工作流管理：模型、方法与系统详解

需积分: 9 142 浏览量更新于2024-07-28 收藏 9.55MB PDF 举报

"Workflow Management_ Models, Methods and Systems" 是一本由Wil van der Aalst和Kees van Hee合著的书籍，主要探讨了工作流管理的模型、方法和系统，质量被高度评价。工作流管理是信息化系统中的一个重要概念，它涉及到组织内任务的自动化协调和控制，以优化业务流程。本书深入研究了这一领域，为读者提供了全面的理解框架。作者们通过丰富的理论和实践案例，详细阐述了工作流管理的核心原理。首先，书中的“模型”部分可能涵盖了各种类型的工作流模型，如活动网络图（Activity Network Diagrams）、Petri网、业务流程建模语言（Business Process Modeling Notation, BPMN）等。这些模型用于描绘业务流程的结构和动态，帮助管理者识别瓶颈，改进效率，并为自动化提供基础。其次，“方法”部分可能涉及工作流设计、分析和优化的方法。这包括流程映射、流程挖掘、流程改进技术（如六西格玛和精益管理），以及如何利用信息技术实现自动化。此外，书中可能还讨论了如何通过模拟和实验来评估和调整流程设计。接着，“系统”部分则可能涵盖了工作流管理系统（Workflow Management System, WfMS）的设计与实现。WfMS是一套支持工作流自动化和管理的软件工具，包括流程定义、任务分配、状态跟踪、报告生成等功能。作者可能会讨论不同类型的WfMS架构，如基于规则的系统、事件驱动的系统，以及云和分布式环境下的工作流解决方案。书中可能还包含了对合作信息系统（Cooperative Information Systems）的讨论，强调在协同工作环境中工作流的重要性。此外，书中的参考文献和索引为读者提供了进一步研究的资源。这本书不仅适合IT专业人士，也适合企业经理和业务分析师，因为它提供了理解、设计和实施工作流管理的强大工具和理论基础。通过阅读本书，读者可以学习如何利用工作流管理提升业务效率，优化组织流程，以及应对快速变化的商业环境。

4 Chapter 1

Each case involves a process being performed. A process consists of a

number of tasks that need to be carried out and a set of conditions that

determine the order of the tasks. A process can also be called a proce-

dure. A task is a logical unit of work that is carried out as a single whole

by one resource. A resource is the generic name for a person, machine or

group of persons or machines that can perform specific tasks. This does

not always mean to say that the resource necessarily carries out the task

independently, but that it is responsible for it. We will examine this sub-

ject more closely in the next section.

As an example of a process, we shall examine how a (fictional) insur-

ance company deals with a claim. We can identify the following tasks:

1. recording the receipt of the claim;

2. establishing the type of claim (for example, fire, motor vehicle, travel,

professional);

3. checking the client's policy, to confirm that it does in principle cover

what has been claimed for;

4. checking the premium, to confirm that payments are up to date;

5. rejection, if task 3 or 4 has a negative result;

6. producing a rejection letter,

estimating the amount to be paid, based upon the claim details;

8. appointment of an assessor to research the circumstances of the dam

age and to establish its value;

9. consideration of emergency measures to limit further damage or re

lieve distress;

10. provision of emergency measures if approved as part of task 8;

11. establishment or revision of amount to be paid and offer to client;

12. recording of client's reaction: acceptance or objection;

13. assessment of objection and decision to revise (task 11) or to take

legal proceedings (task 14);

14. legal proceedings-,

15. payment of claim; and

16. closure of claim: filing.

Here we can see sixteen tasks that do not necessarily need to be per-

formed in the order shown. Two or more tasks that must be performed

in a strict order are called a sequence. For some cases, certain tasks do

not need to be carried out. One example is the appointment of an expert,

if the claim report is clear and the amount of the claim is below a par-

Organizing Workflows 5

ticular value, the involvement of an expert is not necessary. Other tasks

that do not always need to be performed are taking emergency measures,

assessing an objection, or taking legal proceedings. Sometimes, therefore, a

choice between two or more tasks can be made. This we call a selection.

There are also tasks that can be performed in parallel, for exam-

ple checking the policy and checking the premiums. These tasks must

both be completed before the "rejection" task can begin. This is called

synchronization.

This example of a process also includes iteration, or repetition—

namely, the repeated assessment of an objection or the revision of the

amount to be paid. In theory, this could go on forever. Figure 1.2 shows

the order of the tasks as a process diagram: an arrow from task A to task

B means that A must be done before B. We can also see that the diagram

contains more information than the list of tasks. For example, it shows

that a claim can only be closed once any emergency measures required

have been taken. Each task is indicated by a rectangle. If a task has more

than one successor task—that is, if it has more than one arrow leading

from it—then precisely one of these subsequent tasks must be chosen

during the task in question. If a task has more than one predecessor—

more than one arrow leading to it—then all of these must be completed

before that task can begin (synchronization). The circles indicate where

particular workflows meet or split. The gray circles have several precur-

sor tasks and only one subsequent task. They indicate that only one of

the preceding tasks needs to be performed in order to continue. The

black circles have one predecessor and several subsequent tasks. They

show that all the subsequent tasks must be performed. (The circles can be

regarded as "dummy" tasks.) Chapter 2 introduces a process notation

which makes it easier to express such properties.

To summarize, we can identify four different basic mechanisms in

process structures: sequence, selection, parallelization, and iteration. All

are very commonplace in practice, and in principle all processes can

be modeled using these four constructions. We shall consider them in

greater detail in chapter 2.

Some tasks can be performed by a computer without human interfer-

ence. Other tasks require human intelligence: a judgment or a decision.

For instance, a bank employee decides if a client's loan request will be

granted or not. Human workers need knowledge to execute tasks. This

Organizing Workflows 7

knowledge is stored in their minds by experience, the so-called tacit

knowledge. Other forms of knowledge can be obtained by learning and

information retrieval, the so-called explicit knowledge. Knowledge man-

agement is concerned with the acquisition, enrichment, and distribution

of knowledge so that the right knowledge is at the right time with the

person who has to fulfill a task.

A task can also be defined as a process that cannot be subdivided any

further: an atomic process. There is a subjective element in this—what

one person regards as a single task may be a nonatomic one to another.

For an insurance company, for example, the compilation of an assessor's

report of damage to a car is a single task, whereas for the expert himself

it is a process comprising various tasks, such as checking the chassis, en-

gine, and bodywork. A task is therefore an atomic process for the person

defining or ordering it, but for the person carrying it out it is often a

nonatomic one.

A single process is carried out on each case. We call the performance of

a task by a resource an activity. Various cases may have the same pro-

cess, but each case may follow a different route through that process. In

the insurance company, for example, one claim may involve an objection

and another not. The route taken depends upon the specific character-

istics of the case—the case attributes. The number of processes in a

company is (generally) finite and far smaller than the number of cases to

be handled. As a result, a company can develop a routine for performing

processes and thus operate efficiently.

This is clearly seen in the clothing industry: it is much faster to make

one hundred skirts with the same pattern than one hundred skirts using

different patterns. Off-the-rack is cheaper than made-to-measure. What's

more, producing one thousand skirts of the same pattern is less expensive

than ten times making one hundred in that pattern. This is called the

economy of scale: the costs per case fall as the number of cases increases.

Companies therefore endeavor to keep the number of processes small

and to make the number of cases that each can perform as high as pos-

sible—at least, as long as they can earn something from each case. Profit,

after all, is the ultimate objective.

An insurance company wants to keep the number of claims as low

as possible—but this is generally a factor that it cannot control. It will

also try to keep the number of processes low. There is, however, a catch:

剩余374页未读，继续阅读

jianzhanger

粉丝: 87
资源: 58

工作流管理：模型、方法与系统详解

workflow_springbootworkflow_workflowdemo_workflowspringBoot_work

泛微流程引擎表结构

如果数据量很大，下述sql应该把哪些字段建立为索引 ? sql: select workflow_order_id from workflow_record where handle_result = 1 and workflow_id = 1470275488528011266

来一段odoo审批流的核心代码

最新资源