Abstract:

In business intelligence, Agent-based automatic negotiation, which uses various artificial intelligence advantages of Agent to simulate people’s actual business negotiation, has been paid more and more attention, among which multi-attribute decision making is particularly important. In view of the current situation of insufficient research on the weight and perceived value, firstly, it uses the hesitating fuzzy number and makes a new attribute classification. After establishing the corresponding hesitating fuzzy evaluation matrix and the evaluation value specification, it constructs the corresponding subjective weight algorithm. Secondly, combining the objective optimization model and Lagrange function, the corresponding objective weight algorithm is constructed, and then an improved comprehensive weight calculation method is proposed. Then based on prospect theory, it introduces loss aversion factor, proposes to the positive and negative ideal point as a reference point, sets up corresponding algorithm property and the distance from the positive and negative ideal solution is calculated, and uses it as a new parameters to perceived value function, and then puts forward the comprehensive perceived value function based on improved overall dominance algorithm, finally constructs the multiple attribute decision making model based on Agent. Finally, taking the purchase negotiation of experimental teaching equipment in a university as an example, the sensitivity analysis and comparative analysis with relevant research results verify that the model can help Agent make more rapid, reasonable and effective decisions.

Key words: Agent, multi-attribute decision making, comprehensive weight, perceived value, prospect theory, experimental teaching

摘要：

商务智能中，基于Agent的自动谈判利用Agent的各项人工智能优势模拟人们进行实际商务谈判，日益受到重视，其中的多属性决策尤为重要。针对现有研究对其中权重及感知价值研究不够的现状，采用犹豫模糊数，给出新的属性分类，建立相应的犹豫模糊评价矩阵并进行评价值规范后构建相应主观权重算法；结合目标优化模型和拉格朗日函数，构建相应客观权重算法，进而提出改进的综合权重计算法；在前景理论基础上，引入损失规避因子，提出将正负理想点作为双参考点，设定相应算法计算各属性与正负理想解的距离，并将其作为新参数加入感知价值函数，从而提出基于改进综合感知价值函数的总体优势度算法，最终构建出基于Agent的多属性决策模型；以某高校实验教学设备采购谈判为例，通过敏感性分析和与相关研究结果的比较分析，验证了该模型能帮助Agent做出更快速合理有效的决策。

关键词: Agent, 多属性决策, 综合权重, 感知价值, 前景理论, 实验教学