基于行为决策理论的隐性目标决策模型与方法研究

Tacit object decision-making problems are a class of complex decision-making problems in the real world. It is very complex to solve the problems because it has three special characters that the decision objectives are unable or difficult to be defined explicitly in a structured or quantitative way, the decision processes are interacting constantly, and the decision results are influenced by the changed decision scene and decision-maker's behaviors. Based on the previous research, we apply behavioral decision-making and artificial intelligence methods to study the decision-maker's preference rules, and preference model constructions and solving in the tacit object decision-making problems. The specific contents of this project include: Applying behavior decision-making and affective computing methods to study the decision-maker's decision cognitive laws and behavior models on the tacit object decision-making problems, and to build the tacit object decision behavior characteristic model; Studying the decision-maker's psychological space and preference model, the aggregation methods of group psychological space and preference, using interactive artificial intelligence methods to construct the solving model of the tacit object decision-making problems; Studying the interactive aggregation mechanisms and sharing methods in the process of the tacit object decision-making problems, and studying the interactive group decision-making intelligence algorithm based on the methods of multi-population co-evolutionary, adaptive mechanism and intelligence; Discussing the noise suppression methods in the decision-making process and the strategies of enhancing decision-making efficiency; Carrying out experiments and applying our project in some real-world domains. This research can better meet the decision demands of decision-maker, and has theoretical significance and application value.

"隐性目标决策问题"是现实世界中大量存在着一类决策目标难以数量化表征、决策过程需不断交互迭代、决策结果受情景和行为影响的复杂决策问题。在前期研究工作基础上，应用行为决策、人工智能等理论和方法，深入研究隐性目标决策问题中有关决策者行为偏好规律和建模求解方法。具体内容包括：应用行为决策、情感计算等，研究决策者对隐性目标决策问题的决策认知规律和行为模式，分析建立隐性目标决策行为特征模型；研究决策者心理空间与偏好模型，研究群体心理空间与偏好的集成方法，应用交互式人工智能方法构建隐性目标决策问题求解模型；研究隐性目标决策过程的交互集成机制和共享方法，结合多种群协同进化、自适应机制和智能自学习方法等研究交互式群决策智能求解算法；研究决策过程中的噪声抑制方法和提高决策效率策略；并结合具体领域，进行实验和应用研究。本项目研究更加贴近现实决策需求，具有重要的理论意义和应用价值。

“隐性目标决策问题” 是现实世界中大量存在着一类决策目标难以完全数量化、结构化表征，决策过程需不断交互迭代、决策结果受决策情景和决策者行为偏好影响，具有NP难性质的复杂决策问题。本项目结合现实世界实际需求，应用行为决策、多目标/多属性决策、智能决策支持系统、交互式进化计算、人工智能等理论和方法，分析了决策者对隐性目标决策问题的决策认知规律和行为模式，构建了相应的决策行为模型；分析构建了基于决策者偏好的个体、群体隐性目标决策模型和方法；提出了隐性目标决策环境下的多种信息集结方法；讨论了决策交互过程中群体偏好一致性协调机制；提出了多种基于交互式进化计算的建隐性目标决策求解方法，探讨了降低决策者疲劳、噪声抑制和消减等方法和策略；并结合具体领域，进行了原型系统实验和实际应用，取得很好的应用效果。.本项目发表学术论文47篇，其中被SCI/EI检索18篇；培养出博士研究生12名，硕士研究生9名；设计开发出原型系统并应用到实际项目中，获得安徽省科学技术奖二等奖1项。