Applications of Genetic Algorithms and Classifier Systems tothe Analysis of Systems of Social Interaction

遗传算法和分类器系统在社会互动系统分析中的应用

• 批准号: 9223192
• 负责人: John Skvoretz
• 金额: $ 2.94万
• 依托单位: University of South Carolina at Columbia
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1993
• 资助国家: 美国
• 起止时间: 1993-03-15 至 1995-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9223192&HistoricalAwards=false
• 关键词: Applications; Genetic; Algorithms; Classifier; Systems

This research will explore the applicability of genetic algorithms and classifier systems to the analysis of social interaction, seeking to provide a formally-based understanding. The forms of stable interaction range from highly-scripted, instrumentalized social action to informal group processes sustained by the rationality of individual participants. The techniques will be applied to three domains: social exchange, collective action, and action organization. A key advantage of this research is the power of the computer techniques to represent all forms of interaction in terms of an evolving population of genotypical action plans and thus to provide the basis for an integrated theory of social interaction systems that covers the full range of commonly recognized system types. Sociology has been slow to make use of several new approaches in computer technology, and this research will be the first substantial sociological application of the related techniques of genetic algorithms and classifier systems. Genetic algorithms are a way of arriving at a solution to a problem through successive approximation, with the great advantage that it develops many potential solutions simultaneously, successively combining features of different solutions and selecting the best solutions for further development. In the context of this research, classifer systems read computerized strings of symbols, such as the units produced by genetic algorithms, and interpret them as rules for human action under specified circumstances.

这项研究将探讨遗传算法和分类器系统对分析社会互动的适用性，并寻求提供正式的理解。 稳定互动的形式范围从高度标记的，工具化的社会行动到由个别参与者的合理性维持的非正式群体过程。 这些技术将应用于三个领域：社会交流，集体行动和行动组织。 这项研究的一个关键优势是计算机技术的力量，以不断发展的基因型行动计划人群来表示各种形式的互动，从而为社会交互系统的综合理论提供了基础，该理论涵盖了全面公认的公认系统类型。 社会学运用计算机技术的几种新方法的速度很慢，这项研究将是对遗传算法和分类器系统相关技术的首次实质性社会学应用。 遗传算法是通过连续近似来达到问题解决方案的一种方式，其优势是，它同时开发了许多潜在的解决方案，连续地结合了不同解决方案的特征并选择最佳解决方案以进一步开发。 在这项研究的背景下，分类器系统读取符号的计算机字符串，例如遗传算法产生的单元，并将其解释为在特定情况下的人类行动规则。