The Research On Supplier Evaluation And Selection In Emergency Supply Chain Based On The MPGA-ELM

China is one of the fewest countries in the world which have the most serious emergencies. But the lagging emergency industry and emergency supply chain lead to the frequent occurrence of these phenomena such as the insufficient supply of emergency supplies and the dismatch between supply and demand. Emergency suppliers is an important member of the emergency supply chain. So, to explore an effective method of supplier evaluation and selection is of great significance for the formation and and efficient operation of emergency supply chain.Firstly, from the perspective of industrial organization, the paper analyzes the structure of emergency supply chain system, the importance of emergency supplier in emergency supply chain and the supplier selection process. And then, based on the analysis of the new characteristics of the supplier selection in emergency supply chain, the paper analyzes the shortage of the current supplier evaluation indicator system, emphasizes the indicators as product quality, rapid response capability and collaboration compatibility. Then, build a indicator system includes 18 subdivided indicators for supplier evaluation in emergency supply chain.The indicator system of supplier evaluation in emergency supply chain is very intricate and always have inevitable correlation between indicators, so, the neural network method has been widely applied to the supplier evaluation and selection problem relying on its strong nonlinear mapping and self-learning ability. But the traditional neural network methods have many drawbacks such as sensitive to learning rate. Extreme Learning Machine(ELM) can alleviate those problems, but its arbitrariness of weights may cause bad generalization performance. Therefore, the paper uses MPGA optimize the input weights and hidden layer biases of ELM to improve its generalization ability, and proposes a extreme learning machine optimized by multi-population genetic algorithm (MPGA-ELM), Then, verifies the MPGA-ELM has stronger generalization ability than ELM through the simulate calculation of three data sets in UCI database. And then, constructed the supplier evaluation and selection model based on MPGA-ELM in emergency supply chain, and apply it to the numerical example. The calculate result demonstrate that MPGA-ELM has lower test error, and can effectively alleviate the premature problem appeared in ELM optimized by genetic algorithm. Throuth the analysis of the selection result, found that the selected supplier meet the high demand of the rapid response capability and quality for suppliers in emergency supply chain. Therefore, the MPGA-ELM algorithm is effective when applied to supplier evaluation and selection in emergency supply chain.Finally, put forward several corresponding policy recommendations for the development of emergency supply chain and the contents of emergency supplier selection.