From Apr. 2 to Apr. 4, Professor Leyuan Shi’s team in College of Engineering of Peking University named the second in Syngenta Crop Challenge, which was held by 2017 INFORMS Conference on Business Analytics & Operations Research.
The event was funded by the world's leading agriculture technology company Syngenta. Since September 1, 2016, more than 600 teams from Harvard University, Stanford University and others around the world submitted their research programs. The competition is fierce. After two rounds of rigorous screening by committee, five teams were nominated to enter the final on February 24, 2017, and on April 3, the final race was launched in Las Vegas, the United States.
Team delegate Dr. Zhongshun Shi participating in the awards ceremony
In this competition, the paper presented by Professor Leyuan Shi’s team is entitled "A Decision Making Approach for Soy Seed Variety Selection via Hedging Against Weather Risk", which creatively proposed to utilize big data analysis technique and a deep neural network to construct an agricultural production forecasting model. On this basis, the team designed the optimal solution strategy based on mixed integer quadratic programming, which creatively solves the problem of optimal matching ratio of crop selection. The innovation of this scheme is using the covariance matrix to hedge the risks brought by the weather uncertainty, and combine the large data technology, the depth learning technology with the optimization theory and the method to develop the national grain output accurate forecast and the decision analysis supporting tool for the optimization strategy.
The INFORMS Award Committee has unanimously affirmed that this program is strong and extensible. Robin Lougee, the Chairman of the Awards Committee, and the chief scientist of IBM Research and T.J. Watson Research Center agreed with Joseph Byrum, chief scientist of Syngenta on that this program has an important theoretical and practical significance.
In recent years, the research work of Shi’s team has been closely involved with the national development strategy needs, including the major projects of National Natural Science Foundation "High-end equipment manufacturing network collaborative management theory in the internet and large data environment", the national public welfare industry special research " Storage platform data method analysis research in the large data environment" and other important research topics and has, carried out a fruitful research in “Internet + Intelligent” manufacturing, agricultural data analysis and optimization and other fields, greatly recognized by both domestic and foreign counterparts and industry.
The research work was completed in close cooperation between Peking University and the University of Wisconsin Madison (UWM).The authors were: Zhongshun Shi (UWM Assistant Researcher), Yu Zhao (Ph.D. student of College of Engineering), Xi Zhang (associate professor of Department of Industrial Engineering and Management), Leyuan Shi (head and professor of Department of Industrial Engineering and Management).The research got special funding by the National Natural Science Foundation and the national public welfare industry research.
The International Business Analysis and Operations Research Conference is organized by the International Association for Operational Research and Management (INFORMS), held annually and being the top conference in the field of operational research and business analysis. The conference aims to solve practical problems in business analysis with advanced operations research, scientific management methods and large data and machine learning technologies and so on. Thousands of experts and scholars from global operations research, management science and business analysis have participated in this meeting.