基于多属性群决策技术的科技奖励评价与管理体系优化研究

According to the non-uniformity of expert assessment results and the non-validity of evaluation system in the process of scientific and technological reward evaluation, the project which based on multiple attribute group decision-making technology will establish expert composite weights which combine dynamic weights with index weights，analyze the robustness and relative effectiveness of evaluation models, and construct the optimal scientific and technological reward evaluation system. In addition, the project will explore the rule of expert evaluation behavior and the method of control and meta-assessment of expert evaluation, and design methods to control expert evaluation behavior. Finally it will establish optimized scientific and technological reward evaluation and management system. The main content of the project research includes: the index selection, index system optimization, weighting method optimization and optimization of scientific and technological reward evaluation model; the measure and meta-assessment analysis of expert appraisal quality; the software research and development of scientific and technological reward evaluation and management system; the system optimization of scientific and technological reward evaluation.. The optimized scientific and technological reward evaluation and management system can effectively reduce the influence of subjective factors, improve the scientificity, fairness and rationality in the scientific and technological reward evaluating, help improve the competitive mechanism of scientific research, promote the optimal allocation of resources of science and technology, as well as promote continuous improvement and healthy development of the cause of China’s science and technology.

针对科技奖励评价过程中专家评审结果非一致性及评价体系非有效性问题，以多属性群决策技术为基础，建立专家动态权重与指标权重相结合的复合权重，分析评价模型的鲁棒性与相对有效性，构建优化的科技奖励评价体系；探索专家评审行为的规律以及对其进行控制与反评估的方法，通过方法设计来控制专家评审行为；建立优化的科技奖励评价与管理系统。项目研究的主要内容包括：科技奖励评价指标筛选与指标体系优化；指标赋权方法优化；科技奖励评价模型优化；专家评议质量测度及反评估分析；科技奖励评价与管理软件系统的研发；科技奖励制度优化。.优化的科技奖励评价与管理体系能有效减少科技奖励评价工作中主观因素的影响，提高科技奖励评价的科学性、公平性、合理性，有利于健全科研竞争机制，促进我国科技资源的优化配置，推动我国科技事业的持续进步和健康发展。

科技成果评价及奖励是我国科技工作的重要组成部分，研究、完善科技奖励评价与管理体系，在当前具有重要的理论与现实意义。本项目针对科技奖励评价过程中指标权重的主观性及专家评议水平的差异问题，运用聚类分析法确定专家权重，运用熵值法确定指标权重，在此基础上研究构建了专家动态权重与指标权重相结合的复合权重体系；针对科技奖励评价结果的非一致性及相对有效性问题，对科技奖励评价模型进行比较与分析，构建了基于改进TOPSIS法的优化科技奖励评价模型；针对科技奖励评价过程中专家评议质量差异问题，以数理统计学方法为基础，探讨了基于肯德尔和谐系数、离散度、信度系数与一致性系数的对专家评审行为进行控制与反评估的方法，从而达到通过方法设计来控制专家评审行为的目的。实证分析表明，这4种反评估方法均能比较客观地反映专家评议水平，具有可操作性和实用价值；针对科技奖励评价过程的信息化、标准化问题，利用C#编程语言和Access数据库设计开发了科技奖励评价系统软件，测试结果表明，该系统能够高效实现科技成果项目评审专家权重的确定、指标权重的确定、评价指标数据的处理以及科技成果的综合评价与管理工作；针对我国科技奖励制度存在的缺陷，本项目以健全科研竞争机制，促进我国科技资源的优化配置为基本目标，提出了优化我国科技奖励制度的政策建议。