CN112435017A

CN112435017A - Modeling and construction method for cross-organization business process interaction

Info

Publication number: CN112435017A
Application number: CN202110045605.2A
Authority: CN
Inventors: 李传艺; 葛高龙; 骆斌
Original assignee: Nanjing University
Current assignee: Nanjing University
Priority date: 2021-01-14
Filing date: 2021-01-14
Publication date: 2021-03-02

Abstract

The invention relates to a modeling and construction method for cross-organization business process interaction, which comprises the steps of constructing a Token log according to actual business process interaction; preprocessing the Token log; constructing an object network of a cross-organization business process based on the preprocessed non-interactive Token logs; for the interactive Token log, a node projection graph of an object network is constructed according to interactive nodes, and then a system network and a message sequence graph of a cross-organization business process are constructed by utilizing a merging algorithm and interactive details. The method can help enterprises to dynamically change the model according to the actual implementation conditions from the aspect of the actual model, including task change, execution sequence change and resource change, so as to achieve the common goal.

Description

Modeling and construction method for cross-organization business process interaction

Technical Field

The invention relates to a workflow modeling and construction method, in particular to a modeling and construction method based on cross-organization business process interaction, and belongs to the technical field of process mining.

Background

As the external environment within an enterprise changes, more and more enterprises are becoming aware of the importance of improving the efficiency of communications with partners, suppliers, and customers. In recent years, since the cross-organizational workflow has a well-defined definition, which is beneficial for achieving business consensus in collaboration with each level of collaborating companies, the business process of the cross-organization becomes a strategy widely adopted by enterprises. The method has a good effect of solving actual problems in the enterprise supply chain by using process mining related technologies, such as process discovery, conformance inspection, process enhancement and the like. Many scholars have done a lot of important research work.

Process mining techniques are mainly divided into three major categories. The first category is process discovery. The process discovery is to generate a process model from event data from the viewpoint of control flow, and a plurality of algorithms are gradually owned for process mining, such as α, α + +, α #, β, and the like. The second type is a compliance check. The compliance check is intended to detect inconsistencies between event logs of a process model and its corresponding execution model. The third category is process enhancement. The idea of process enhancement is to extend or improve existing flow models using a log of related events recorded in the actual flow. In addition to process mining from event logs, there are also work-based process mining and compliance checking and extended research directions such as process similarity metrics, process repository management, process monitoring and event log generation, etc.

Among them, process discovery is extremely important. In a cross-organizational business process, the process finds a business process model mined from event data from a control flow perspective. For an enterprise, process discovery may mine a business process model representing the most recent structure across organizational processes from the logs of their respective process systems. The business process model may also clearly capture the entire collaboration graph between organizations as well as the detailed information of the communication nodes for the business manager.

The process models in process discovery are typically modeled using Petri nets, and also using extensions such as workflow nets (WF-net), Business Process Models and Notation (BPMN) and another workflow language (YAWL). With the proposal of the workflow view concept, the model under the view concept sets different authorities for different organizations to check the workflow and the resources thereof, thereby providing powerful functions for the workflow system between the organizations. To this end, some modeling approaches model the inter-organizational workflow system by combining the concept of a common view with an object Petri Net model. The method comprises the steps of firstly adopting a public view of each workflow to construct an information flow network of cross-organization workflows, then establishing an internal process of an organization through an object Petri network, and finally displaying a synchronization relation between communication structures. While making significant progress, they only consider the constant workflow model, i.e., no changes between the business process itself and the interactions of the business process across the organization. In practice, the flexibility of quickly implementing new processes and quickly adapting to existing changes to cope with environmental changes (e.g., business services, rules, and partners) across an organization's business processes is not as flexible as in a single business process. In the invention, we discuss the problem of modeling the cross-organization process along with the business service and rule change, combine the basic object system (EOS), the inter-organization workflow (IOWF) and the Message Sequence Chart (MSC), and pay attention to the interaction of the business process, and focus on researching a modeling and construction method of cross-organization business process interaction.

Disclosure of Invention

The invention relates to a modeling and constructing method for cross-organization business process interaction, which comprises the steps of constructing a Token log according to a log output by actual business process interaction; preprocessing the Token logs to distinguish non-interactive Token logs from interactive Token logs; constructing an object network of a cross-organization business process based on the preprocessed non-interactive Token logs; for the interactive log, firstly, a node projection graph of an object network is constructed according to interactive nodes, and then a system network and a message sequence graph of a cross-organization business process are constructed by utilizing a merging algorithm and interactive details. The method can help enterprises to dynamically change the model according to the actual implementation conditions from the aspect of the actual model, including task change, execution sequence change and resource change, so as to achieve the common goal.

1. A modeling and construction method for cross-organizational business process interaction is characterized by comprising the following steps:

constructing a Token log by using historical interactive data of a business process, wherein the Token log comprises an interactive Token log and a non-interactive Token log;

preprocessing the Token log;

mining an object workflow network participating in interaction from a non-interactive Token log by using a process mining algorithm;

constructing a projection graph of an interaction node by combining the interaction Token log and the object workflow network, and integrating a message sequence graph from the interaction Token log;

and integrating the projection graphs of different object networks according to a specific algorithm to form a system network, and combining the system network, the object network and the message sequence graph into a cross-organization business process interaction model.

2. The method of claim 1, the tokken log being: each record represents consumption of a resource or transmission of a message, and each record comprises a production task, production time, a running instance identifier (PCID) of a process to which the production task belongs, a consumption task, consumption time, and a running instance identifier (CCID) of a process to which the consumption task belongs.

3. The method of claim 1, wherein preprocessing the tokken log comprises:

dividing the log into an interactive log or a non-interactive log according to whether the PCID and the CCID in the log are the same, wherein the non-interactive log is identified by the same identifier, and the interactive log is not identified;

carrying out duplicate removal processing on the non-interactive log, then removing PCID information in the non-interactive log according to a quintuple Token log format required by a mining algorithm, and converting the CCID into a format required by a Token quintuple in the non-interactive Token log;

and dividing each interaction record in the interaction log into a synchronous interaction set or an asynchronous interaction set according to whether the two tasks have interaction.

4. The method of claim 1, wherein mining the network of object workflows participating in the interaction from the non-interactive tobken logs using a process mining algorithm comprises:

according to the sequence number of the recorded business process in the non-interactive log, obtaining a Token log related to each business process;

carrying out model mining on the single object net by using a disclosed process mining algorithm tau to obtain a single object net model;

and combining the mining results of the object networks to obtain all object network models interacting across the organizational business process.

5. The method of claim 1, wherein constructing a projected graph of interaction nodes in conjunction with the interaction tokken log and the object workflow network, and wherein integrating the message sequence graph from the interaction tokken log comprises:

setting the information that the production task and the consumption task are the same or opposite in the synchronous interaction set obtained in the claim 3 as a synchronous interaction node;

generating a message sequence diagram according to each synchronous interaction set, wherein the producer and the consumer of the Token are used as two entities, each Token represents one-time message transmission, and the message transmission is drawn according to the occurrence sequence of the Token to form the message sequence diagram;

setting the same information of the production task and the consumption task in the asynchronous interaction set obtained in the claim 3 as an asynchronous interaction node, and recording the interaction direction in the node;

removing tasks which are not in the interactive node record set in the mined object network model;

and projecting the interactive nodes in the object network operated in the previous step according to synchronization or asynchronization, wherein the synchronous nodes give no-arrow task connection lines, and the asynchronous nodes give direction connection lines.

6. The method of claim 1, wherein the combining the projection views of different object networks according to a specific algorithm to form a system network, and the combining the system network, the object network and the message sequence graph into a cross-organizational business process interaction model comprises:

defining the condition of deadlock generated in the interactive process, performing deadlock detection on related interactive nodes in the projection network according to the definition of deadlock, marking the interactive nodes forming deadlock, ensuring that the deadlock nodes do not participate in combination, stopping the whole process if deadlock exists, and returning to the construction failure;

if no deadlock exists, acquiring the structural relationship among the merging nodes, including the sequence, concurrence and selection relationship;

combining the related nodes in each service process in the projection network according to a combination rule to form a system network;

and according to the standardized specifications of the system network, the object network and the message sequence diagram, providing graphic display and text display of the modeling result.

Compared with the prior art, the invention has the following remarkable advantages: the advantages of the three models of the basic object system, the inter-organization workflow and the message sequence diagram are combined, and the layered cross-organization workflow modeling method is constructed. Using this modeling approach, an enterprise can mine the latest structure of the cross-organizational process from the logs of its respective process system. At the same time, the service manager can clearly capture the entire collaboration graph between organizations and the detailed information of the communication nodes.

Drawings

FIG. 1 is a flow chart of a modeling and construction method for cross-organizational business process interaction

FIG. 2 specific format of the Token Log

FIG. 3 exemplary diagram of a Token Log

FIG. 4 interaction model construction schematic

FIG. 5 message sequence chart construction

FIG. 6 constitutes all possible combinations of deadlocks

FIG. 7 all possible merged results for two nodes

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.

The invention provides a modeling and construction method for cross-organization business process interaction, which comprises the steps of constructing a Token log according to actual business process interaction; preprocessing the Token log; constructing an object network of a cross-organization business process based on the preprocessed non-interactive Token logs; for the interactive log, firstly, a node projection graph of an object network is constructed according to interactive nodes, and then a system network and a message sequence graph of a cross-organization business process are constructed by utilizing a merging algorithm and interactive details. The method can help enterprises to dynamically change the model according to the actual implementation conditions from the aspect of the actual model, including task change, execution sequence change and resource change, so as to achieve the common goal. The invention mainly comprises the following steps:

step (1), constructing a Token log;

step (2), preprocessing the Token log;

step (3) digging an object network;

step (4), constructing a node projection graph and a message sequence graph;

and (5) constructing a system network.

The detailed work flow of the modeling and construction method for cross-organizational business process interaction is shown in fig. 1. The above steps will be described in detail herein.

1. Because the method is based on modeling and construction of the Token log, data is needed for mining. The method models the cross-organizational business process interaction and provides a relevant Token log format, as shown in FIG. 2.

2. The tokken log is the content gathered from different business processes, and cannot be directly used for mining, and the tokken log needs to be preprocessed in the step 2. The method comprises the following specific steps:

and (2.1) separating the interactive log and the non-interactive log. The interactive log refers to information recorded when tasks in different business processes are interacted, and the non-interactive log refers to interactive records among the tasks in the same business process. Therefore, according to the characteristic, whether the unique identifiers of the running examples of the system where the production task and the consumption task are located are the same (namely whether the PCID and the CCID in the Token log are the same) is judged, the same records are divided into non-interactive logs, and the same records are divided into interactive logs if the unique identifiers are different. Fig. 3(a) is an original tobken log, and fig. 3(b) is an interactive tobken log and a non-interactive log after separation.

And (2.2) preprocessing the non-interactive log. After the non-interactive log is obtained, we preprocess it. The preprocessing is mainly divided into two steps, the first step is to remove the duplicate of the non-interactive log, and as the mining object network does not need absolute time information, other characteristics are compared, and the same records can be deleted. And the second step is that the non-interactive log is removed according to a five-tuple Token log format required by a mining algorithm, and because the CCID and the PCID in the non-interactive log are the same, the CCID and the T (time stamp of absolute time) in the log can be deleted. Meanwhile, the PCID format is converted into the format required by the quintuple Token log.

And (2.3) interactive log preprocessing. According to different constraints of the synchronous interactive node and the asynchronous communication node, the node information contained in the external token log can be classified into synchronous communication or asynchronous communication. Dividing each interaction record in the interaction log into a synchronous interaction set or an asynchronous interaction set according to whether mutual interaction exists between the two tasks;

3. in the step, the preprocessed non-interactive log is used for mining the object network.

And (3.1) separating the single object net logs. Since the PCID of the preprocessed non-interactive log indicates the target network where the production tokken log is located, the tokken log is divided into the tokken logs of n target networks according to the serial number.

And (3.2) digging the object network. Model mining is carried out on the single object network Torken log by using a process mining algorithm tau, so that a Petri network model of the single object network is obtained;

and (3.3) constructing an object network set. And combining the model results of the single object network in the previous step into a graph, thereby obtaining the object network model interacting across the organizational business process. Fig. 4(a) shows an example of an object web model.

4. After obtaining the model of the object network of the business process itself, we also project the interactive nodes into the object network.

And (4.1) constructing a synchronous interactive node set. According to the definition of the interactive node:

interactive node is type, ts, N

Wherein:

(1)

set of communication types, including asynchronous communication

And synchronous communication

(2)ts＝{t₁，...，t_m} (m ≧ 2) a set of communication-related tasks in which

(3)

And collecting the networks where the communication tasks are located.

(4) When in use

Then

(5) When in use

Then

If it is satisfied with

Conditions of (1) then

Setting the same or opposite information of the production task and the consumption task in the synchronous interaction set obtained in the step two as a synchronous interaction node;

and (4.2) constructing an asynchronous interactive node set. Setting the same information of the production task and the consumption task in the asynchronous interaction set obtained in the step two as an asynchronous interaction node according to the definition of the interaction node;

and (4.3) removing useless tasks. Since the node projection graph is used by all business personnel related to the enterprise, tasks which do not relate to interaction are removed from the nodes in advance in order to ensure the safety of a business system. If the task is a head-end task, deleting and redefining the head and the tail of the object network model, and if the task is an intermediate task, removing the task and connecting the front task and the rear task to ensure the integrity of the model.

And (4.4) node projection. And according to the name sorting of the object network, finding the tasks related to the object network from the synchronous interaction set and the asynchronous interaction set in sequence, and projecting. And projecting the interactive nodes in the object network operated in the previous step according to synchronization or asynchronization, wherein the synchronous nodes give no-arrow task connection lines, and the asynchronous nodes give direction connection lines. FIG. 4(b) shows an example of a nodal projection network;

and (4.5) constructing a system network interaction record. Because a plurality of synchronous interactive nodes can be combined into the same node in the system network, the synchronous node information in the system network is recorded based on the system network and the synchronous interactive node set.

And (4.6) constructing a message sequence chart. According to the system network interaction record of the last step and the definition of the message sequence chart:

message sequence chart (ON, T, M)

Wherein:

(1)T＝{t₁，...，t_mset of communication tasks, and ON ═ ON₁，...，ON_mIs the set of object nets for the corresponding task in T.

(2)M＝{m₁，...，m_nIs the set of messaging for the communication task.

(3)

Must satisfy t_s∈T，t_r∈T。

And (3) each system network synchronous interactive node gives graphic display and text display according to the standardized specification of the message sequence diagram. Fig. 5(a) gives a graphical representation of an example, and fig. 5(b) gives a textual representation of an example.

5. After the nodes are projected, the nodes are merged according to rules to form a system network, and a message sequence chart is used for displaying the specific information of synchronous interaction.

And (5.1) deadlock detection. Before node merging we first do deadlock detection. Because the interaction of a plurality of nodes is carried out in two business processes, and the interaction and the original execution sequence in the business process easily form a closed loop to form deadlock, the work order of the business process is seriously influenced. We give all the possibilities of constituting deadlock in two business processes, as shown in fig. 6, and perform deadlock detection on the relevant interactive nodes in the projection network. And if the deadlock is detected, directly prompting without carrying out subsequent merging work.

And (5.2) acquiring the structural relationship between the merging nodes. The merging principle is to determine the merged structure according to the structural relationship between the nodes, so we use the TS invariant to obtain it.

And (5.3) constructing a system network. And combining the node projection graph obtained in the step 4 and the structural relationship between the combined nodes to perform node combination, wherein all possible forms of combination are given according to full arrangement, as shown in fig. 7. Firstly, selecting two object networks with interactive relation, merging all related nodes to form a temporary system network, then selecting another object network related to the temporary system network to merge, repeating the steps until no object network which is not merged yet exists, and obtaining the final system network.

A cross-organizational business process interaction modeling and construction method implemented in accordance with the present invention has been described in detail with reference to the accompanying drawings, FIG. 4 is a sample of an interaction business process and an operational flow for illustrating a method implementation. The invention has the following advantages: the advantages of the three models of the basic object system, the inter-organization workflow and the message sequence diagram are combined, and the layered cross-organization workflow modeling method is constructed. Using this modeling approach, an enterprise can mine the latest structure of the cross-organizational process from the logs of its respective process system. At the same time, the service manager can clearly capture the entire collaboration graph between organizations and the detailed information of the communication nodes.

It is to be understood that the invention is not limited to the specific arrangements and instrumentality described above and shown in the drawings. Also, a detailed description of known process techniques is omitted herein for the sake of brevity. The present embodiments are to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Claims

preprocessing the Token log;

3. The method of claim 1, wherein preprocessing the tokken log comprises: