工作流管理：模型与方法深入探索

需积分: 16 186 浏览量更新于2024-07-29 收藏 9.71MB PDF 举报

"工作流管理：模型、方法和系统"是一本由Wil van der Aalst和Kees van Hee合著的专业书籍，专注于探讨工作流的理论与实践。该书由MIT Press出版，最初在1997年以荷兰语发布，2002年被翻译成英文版。书中深入探讨了工作流管理的各个方面，包括模型、方法和系统，为理解和应用工作流提供了全面的理论基础。工作流管理是信息技术领域的一个关键概念，它涉及组织内任务和过程的自动化与协调。本书首先介绍了工作流的基本概念，解释了工作流管理系统如何支持企业运营中的业务流程。作者通过详细讨论不同的工作流模型，帮助读者理解工作流设计的各种可能性，如顺序、分支、并行和循环等结构。书中涵盖了工作流建模的方法，如Petri网、活动网络图（Activity Network Diagrams）以及业务流程建模语言（Business Process Modeling Notation, BPMN）。这些建模工具和技术有助于将复杂的业务逻辑转化为可视化模型，便于分析和优化。此外，书中还讨论了如何使用这些方法来分析和改进现有的工作流程，以提高效率和响应速度。在系统方面，作者探讨了工作流管理系统的设计与实现，包括系统架构、流程执行引擎、用户界面和与其他系统的集成。他们强调了适应性、可伸缩性和安全性等关键特性，并介绍了实际案例研究来展示工作流管理系统的应用效果。除了理论内容，本书还包括了参考文献和索引，为读者提供了进一步学习和研究的资源。无论对学术研究者还是对企业实践者，这本书都是一份宝贵的参考资料，可以帮助他们在工作流管理领域深化理解，提升专业能力。 "工作流管理：模型、方法和系统"是一本深入且全面的著作，对于想要掌握工作流管理核心原理的人来说，是一本不可或缺的读物。尽管原文为英文，但其丰富的信息和实用的价值使其值得挑战，尤其对于那些在IT行业从事流程自动化、系统集成或业务流程优化工作的专业人士来说。

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

wtsky

粉丝: 2
资源: 8

工作流管理：模型与方法深入探索

[工作流管理技术基础].范玉顺.文字版.pdf

工作流管理技术基础：概念、系统与发展趋势

[信息系统开发与项目管理].章宁.文字版.pdf

Node Cookbook英文文字版pdf和随书源码

Data_analysis_study：资料库和模型库的资料库

UNIX环境高级编程 英文第三版 文字版

2008年下半年系统分析师试题(PDF文字版)

并行编程模型：分布存储系统与MPI、HPF简介

数据库系统概念：面向对象模型与新型数据库探索

数据库系统概念：解析关系数据模型与实现技术

最新资源

UNIX环境高级编程英文第三版文字版