赫伯特·亚历山大·西蒙 Herbert Alexander Simon

来自集智百科 - 复杂系统|人工智能|复杂科学|复杂网络|自组织
Qige96（讨论 | 贡献）2021年5月29日 (六) 15:52的版本 （→‎人工智能）
跳到导航 跳到搜索

此词条暂由彩云小译翻译，翻译字数共3172，未经人工整理和审校，带来阅读不便，请见谅。


赫伯特·西蒙是一位美国经济学家、政治科学家、认知心理学家、和计算机科学家，他的主要研究兴趣是组织内部的决策，他最著名的理论是“有限理性（Bounded Rationality）”和“满足（Satisficing）”[1]。他分别于1975年获得图灵奖，和1978年获得诺贝尔经济学奖[2][3]。他的研究以其跨学科性质而著称，跨越了认知科学、计算机科学、公共管理、管理和政治科学等领域[4] 。从1949年到2001年，他职业生涯的大部分时间都在卡内基梅隆大学度过。[5]

值得注意的是，西蒙是几个现代科学领域的先驱之一，如人工智能、信息处理、决策、问题解决、组织理论和复杂系统。他也是最早分析复杂性结构，并提出优先连接（Preferential Attachment）机制来解释幂律分布的人之一。[6][7]

基本信息

类别 信息
姓名 赫伯特·亚历山大·西蒙
Herbert A. Simon
出生日期 1916年6月15日
出生地 美国威斯康辛州密尔沃基
所在机构 卡内基梅隆大学
主要研究方向 决策理论、认知心理、行政管理、人工智能、复杂系统等领域
主要贡献 有限理性理论、通用问题求解器、层次化复杂性等

早期生活

赫伯特·亚历山大·西蒙（Herbert Alexander Simon）于1916年6月15日出生于威斯康星州的密尔沃基。西蒙的父亲，亚瑟·西蒙 (1881-1948) ，是一位犹太电气工程师，在慕尼黑工业大学获得工程学位，1903年从德国来到美国[8]。作为一名发明家，亚瑟也是一名独立的专利律师。西蒙的母亲埃德娜 · 玛格丽特 · 默克尔(Edna Marguerite Merkel，1888-1969)是一位杰出的钢琴家，她的祖先来自布拉格和科隆。西蒙的欧洲祖先是钢琴制造商、金匠和酒商。和他的父亲一样，西蒙的母亲也来自一个有着犹太、路德教和天主教背景的家庭。与大多数孩子不同，赫伯特·西蒙的家人向他介绍了人类行为可以被科学研究的观点。通过哈罗德关于经济学和心理学的书，西蒙了解到了社会科学。在他最早的影响中，西蒙引用了诺曼 · 安吉尔的《大幻觉》和亨利 · 乔治的《进步与贫穷》[9] 。

教育经历

1933年，西蒙进入芝加哥大学学习。在他早期所受到的影响下，他决定学习社会科学和数学。西蒙对生物学很感兴趣，但由于“色盲和实验室里的笨拙” ，他选择了不从事这一领域的研究。西蒙很小的时候就知道自己是色盲，并且发现外部世界和感知的世界是不一样的。在大学期间，西蒙专注于政治科学和经济学。西蒙最重要的导师是亨利 · 舒尔茨，一位计量经济学家和数学经济学家。在报名参加了“测量市政府”的课程后，西蒙成为了克拉伦斯·里德利的研究助理，两人合著了《测量市政活动: 1938年评估行政管理建议标准调查》。这项研究引导他进入组织决策领域，也成为了他博士论文的主题。1943年，西蒙在芝加哥大学获得政治科学博士学位(Ph.D in Political Science)。

任职经历

从1942年到1949年，西蒙是一名政治学教授，同时也是芝加哥伊利诺理工学院的系主任。在那里，他开始参加考尔斯委员会工作人员举办的研讨会，这些工作人员当时包括雅各布·马沙克，和特亚林·科普曼斯。于是，他开始了在制度主义（Institutionalism）经济学的深入研究。马沙克请西蒙来协助他当时正在与山姆 · 舒尔进行的“原子能的未来经济效应”研究[10]。

1949年，西蒙成为卡内基理工学院(卡内基梅隆大学的前身)的行政学教授和工业管理系主任。西蒙后来还在同一所大学教授心理学和计算机科学。从1949年到2001年，西蒙都在在卡内基梅隆大学任教，当然有时候也会访问其他大学和科研机构[11]。

为了取代高度简化的经典经济模型方法，西蒙在他的著作《管理行为（Administrative Behavior）》中以提出了他著名的公司决策理论（Theory of Corporate Decision）。在这本书中，他提炼出了一系列概念，以及一套可以识别出多种影响决策制定因素的方法。他在组织和行政方面的兴趣使他不仅三次担任大学系主任，而且在1948年经济合作署的成立中发挥了重要作用，也参与了马歇尔计划中管理援助的行政团队，还在林登·约翰逊总统的科学咨询委员会和国家科学院任职[10]。

学术贡献

决策

Administrative Behavior,[12] first published in 1947 and updated across the years, was based on Simon's doctoral dissertation.[13] It served as the foundation for his life's work. The centerpiece of this book is the behavioral and cognitive processes of humans making rational decisions. By his definition, an operational administrative decision should be correct, efficient, and practical to implement with a set of coordinated means.[13]

Simon recognized that a theory of administration is largely a theory of human decision making, and as such must be based on both economics and on psychology. He states:

[If] there were no limits to human rationality administrative theory would be barren. It would consist of the single precept: Always select that alternative, among those available, which will lead to the most complete achievement of your goals.[13] (p xxviii)

Contrary to the "homo economicus" stereotype, Simon argued that alternatives and consequences may be partly known, and means and ends imperfectly differentiated, incompletely related, or poorly detailed.[13]

Simon defined the task of rational decision making is to select the alternative that results in the more preferred set of all the possible consequences. Correctness of administrative decisions was thus measured by:

  • Adequacy of achieving the desired objective
  • Efficiency with which the result was obtained

The task of choice was divided into three required steps:[14]

  • Identifying and listing all the alternatives
  • Determining all consequences resulting from each of the alternatives;
  • Comparing the accuracy and efficiency of each of these sets of consequences

Any given individual or organization attempting to implement this model in a real situation would be unable to comply with the three requirements. Simon argued that knowledge of all alternatives, or all consequences that follow from each alternative is impossible in many realistic cases.[12]

Simon attempted to determine the techniques and/or behavioral processes that a person or organization could bring to bear to achieve approximately the best result given limits on rational decision making.[13] Simon writes:

The human being striving for rationality and restricted within the limits of his knowledge has developed some working procedures that partially overcome these difficulties. These procedures consist in assuming that he can isolate from the rest of the world a closed system containing a limited number of variables and a limited range of consequences.[15]

Therefore, Simon describes work in terms of an economic framework, conditioned on human cognitive limitations: Economic man and Administrative man.

Administrative Behavior addresses a wide range of human behaviors, cognitive abilities, management techniques, personnel policies, training goals and procedures, specialized roles, criteria for evaluation of accuracy and efficiency, and all of the ramifications of communication processes. Simon is particularly interested in how these factors influence the making of decisions, both directly and indirectly.[citation needed]

Simon argued that the two outcomes of a choice require monitoring and that many members of the organization would be expected to focus on adequacy, but that administrative management must pay particular attention to the efficiency with which the desired result was obtained.[citation needed]

Simon followed Chester Barnard, who stated "the decisions that an individual makes as a member of an organization are quite distinct from his personal decisions".[16] Personal choices may be determined whether an individual joins a particular organization and continue to be made in his or her extra–organizational private life. As a member of an organization, however, that individual makes decisions not in relationship to personal needs and results, but in an impersonal sense as part of the organizational intent, purpose, and effect. Organizational inducements, rewards, and sanctions are all designed to form, strengthen, and maintain this identification.[citation needed]

Simon[13] saw two universal elements of human social behavior as key to creating the possibility of organizational behavior in human individuals: Authority (addressed in Chapter VII—The Role of Authority) and in Loyalties and Identification (Addressed in Chapter X: Loyalties, and Organizational Identification).

Authority is a well-studied, primary mark of organizational behavior, straightforwardly defined in the organizational context as the ability and right of an individual of higher rank to guide the decisions of an individual of lower rank. The actions, attitudes, and relationships of the dominant and subordinate individuals constitute components of role behavior that may vary widely in form, style, and content, but do not vary in the expectation of obedience by the one of superior status, and willingness to obey from the subordinate.[citation needed]

Loyalty was defined by Simon as the "process whereby the individual substitutes organizational objectives (service objectives or conservation objectives) for his own aims as the value-indices which determine his organizational decisions".[17] This entailed evaluating alternative choices in terms of their consequences for the group rather than only for oneself or one's family.[18]

Decisions can be complex admixtures of facts and values. Information about facts, especially empirically-proven facts or facts derived from specialized experience, are more easily transmitted in the exercise of authority than are the expressions of values. Simon is primarily interested in seeking identification of the individual employee with the organizational goals and values. Following Lasswell,[19] he states that "a person identifies himself with a group when, in making a decision, he evaluates the several alternatives of choice in terms of their consequences for the specified group".[20] A person may identify himself with any number of social, geographic, economic, racial, religious, familial, educational, gender, political, and sports groups. Indeed, the number and variety are unlimited. The fundamental problem for organizations is to recognize that personal and group identifications may either facilitate or obstruct correct decision making for the organization. A specific organization has to determine deliberately, and specify in appropriate detail and clear language, its own goals, objectives, means, ends, and values.[citation needed]

Simon has been critical of traditional economics' elementary understanding of decision-making, and argues it "is too quick to build an idealistic, unrealistic picture of the decision-making process and then prescribe on the basis of such unrealistic picture".[21]

人工智能

西蒙的一些经济学研究主要是为了理解技术变革，特别是信息处理革命。

西蒙是人工智能领域的先驱，他与艾伦·纽厄尔合作开发了逻辑理论机器(Logic Theory Machine)(1956)和通用问题求解器(General Problem Solver)(1957)程序。通用问题求解器可能是第一种将问题解决策略与特定问题的信息分离开来的方法。这两个程序都是使用由纽厄尔，克里夫肖和西蒙发明的信息处理语言(Information Processing Language)(1956年)来开发的。1957年，西蒙预测计算机国际象棋将在“十年内”超过人类的国际象棋能力，而实际上，这种转变需要大约四十年。[22]

20世纪60年代早期，心理学家乌尔里克·内瑟断言，虽然机器能够复制“冷认知”行为，如推理、计划、感知和决策，但它们永远不能复制“热认知”行为，如疼痛、快乐、欲望和其他情绪[23]。西蒙在1963年写了一篇关于情感认知的论文，回应了内瑟的观点。他在1967年对论文进行了更新，并发表在《心理学评论》上[24] 。几年来，西蒙在情绪认知方面的工作在很大程度上被人工智能研究界所忽视，但是斯洛曼和皮卡德后来在情绪方面的工作帮助人们重新关注西蒙的论文，并最终使其对这一主题产生了很大影响。

西蒙与与艾伦·纽厄尔一起合作提出了一个理论，使用产成式规则模拟人类解决问题的行为[25] 。研究人类如何解决问题需要一种新的测量方法，西蒙和安德斯·埃里克森一起开发了口头协议分析（Verbal Protocol Analysis）的实验方法。[26]西蒙对知识在专业技能中的作用很感兴趣。他认为，要成为一个领域的专家需要大约十年的经验。他和同事估计，专业知识是学习大约50,000块（Chunk）信息的结果。据说，一位国际象棋专家已经掌握了大约50,000个棋子或棋位模式[27]。

心理学

西蒙对人类如何学习很感兴趣，且在爱德华·费根鲍姆的帮助下，他提出了最早的把学习作为计算机程序来实现的理论之一——初级感知者和记忆者（Elementary Perceiver And Memorizer，EMPA）理论，能够解释言语学习领域中的大量现象[28] 。这个模型的后来版本被应用于研究概念形成和专业知识的获取。在弗尔南多·戈贝特的帮助下，他将 EPAM 理论扩展到了 CHREST 计算模型[29]。这个理论解释了简单的信息块是如何形成图式结构，一种更复杂的结构。CHREST模型被成功用于模拟国际象棋专业技能的各个方面。

社会学与经济学

Simon has been credited for revolutionary changes in microeconomics. He is responsible for the concept of organizational decision-making as it is known today. He also was the first to discuss this concept in terms of uncertainty; i.e., it is impossible to have perfect and complete information at any given time to make a decision. While this notion was not entirely new, Simon is best known for its origination. It was in this area that he was awarded the Nobel Prize in 1978.[30]

At the Cowles Commission, Simon's main goal was to link economic theory to mathematics and statistics. His main contributions were to the fields of general equilibrium and econometrics. He was greatly influenced by the marginalist debate that began in the 1930s. The popular work of the time argued that it was not apparent empirically that entrepreneurs needed to follow the marginalist principles of profit-maximization/cost-minimization in running organizations. The argument went on to note that profit maximization was not accomplished, in part, because of the lack of complete information. In decision-making, Simon believed that agents face uncertainty about the future and costs in acquiring information in the present. These factors limit the extent to which agents may make a fully rational decision, thus they possess only "bounded rationality" and must make decisions by "satisficing", or choosing that which might not be optimal, but which will make them happy enough. Bounded rationality is a central theme in behavioral economics. It is concerned with the ways in which the actual decision making process influences decision. Theories of bounded rationality relax one or more assumptions of standard expected utility theory.

Further, Simon emphasized that psychologists invoke a "procedural" definition of rationality, whereas economists employ a "substantive" definition. Gustavos Barros argued that the procedural rationality concept does not have a significant presence in the economics field and has never had nearly as much weight as the concept of bounded rationality.[31] However, in an earlier article, Bhargava (1997) noted the importance of Simon's arguments and emphasized that there are several applications of the "procedural" definition of rationality in econometric analyses of data on health. In particular, economists should employ "auxiliary assumptions" that reflect the knowledge in the relevant biomedical fields, and guide the specification of econometric models for health outcomes.

Simon was also known for his research on industrial organization.[32][33] He determined that the internal organization of firms and the external business decisions thereof, did not conform to the neoclassical theories of "rational" decision-making. Simon wrote many articles on the topic over the course of his life, mainly focusing on the issue of decision-making within the behavior of what he termed "bounded rationality". "Rational behavior, in economics, means that individuals maximize their utility function under the constraints they face (e.g., their budget constraint, limited choices, ...) in pursuit of their self-interest. This is reflected in the theory of subjective expected utility. The term, bounded rationality, is used to designate rational choice that takes into account the cognitive limitations of both knowledge and cognitive capacity. Bounded rationality is a central theme in behavioral economics. It is concerned with the ways in which the actual decision-making process influences decisions. Theories of bounded rationality relax one or more assumptions of standard expected utility theory".模板:Cite quote

Simon determined that the best way to study these areas was through computer simulations. As such, he developed an interest in computer science. Simon's main interests in computer science were in artificial intelligence, human–computer interaction, principles of the organization of humans and machines as information processing systems, the use of computers to study (by modeling) philosophical problems of the nature of intelligence and of epistemology, and the social implications of computer technology.[citation needed]

In his youth, Simon took an interest in land economics and Georgism, an idea known at the time as "single tax".[34] The system is meant to redistribute unearned economic rent to the public and improve land use. In 1979, Simon still maintained these ideas and argued that land value tax should replace taxes on wages.[35]

Some of Simon's economic research was directed toward understanding technological change in general and the information processing revolution in particular.[citation needed]

奖项和荣誉

西蒙一生获奖诸多，包括但不限于：

  • 1959年美国艺术与科学学会院士 [36]
  • 1967年入选国家科学院院士[37]
  • 1969年APA杰出科学贡献奖
  • 1975年ACM图灵奖“为人工智能、人类认知心理学和列表处理做出基本性的贡献”
  • 1978年诺贝尔经济学奖“在经济组织的决策过程领域做出了先驱式的研究”(1978)
  • 1986年国家科学奖章
  • 1993年APA终身成就奖
  • 1995年IJCAI卓越研究奖

精选刊物

赫伯特·西蒙是一个非常高产的写作者。他参与撰写了27本书，和上千篇文章。截至2016年，西蒙是Google Scholar中认知心理学领域获得最多引用的学者。毫无疑问，赫伯特·西蒙是20世纪科学界最有影响力的人物之一。

书籍

– 4th ed. in 1997, The Free Press
  • 1957. Models of Man. John Wiley. Presents mathematical models of human behaviour.
  • 1958 (with James G. March and the collaboration of Harold Guetzkow). Organizations. New York: Wiley. the foundation of modern organization theory
  • 1969. The Sciences of the Artificial. MIT Press, Cambridge, Mass, 1st edition. Made the idea easy to grasp: "objects (real or symbolic) in the environment of the decision-maker influence choice as much as the intrinsic information-processing capabilities of the decision-maker"; Explained "the principles of modeling complex systems, particularly the human information-processing system that we call the mind."
- 2nd ed. in 1981, MIT Press. As stated in the Preface, the second edition provided the author an opportunity "to amend and expand [his] thesis and to apply it to several additional fields" beyond organization theory, economics, management science, and psychology that were covered in the previous edition.
- 3rd ed. in 1996, MIT Press.
  • 1972 (with Allen Newell). Human Problem Solving. Prentice Hall, Englewood Cliffs, NJ, (1972). "the most important book on the scientific study of human thinking in the 20th century"
  • 1977. Models of Discovery : and other topics in the methods of science. Dordrecht, Holland: Reidel.
  • 1979. Models of Thought, Vols. 1 and 2. Yale University Press. His papers on human information-processing and problem-solving.
  • 1982. Models of Bounded Rationality, Vols. 1 and 2. MIT Press. His papers on economics.
- Vol. 3. in 1997, MIT Press. His papers on economics since the publication of Vols. 1 and 2 in 1982. The papers grouped under the category "The Structure of Complex Systems"– dealing with issues such as causal ordering, decomposability, aggregation of variables, model abstraction– are of general interest in systems modelling, not just in economics.
  • 1983. Reason in Human Affairs, Stanford University Press. A readable 115pp. book on human decision-making and information processing, based on lectures he gave at Stanford in 1982. A popular presentation of his technical work.
  • 1987 (with P. Langley, G. Bradshaw, and J. Zytkow). Scientific Discovery: computational explorations of the creative processes. MIT Press.
  • 1991. Models of My Life. Basic Books, Sloan Foundation Series. His autobiography.
  • 1997. An Empirically Based Microeconomics. Cambridge University Press. A compact and readable summary of his criticisms of conventional "axiomatic" microeconomics, based on a lecture series.
取自“https://wiki.swarma.org/index.php?title=赫伯特·亚历山大·西蒙_Herbert_Alexander_Simon&oldid=22636”