Abstract: The traditional empirical method for crop breeding material evaluation is difficult to meet the requirements of large-scale, programmed and data based commercial breeding, and breeders need urgently a scientific, fast and efficient method to evaluate crop breeding material. This paper proposed a method for evaluating crop breeding material based on improved gray correlation degree and TOPSIS, which could quickly evaluate breeding material according to the comprehensive performance of its traits. The proposed method consisted of the following steps: First, a comprehensive evaluation index system of breeding material was constructed by gray correlation analysis of crop traits. Then, the weight of each trait index was obtained by combining the information entropy method and expert scoring method. Finally, the comprehensive score and ranking of each material was obtained based on the TOPSIS. The trait data of 2015 National Rice Variety Regional Test Huanghuai Japonica Rice Group A was chosen to verify the effectiveness of our proposed method. The results showed that the proposed method could effectively identify the excellent varieties with good comprehensive performance from all tested varieties and was superior to the entropy method. The proposed method had been integrated into large-scale commercial breeding software “Golden Seed Breeding Platform (GSBP)”, which had been successfully implemented in 2 seed leading enterprises of China.

Key words: gray correlation degree, technique for order preference by similarity to an ideal solution (TOPSIS), breeding material, evaluation method

摘要： 传统基于经验的作物育种材料评价方法，难以适应大规模、程序化、数据化的商业化育种需求，迫切需要一种科学、快速、高效的育种材料评价方法。提出了一种灰色关联度与TOPSIS相结合的作物育种材料评价方法，可根据育种材料的性状表现快速作出综合评价。首先，通过对作物性状进行灰色关联度分析，筛选出与育种目标关联度较大的性状，以此构建综合评价指标体系；然后，组合运用熵权法与专家打分法确定指标权重；最后，基于TOPSIS方法计算每个材料的综合评分，并按评分高低对材料进行排序。选取2015年国家水稻品种区域试验黄淮粳稻A组试验数据进行模型验证，结果表明：所提方法的评价结果与品种区域试验年会审议意见趋于一致，表明所提方法能有效识别出优良品种。该方法已集成到大型商业化育种软件“金种子育种平台”，并在国内两家大型种业企业成功实施应用。

关键词: 灰色关联度, TOPSIS, 育种材料, 评价方法