人工生命之园

什么是人工生命？

简单教程（ what-is-alife.zip,new version,10Kb）

历史和起源

　　1987年9月21日，美国罗沙拉莫斯的人工生命研讨会中人工生命作为一门新的研究学科诞生了。C.C.Langton 是人工生命这一概念的创始人，也就是公认的“人工生命之父”。

　　其实，关于生命模拟的尝试早在70年代就已经出现，例如最初向世人展现人工生命的英国数学家约翰?康卫(John Horton Conway)开发了一个"生命游戏"软件，这个软件流行广泛，因为其具有不可抗拒的诱惑力，它象是一个演化中的微型宇宙。开始时，计算机屏幕上只出现这个宇宙的一个影象：一个平面坐标方格上布满了"活着的"黑方块和"死了的"白方块，最初的图案可以任你摆布。但一旦你开始运作这个游戏后，这些方块就会根据很少几条简单规律活过来或死过去。每一代的每一个方块首先要环顾其四周的近邻，如果近邻中就有太多活着的方块了，则这个方块的下一代就会因为数额过剩而死去。如果近邻中存活者过少，则这个方块就会因为孤独而死去。但如果其邻近中存有两到三个"活着的"方块，比例恰到好处，则这个方块的下一代就能存活下去。也就是说，要么是这一代已经活着，能够继续存活下去，如果不是这样，就会产生新的一代。虽然这些规则只是一些简单的漫画式的生物学，但是人们发现了很多有趣的图案，就象一个复杂的世界。

定义

象许多新兴学科一样，人工生命尚无统一的定义，要给人工生命下个准确的定义是困难的。不同科学或科学背景的学者对人工生命有不同的理解。这里，我们给出一些比较典型的定义：

定义1

"研究具有自然生命系统行为的人造系统。"

C. G. Langton. "Artificial Life." In C. G. Langton, editor. Artificial Life, Volume VI of SFI Studies in the Sciences of Complexity, pages 1-47, Addison-Wesley, Redwood City, CA, 1989.

定义2

"人工生命是研究怎样通过抽取生物现象中的基本动力规则来理解生命，并且在物理媒体如计算机上重建这些现象，使它们成为一种新的实验方式和受操纵。

C. G. Langton. "Preface." In C. G. Langton, C. Taylor, J. D. Farmer, and S. Rasmussen, editors, Artificial Life II, Volume X of SFI Studies in the Sciences of Complexity, pages xiii-xviii, Addison-Wesley, Redwood City, CA, 1992.

定义3
"在人工生命中的所有存在或将会存在的事物中，我们至少可以说这一领域从总体来说，代表了一种尝试，就是加重了生物学中合成理论的分量。"

C. G. Langton. "Editor's introduction." Artificial Life Journal, Volume 1, Number 1/2, pages v-viii, 1994. The MIT Press, Cambridge, MA.

定义4
"人工生命模型有足够强大的功能来获取复杂系统中更多的认知。这种方式较之自然系统更容易被操纵、重复和精确控制实验。"

C. Taylor and D. Jefferson. "Artificial life as a tool for biological inquiry." Artificial Life Journal, Volume 1, Number 1/2, pages 1-13, 1994. The MIT Press, Cambridge, MA.

目前状况

　　在这十几年来，人工生命引起了越来越多的人的关注，进行了大量的相关研究。除了计算机领域的科研工作站外，研究者还来自于其它学科领域，例如生物，经济，物理，化学等等。进行人工生命研究的著名研究所有圣塔非研究所，CalTech的Avida Artificial Life Group，意大利G.R.A.L研究所的 Research Group on Artificial Life…

　　1996年，我国举行了第一届人工生命研讨会，由中科院自动化研究所与中科院系统研究所联合举办，邀请参加的专家来自美国的SFI、MIT、加拿大、日本与中国。

研究任务

简单的说，人工生命研究有2方面：如何用计算机帮助生物学和如何用生物学帮助计算机。

可以说人工生命是一门相对新兴的混合了计算技术和生物学的高级学科。从程序员的角度上说：是运用人工智能技术作为指导，开发软件模型，模拟当今最聪明的计算机系统，以观察它的行为是否能象人类一样表现出自我学习和适应环境。对于某些人来说，是试图用计算机来模拟生物学。

相关会议

7th International Conference on Artificial Life （The Seventh International Conference on the Simulation and Synthesis of Living Systems ）Reed College, Portland, Oregon, USA .1-6 August 2000
"Looking Backward, Looking Forward"

6th International Conference on Artificial Life

4th International Conference on Artificial Life

3rd International Conference on Artificial Life

2nd International Conference on Artificial Life

1st International Conference on Artificial Life

相关刊物

Paper Journals

Artificial Life (MIT Press)

Adaptive Behavior (MIT Press)

Artificial Life and Robotics (Springer Verlag)

Advances in Complex Systems (formerly Journal of Complex Systems) (Editions Hermes)

Biological Cybernetics (Springer Verlag)

Complexity (Wiley)

Cybernetics & Systems Analysis (Plenum)

Evolution (Society for the Study of Evolution)

Evolution of Communication (John Benjamins Publishing Co.)

Evolutionary Computation (MIT Press)

IEEE Transactions on Evolutionary Computation (IEEE)

IEEE Transcations on Fuzzy Systems (IEEE)

IEEE Transactions on Neural Networks (IEEE)

IEEE Transactions on Systems, Man, and Cybernetics B: Cybernetics (IEEE)

Journal of Complexity (Academic Press)

Journal of Social and Evolutionary Systems (JAI)

Journal of Theoretical Biology (Academic)

Kybernetes (MCB)

Open Systems & Information Dynamics (Kluwer)

Online Electronic Journals

Complexity International: an electronic journal of complex systems research

Complexity Digest: an news bulletin of complex systems research

Intelligent Data Analysis-an International Journal (European mirror)

Journal of Artificial Societies and Social Simulation (full text available)

Journal of Memetics: Evolutionary Models of Information Transmission

Noetica: A Cognitive Science Forum. A WWW journal devoted to the interdisciplinary field of cognitive science (Australian mirror)

（相关刊物这一部分是从alife.org拷贝的）

How to study to become an ALife scientist? 　(from alife.org)

怎么成为人工生命的研究者

“我的经历是从计算机科学开始，然后选修了一些自适应系统和人工智能的课程，以及生物学课程。”

基础：计算机科学

“由于要做Alife的计算机实验，所以基本的编程基础是很需要的。建议学习：编程技术、计算机科学一般原理、面向对象的程序设计、数学。学习一些人工智能方面的课程，例如：AI、自适应系统、机器学习、机器人学和电子学、系统学、搏奕论、计算机图形学、分布式系统。”

加上一些生物学

“毕竟，人工生命适与生物学有关的。所以我还学习了：生态学、行为生物学、进化学、基因学、生态人口学。”

但是，这样还是不够的

... ...

A lot of people often ask me how they should choose their studies to prepare for a career in Artificial Life. I myself have done a study, of which I think, is a good preparation for further work in Artificial Life. I started with a basis in computer science, then choose some adaptive systems and artificial intelligence courses from our computer science department, and added a couple of courses from the biology faculty to complete my studies. Of course, my thesis was about Artificial Life. A complete review of my curriculum is below, and I would recommend a similar package to all students preparing for Artificial Life. 　“我的经历是从计算机科学开始，然后选修了一些自适应系统和人工智能的课程，以及生物学课程。”

The basics: Computer Science　基础：计算机科学

Artificial Life is about creating life in artificial media, with the computer on the first place. So you'll need all the basic notions of computer science to even be able to start programming on ALife experiments. If you're able to choose a package of courses at you university, make sure you have the following courses in that package:
- Programming techniques: A good introduction to computer programming. You're going to need that for preparing your computer experiments in Artificial Life.
- Formal methods in computer science: If you don't know the theoretical basics needed for computational experiments, you're nowhere in writing simulations.
- Object oriented programming: Courses in object-oriented programming offer you the ability to get good programming style. A lot of the simulations written by ALife scientists are simply better when written from an Object Oriented basis.
- Mathematics: This is maybe even more important than the computer scientific classes. make sure there's differential equations (and not just differentials) in the mathematics courses, because a lot of the basic notions of complex systems are based on these differential equations.
“由于要做Alife的计算机实验，所以基本的编程基础是很需要的。建议学习：编程技术、计算机科学一般原理、面向对象的程序设计、数学。”

Most of the computer science faculties around the world offer some Artificial Intelligence related courses. Make sure you take a couple of these too, especially look out for universities offering the following courses:
- Artificial Intelligence: Offering you the basics of Artificial Intelligence, this is sometimes combined with a couple of topics from robotics and cognitive sciences (brain, language, ...)
- Adaptive Systems: Giving you the basics of neural networks, genetic algorithms and the alike. this is a very important part in becoming an Artificial life scientist
- Machine Learning: Different formal methods in learning. Things that should be covered are bayesian learning and the alike.
- Robotics and Electronics: If you're planning to work with robots, make sure you have basic notions of robotics and electronics
- Systems theory: If your university offers a course that includes systems theory, be sure to have that one too. A good introduction to complex and dynamical systems is needed to be able to discuss several topics in Artificial Life.
- Game theory: i personnally think that game theory will have a serious impact on Artificial Life in the near future, so be sure to have some of that too. Most game theory courses offer you the first relations between biology and formal thinking.
- Computer graphics: this course is not necessary, but can give you a couple of techniques that could come in handy when programming Artificial Life projects.
- Distributed systems: this isn't necessary either, but can give you some notions on how to work with multiple machines on one simulation project, and explain to you how the agents in your simulations should communicate with each other. I found it necessary.
“学习一些人工智能方面的课程，例如：AI、自适应系统、机器学习、机器人学和电子学、系统学、搏奕论、计算机图形学、分布式系统。”

In most universities you also need to do a couple of programming projects. Try to get a project that is related to Artificial Life. Many professors don't mind that you propose your own programming projects. Try to do something difficult with genetic algorithms to get the feeling of it.

Add some biology to that 　加上一些生物学

In my university, we had to choose a couple of courses to complete our study. I also had the possibility to add courses from the biology faculty to my curriculum in the computer science faculty. If you're looking for a basis in Artificial Life, be sure to ask whether your university offers you the possibility to add a couple of courses from the biology department to your curriculum.
After all, Artificial Life is about biology...

And if you're looking for biology classes, be sure to pick out the more theoretical classes. this is the list of biology classes I added to my CS curriculum:
- Ecology: this gives you most of the definitions that could come in handy when doing research in Artificial Life.
- Behavioral biology: How do animals react on their environment? A lot of the ALife projects are based on behavioral techniques taken from biology.
- Evolution: A lot of biology is based on evolutionary thinking. So a good thorough introduction to evolution is needed to do something useful for biology. And don't you think that evolution is just about Darwin and dinosaurs, it's a bit harder than that.
- Genetics: When you're involved with genetic algorithms and evolutionary techniques in Artificial Life, a pure biological class on genetics can come in handy too!
- Population ecology: This is about how dynamics on the level of population is studied. A lot of basic notions of "emergent behavior" can be studied in the population ecology course. Plus it's a very theoretical start for doing formal biology.
“毕竟，人工生命适与生物学有关的。所以我还学习了：生态学、行为生物学、进化学、基因学、生态人口学。”

But that's not it 　但是，这样还是不够的

Doing Artificial Life is not just about studying the former list of courses, it's a lot more. You'll have to be prepared for a couple of other things...
Be sure that you want to do something with Artificial Life. It could be a very interesting new topic in science, but you'll need a lot of motivation to get there. Some of the courses (like the mathematics and formal methods) would at first be very hard to get through but you'll be needing them later.

And it's not just about studying. You'll be doing a lot of other homework too. Make sure to follow a lot of the work that's already been done in Artificial Life. Buy the proceedings of a couple of recent conferences (such as Artificial Life or ECAL or SAB conferences) and try to read as much books and papers you can find. Some books on Biology are good too. Philosophy of science books in biology (e.g. the heavy weight books by Mayr) are a good read. Maybe it's also a good idea to try and prepare a paper for a conference yourself. Publishing can give you some confidence in starting an Artificial Life career. “阅读大量文章，例如Artificial Life or ECAL or SAB 的会议论文等”

If you're really motivated, make sure that most of your studies will bring you something new about Artificial Life. If you have to do anything like projects, try to talk to the professor about what you're planning to do and they'll be happy to give you a chance in doing a project that will gain you some techniques needed in Artificial Life.

It's hard to get to the finishing line, but if you studied the courses, you'll have a good basis in doing research as an Artificial Life scientist. By the way, I did it so anyone can do it :)

一些文章

(以下1-8摘自http://ai.ia.ac.cn)

1.人工生命，周登勇、戴汝为，模式识别与人工智能. Vol. 4, 1998(3): 412-418。

2.The Control Structure of Artificial Creatures，Zhou Dengyong, Dai Ruwei，to appear in Artificial Life and Robotics，Vol.5,2000。

3.适应性行为与仿真，周登勇、戴汝为，已被《系统仿真学报》接受发表。

4.一个基于多主体的仿真工具系统-COMPLEXITY，周登勇，已被《模式识别与人工智能》接受发表。

5.智能控制与适应性，戴汝为、周登勇，第三届全球智能控制大会，2000，安徽。

6.遗传算法，董向辉，2000.3，开题报告。

7.从复杂性的观点来理解智能行为，周登勇、戴汝为，2000.6。( 请使用gsview阅读 )(新增)

8.从复杂性的观点来理解动物行为，周登勇、戴汝为，2000.6。

9.人工生命的研究,沈学华,杨献春,周志华等。南大学报，2000，11。

最后更新: