degree of each service by using a selfish degree definition formula, and updating in time when running;
s2, selecting a proper candidate service mode; extracting commonalities in the historical combination by using the demand mode, the service mode and the incidence matrix; the node first uses a demand pattern to match demands, called demand matching, which breaks the demands into a set of matched demand patterns and a set of unmatched sub-demands; then, based on the current environment, the demand modes are mapped to service modes through an association matrix, so that the association probability of each service mode is given, and the service modes with high association probability are used for subsequent combination;
s3, reducing the search space of the candidate service; training a multi-label classification model under the federal learning framework, and finding a proper service for the unmatched sub-requirements in the step S2;
s4, coordinating all nodes to construct a service scheme; utilizing a centralized service combination method to optimize the candidate service modes and services in the step S2 and the step S3, constructing an initial service solution, and providing service use requests to other nodes, wherein the other nodes selectively provide services according to the selfiness of the other nodes, and the responses of the other nodes are divided into three types: agreeing to provide, disagreeing to provide but providing other services and disagreeing to provide; the node receiving the demand reconstructs the initial service solution from the response, and when all services in the initial service solution are available, a final service solution is obtained.
2. The method according to claim 1, characterized in that: the selfish degree comprises objective selfish degree and subjective selfish degree, wherein the objective selfish degree is influenced by objective conditions of the nodes, and the subjective selfish degree follows subjective willingness of the nodes; the selfish S is defined as follows:
S＝f(S R ,S P ,ω)
wherein S is R E (0, 1) is the available service resource factor, S P ∈[0,1]Is a benefit factor of service execution, ω is a subjective willingness factor of the node, and f is a linear function; the first two factors are defined as follows:
where R is the number of available service resources, ω R And κ is a parameter that measures the influence of R on the selfiness, P is the service execution benefit, ω P And P 0 Is a parameter for measuring the influence of P on the selfish degree.
3. The method according to claim 2, characterized in that: each node sets a threshold for the selfiness of a service, and when the selfiness of the service is less than the threshold, the node is willing to provide the service.
4. The method according to claim 1, characterized in that: when services in a service network cooperate to meet the demand, there are often fixed collocations called service patterns, and the service patterns are of two types, one is designed by a domain expert, and the other is a potential service pattern, and need to be mined through historical combination or service knowledge patterns.
5. The method according to claim 1, characterized in that: defining pairs to represent the association relation between the demand mode and the service mode, and converging all pairs together to obtain a three-dimensional matrix containing the demand mode, the service mode and the situation, which is called an association matrix.
6. The method according to claim 1, characterized in that: the multi-label classification model is divided into three parts:
the demand part: firstly, extracting functional and non-functional requirements in user requirements, modeling the functional and non-functional requirements as a constraint graph, and then, using a graph convolution network to learn a requirement representation;
service part: firstly, extracting service characteristics, modeling the service as a service vector, then using a multi-layer perceptron to represent the service, and using a transducer encoder to find the dependency relationship between the services to obtain service embedding;
the transducer section: using a cross-attention module in the transformer decoder, services are embedded as queries, features associated with the services are detected from the demand representation, and the probability that each service is used to satisfy the demand is predicted.
7. The method according to claim 1, characterized in that: the result of demand matching consists of two parts: a matched set of demand patterns and an unmatched set of sub-demands; for each demand pattern, finding a service pattern with high association probability from the association matrix based on the current user context; for the sub-requirement set, the search space will be further reduced by step S3.
8. The method according to claim 1, characterized in that: the centralized service combination method is a centralized service combination method based on a genetic algorithm.
9. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1-8 when the computer program is executed.
10. A computer readable storage medium storing computer instructions which, when executed by a processor, implement the steps of the method of any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310248327.XA CN116471281A (en) | 2023-03-15 | 2023-03-15 | Decentralised service combination method considering node selfiness |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310248327.XA CN116471281A (en) | 2023-03-15 | 2023-03-15 | Decentralised service combination method considering node selfiness |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116471281A true CN116471281A (en) | 2023-07-21 |
Family
ID=87183251
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310248327.XA Pending CN116471281A (en) | 2023-03-15 | 2023-03-15 | Decentralised service combination method considering node selfiness |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116471281A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117171551A (en) * | 2023-11-02 | 2023-12-05 | 山东港口科技集团烟台有限公司 | Large-scale industrial equipment data analysis and intelligent management method |
-
2023
- 2023-03-15 CN CN202310248327.XA patent/CN116471281A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117171551A (en) * | 2023-11-02 | 2023-12-05 | 山东港口科技集团烟台有限公司 | Large-scale industrial equipment data analysis and intelligent management method |
| CN117171551B (en) * | 2023-11-02 | 2024-01-30 | 山东港口科技集团烟台有限公司 | Large-scale industrial equipment data analysis and intelligent management method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11227190B1 (en) | Graph neural network training methods and systems | |
| CN114048331A (en) | Knowledge graph recommendation method and system based on improved KGAT model | |
| CN109255586B (en) | Online personalized recommendation method for e-government affairs handling | |
| CN112966114B (en) | Literature classification method and device based on symmetrical graph convolutional neural network | |
| EP3482353A1 (en) | Machine learning and prediction using graph communities | |
| CN110852447A (en) | Meta learning method and apparatus, initialization method, computing device, and storage medium | |
| CN107038184B (en) | A kind of news recommended method based on layering latent variable model | |
| US20220138502A1 (en) | Graph neural network training methods and systems | |
| CN115293919B (en) | Social network distribution outward generalization-oriented graph neural network prediction method and system | |
| CN112131261B (en) | Community query method and device based on community network and computer equipment | |
| CN110321492A (en) | A kind of item recommendation method and system based on community information | |
| CN116471281A (en) | Decentralised service combination method considering node selfiness | |
| CN115964568A (en) | Personalized recommendation method based on edge cache | |
| CN115271980A (en) | Risk value prediction method and device, computer equipment and storage medium | |
| Jahandideh et al. | Allocating duplicate copies for IoT data in cloud computing based on harmony search algorithm | |
| CN114358111A (en) | Object clustering model obtaining method, object clustering method and device | |
| CN113609337A (en) | Pre-training method, device, equipment and medium of graph neural network | |
| CN115438755B (en) | Incremental training method and device for classification model and computer equipment | |
| CN116308644A (en) | Recommendation method, device, equipment, storage medium and product of target product | |
| CN112927810B (en) | Smart medical response method based on big data and smart medical cloud computing system | |
| CN114898184A (en) | Model training method, data processing method and device and electronic equipment | |
| Xu et al. | [Retracted] A Multi‐RNN Research Topic Prediction Model Based on Spatial Attention and Semantic Consistency‐Based Scientific Influence Modeling | |
| CN113343100A (en) | Wisdom city resource recommendation method and system based on knowledge graph | |
| Bose | Data mining in tourism | |
| Azondekon | Modeling the Complexity and Dynamics of the Malaria Research Collaboration Network in Benin, West Africa: papers indexed in the Web Of Science (1996—2016) |
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 |