Research On Modeling Methods For Goal-Oriented Migrating Workflow

According to the definition of Workflow Management Collation (WfMC), the workflow is defined as follows:the automation of business process, in whole or part, during which documents, information or tasks are passed from one participant to another for action, according to set of procedural rules. The model of workflow management includes two phases:modeling and execution, where the main task of modeling is to fulfill the definition of business process and the main task of workflow execution is to instance the business process and to drive its execution of workflow engine.Migrating workflow is an emerging technology that applies the mobile-agent computing paradigm to workflow management. According to the definition of migrating workflow framework presented by Professor Zeng Guangzhou, migrating instance (MI), work place and migration workflow management engine constitute the three most important elements, where (1) MI is the mobile agent migration paradigm and is the main part of execution, which can move between work places followed by its instructions and fulfill one or more tasks using the service supplied by the workplace, multiple MIs can collaborate to complete a business process; (2) work place is the agent of workflow participant, which represents the bureau or individuals to provide run-time instances of services and workflow services; (3) Migrating workflow management engine not only supports the workflow definition, and support the creation of migrating instance, dispatch and monitoring.In migrating workflow model, the description of MI, either for business process-oriented definition or business goal-oriented definition, and were known as the migration process-oriented or goal-oriented workflow system. The advantages of process-oriented description of MI is that the tasks and rules are intuitionistic, such as Wf-Net, Petri nets, the disadvantage is that the process of pre-defined path is not only a lack of flexibility, but also for users master knowledge of complete business process. It is very difficult for inter-agency large-scale collaborative business processes to master complete business process knowledge. Studies show that it is more easy for describe a goal than a process. So; let a MI carry description of goal-oriented, not only can reduce the migration of instances of the knowledge creator burden, but also increase the workflow path flexibility by finding services or the location of the goal service work and other mechanisms recommended. In this paper, we focus on the study of modeling methods for goal-oriented migrating workflow system, called goal-oriented migrating workflow modeling methods. The main contributions of this thesis are described as follows:1. Study of goal and/or graph based on timing constraints and dependencies of goal satisfiability.The basic method of goal planning is to decompose into an "and/or" tree, which can clearly describe "is-a" relation and "and/or" constraints among goals and sub-goals. But it can not show the time constraints and dependencies relation of satisfiability among them. So, it is lack of capacity for describing a goal.In this paper, we improve the method in two ways:(1) importing timing relationship between goals, and turn goal tree into goal graph, established a timing constraints "and/or" goal graph; (2) importing goal satisfiability/deniability among goals based on time constraints goal graph, established a timing constraints "and/or" goal graph with satisfiability/deniability dependencies, put forward an algorithm for satisfiability/deniability transmission. Analysis and application show that the goal graph with time constraints and satisfiability dependencies has a good workflow semantics, goals descriptive capability and power of goal satisfiability reasoning.2. Study of Goal Description Logics methods.In the fields of intelligent agent, the goal is divided into two types:declarative goal and procedural goal, the former one represents the expected state to achieve, has the integrity and stability, the latter one shows the expectations of the action agent (Goal to do), with partial and dynamic. Goal description of the existing methods, such as 3APL, GOAL, focus on the declared type of goal, or focus on process-based goals, are lack of comprehensive descriptive ability.In this paper, we construct a framework with explicit representation and formal semantics of goals-Goal Description Logics (GDLs), which integrate two types'goal: declarative goal and procedural goal into one concept based on Description Logics (DLs) by importing goal inclusion, goal equation and goal plan rules. On the other hand, some reasoning problems, such as goal consistency and goal satisfiability, are discussed. It is shown that GDLs can effectively describe the workflow goals and goal plan process with a good goal consistency determinant, satisfiability determinant and implementation of exception handling capabilities.3. Multi-goals optimization methods based on goal plan combination.In goal-oriented migrating workflow system, MI may simultaneously implement a number of business goals, where there may be some of the same goal planning steps (actions). So, in the phrase of modeling, it is significant for improving the efficiency of migration by optimizing the goal planning.In this paper, based on GDLs, we decomposed effects into definite effects and possible effects, and discuss the optimization of intelligent agents'behaviors, including DD, DP and PP. An algorithms for goal combination is put forward and proved time complexity O(n3) theoretically, n is the number of MIs. Analysis and application show that the multi-goal planning model for optimization reduces migration and improve its efficiency.4. Study of modeling for migrating workflow System based on GDLs.Aimed to the requirement of GDLs, we extend the structure of migrating workflow system proposed by Professor Zeng Guangzhou and illustrates goals determine satisfiability, multi-goal optimization model of the modeling process using travel services booking. Application and example shows that, the migration to goal-oriented workflow modeling using GDLs provides a well formal basis and good practical value.The main innovative contributions of this thesis are:1. A timing constraints and dependencies of the goal "and/or" graph with satisfiability is proposed.Time constraints define the order of the realization of sub-goals, satisfiability dependency relations implies transmit. Contrast to traditional "and/or" tree, the goal graph with time constraints and dependencies of goal satisfiability, not only enriches the semantic object description, and supports to meet the objectives of the transport chain reasoning, there is a strong description and goal adaptability workflow modeling.2. A goal description method including declarative goal and procedural goal is proposed.In the application of goal-oriented migrating workflow system, declared goal describe expectations of the final state, and procedural goal describe the stage that MI migration using location services. So, those, only focus on the type of description logic approach, can not support a goal-oriented workflow modeling applications, In other words, goal description method, the coverage of declarative goal and procedural goal based Goal Description Logic, is innovative.3. A multi-goal plan optimization method based on goal combination is proposed.In the workflow description based on GDLs, the goal planning is used to guild the behavior and service utilization. When a migrating instance need to send multiple business goals, there may exist between some of the same goal plan steps, it can eliminate redundancy migration migrating instance and improve the efficiency by using goal combination.The author proposes the following future works:(1) Goal Description Logics can be combined with resource logic, inculuding the establishment of resource consumption, reuse resources, the aim of which is to improve goal-oriented workflow system reliability and robustness; (2) to establish the mechanisms of consultation, control and exception handling, increasing process of fuzzy goals to improve the system's utility and efficiency.