基于自组织I/O映射神经网络的产业链核心竞争力

Core competitiveness of industrial chain based on self-organizing input/output mapping neural network

为有效提升产业链竞争优势,合理、高效地配置产业链的内部资源,对产业链核心竞争力进行了研究。分析了其构成要素,建立了数学模型与指标体系,并针对产业链核心竞争力高维指标体系内存在的部分相关性高、影响度低的冗余因素,提出一种基于BP神经网络、遗传算法、粒子群算法融合的自组织I/O映射神经网络模型。基于该模型,对指标体系内的各复杂因素进行了筛选,过滤掉冗余指标因素,识别出最能反映输入输出映射关系的因素变量,并在此基础上构建了关键因素识别模型,以识别构成产业链核心竞争力的关键因素。

To enhance the competitive advantage of industrial chain effectively,and to allocate resource belonging to industrial chain quickly and reasonably,the related mathematical model and index system were established through researching the core competitiveness of industrial chain and analyzing the components of core competitiveness.According to some redundant factors with high correlation and low impact in the high-dimensional index system of industrial chain＇s core competitiveness,a Self-Organizing Input/Output Mapping Neural Network（SIOM-NN）based on integration of BP neural network,particle swarm optimization algorithm and genetic algorithm was proposed.On this basis,the various complex factors in index system were screened to filter out redundant index factors,and the satisfactory factors that reflected the mapping relationship between input and output were identified.An identification model of key factor was constructed,and the key factors which constituted core competitiveness of industrial chain were identified.