Abstract: Aiming at the problem that uncertainty, group consensus and irrational psychological factors in the expert evaluation process make it difficult to evaluate existing agricultural product intelligent grading equipment, an IVFFN—CNTR—PT group decision‑making method combining Complex Network of Trust Relationship (CNRT) and Prospect Theory (PT) under interval Fermatean fuzzy(IVFFN) environment was proposed. Based on the characteristics of the evaluation of agricultural product intelligent grading equipment, a two‑stage evaluation model was established based on the evaluation of the maturity of intelligent grading as a prerequisite and then a comprehensive performance evaluation. Firstly, based on the actual production and operation of the equipment, a performance evaluation indicator system was established. Secondly, the manufacturing maturity, complexity and multi‑disciplinarity of the intelligent classification equipment for agricultural products made the evaluation uncertain and  the group opinions difficult to reach consensus. IVFFN was used to fully express the uncertainty of expert evaluation information. The Dombi operator aggregated expert evaluation information, and an improved personalized feedback adjustment mechanism based on CNTR was proposed to solve the group consensus problem, then a preliminary evaluation of the equipment's intelligent classification maturity based on group consensus was conducted. After that, a comprehensive evaluation of the preliminary evaluation equipment was conducted, and at the same time, in order to solve the irrational psychology caused by experts' psychological preferences and related interests, the correlation coefficient entropy method and the optimum order method based on game theory were proposed to obtain the combined weight of indicator information, and the prospect theory was used to simulate and analyze the experts' psychological factors for decision‑making. Finally, the comprehensive performance evaluation of five different flue‑cured tobacco intelligent graders was used as a case study to verify. The results showed that in the first stage of equipment intelligence classification maturity, the level of the first four manufacturers was Ⅱ, and the fifth manufacturer was level I, with scores of 0.613 0(Ⅱ), 0.533 4(Ⅱ), 0.536 9(Ⅱ), and 0.528 5(Ⅱ), 0.421 1(Ⅰ), the comprehensive performance evaluation values of the second stage were 1.691 5, 0.922 4, 0.972 9, 1.012 7 and 0, the first manufacturer's equipment was the best. And the feasibility and superiority of this method were verified through comparative analysis of sensitivity and methods.

Key words: agricultural product intelligent grading equipment, uncertainty, group consensus, irrational psychology,  , two?stage

摘要： 针对专家评价过程中的不确定性、群体共识和非理性心理因素导致现有农产品智能分级设备难以评价的问题，提出区间Fermatean(IVFFN)模糊环境下结合复杂信任关系网络(CNRT)与前景理论(PT)的IVFFN—CNTR—PT群决策方法，并针对农产品智能分级设备评价的特点，以智能分级成熟度评价为先决条件，再进行综合性能评价，建立两阶段评价模型。首先，根据农产品智能分级设备的实际生产运作情况，建立性能评价指标体系；其次，农产品智能分级设备的制造成熟度、复杂性和多学科性，使得评价存在不确定性和群体意见难以共识问题，利用IVFFN充分表达专家评价信息的不确定性，Dombi算子聚合专家评价信息，提出基于CNTR改进的个性化反馈调整机制进行群体共识，并基于群体共识对设备智能分级成熟度进行初评；然后，对初评设备进行综合评价，同时为解决专家心理偏好和相关利益性产生的非理性心理，提出基于博弈论的相关系数熵法和优序法获取指标组合权重，并运用前景理论模拟分析专家心理因素进行决策；最后，以5种不同烤烟智能分级机的综合性能评价为研究案例进行验证。结果表明：第一阶段设备智能分级成熟度前4个厂商等级为Ⅱ级、第5个厂商为Ⅰ级，评分为0.613 0(Ⅱ)、0.533 4(Ⅱ)、0.536 9(Ⅱ)、0.528 5(Ⅱ)、0.421 1(Ⅰ)，第二阶段综合性能评价值为1.691 5、0.922 4、0.972 9、1.012 7和0，第一个厂商设备最优，并通过灵敏度和方法对比分析，验证本方法的可行性与优越性。

关键词: 农产品智能分级设备, 不确定性, 群体共识, 非理性心理, 两阶段