时变论域下的词计算与语言动力系统理论及应用

Many problems of complex systems are closely related to human actions, and which always manifest as perceptual information，and the analysis and process of these problems are mainly by the means of natural language，then to deal with these perceptual information by using natural language is the keypoint to solve these matters. Therefore, linguistic dynamic theory was developed to manage the prolems, which have been applied in management systems and electric systemsm，and type-2 fuzzy theory has involved. Based on the early works, the project will conduct the the theories of computing with words and linguistic dynamic systems based on time-varying universe discourse, including the following works: 1) the mechanism for variation of universe changing with time variation; fuzzy sets and their operations based on time-varying universe of discourse; dynamic fuzzy rules; the theory and menthod of computing with words and linguistic dynamic systems based on time-varying universe; 2)The key technologies, including adaptive dynamic programming, fuzzy control, multi-objective optimization and parallel control technology. The research of the project is to mimic the human ability of manipulating language to discribe the evolutionary mechanism and dynamic developing process by means of natuarl language, dynamically and effectively take advantage of information on the language level to solve the problems of modeling,analysis,control,prediction and evaluation for complex systems, and the results of the project will push forward the reseach of complex systems.

复杂系统中的诸多问题与人类行为密切相关，这类问题往往表现为感知信息，其分析和处理主要采用自然语言的方式，因此运用自然语言来处理感知信息成为解决该问题的关键。为此，本项目提出了解决复杂系统相关问题的语言动力学理论体系，已经在管理系统的评估中得到初步实践。在前期工作的基础上，本项目拟研究基于时变论域的词计算与语言动力学建模与分析方法，包括论域随时间变化的机理，时变论域上的模糊集合及其运算，动态模糊规则的建立，基于时变论域的词计算理论与方法，时变论域下的语言动力学理论；对应的关键技术体系，包括自适应动态规划，模糊控制，多目标优化及平行控制技术等。本项目通过模拟人类使用文字语言的能力，以自然语言的方式描述来复杂系统演化机制与动态发展过程，在语言的层次上动态有效利用信息来解决复杂系统建模、分析、控制、预测和评估问题, 将复杂系统研究推向新阶段。

与人类行为相关的问题主要表现为感知信息，其分析与处理主要是采用自然语言的方式进行的，因此如何运用自然语言 来处理感知信息成为解决这类问题的关键。本项目提出了基于时变论域的词计算与语言动力学建模与分析方法，包括时变论域上的模糊集合及运算，动力模糊规则，基于时变论域的词计算理论与方法，时变论域下的语言动力系统理论，及二型模糊集合与逻辑，二型模糊综合评判等方法。同时将相应的理论与方法应用于平行红绿灯的动态配时设计，主干道的交通流预测，停车场停车位的选取及停车收费策略，员工胜任力的综合评判以及心理健康状态的评估及心理危机的干预等。本项目通过模拟人类使用文字语言的能力，以自然语言的方式来描述系统演化机制与动态发展过程，在语言的层次上动态有效利用信息来解决复杂系统建模、分析、控制、预测和评估问题。自然语言的精确处理被誉为人工智能王冠上的一个明珠，模糊集合通过给定的精确隶属函数为实现自然语言的精确理解奠定了基础。