Distribution Estimation & Efficient Optimization for Portfolio Management in Agribusiness

2020 Wagner Prize Winner
We worked with an agribusiness firm to develop new mathematical algorithms to manage the firm’s annual $1 Billion portfolio of more than 200 seeds offered to farmers in the continental U.S. The algorithms included, (i) optimization-development to deduce probability distributions for production yields for new products when managers provide partial information for the distributions, and (ii) a decomposition based development to optimize the production plan for the firm’s portfolio with more than 600 uncertainties. The algorithms have yielded substantial monetary savings and intangible benefits.