A0229
Title: Climate-driven inventory optimization using generalized newsvendor model with random supply and demand
Authors: Soham Ghosh - Indian Institute of Technology Indore (India) [presenting]
Sujay Mukhoti - Indian Institute of Management Indore (India)
Pritee Sharma - Indian Institute of Technology Indore (India)
Abstract: A decision-theoretic framework is developed to optimize humanitarian food supply in drought-prone African countries using an extended newsvendor model that incorporates climate and production risk. Agricultural output is modeled as a function of temperature and latent volatility in precipitation, derived from a Bayesian stochastic volatility structure applied to the autoregressive standardized precipitation index (SPI). MIDAS (mixed data sampling) regressions are employed with beta polynomial weighting to link high-frequency climate risk measures to annual production. The expected production is embedded within a piecewise-linear cost function that penalizes both excess supply and shortfalls. Using particle filtering, the distribution of latent volatility is approximated, and nested expectations that define the cost function are evaluated. The proposed model captures the asymmetric and nonlinear influence of climate risk on food security outcomes and yields optimal inventory decisions under uncertainty. This framework supports adaptive food aid planning by integrating statistical forecasting, volatility modeling, and supply chain optimization. The approach is particularly suited for use by institutions such as the World Food Program, enabling more efficient and responsive resource allocation. The methodology contributes to the intersection of climate econometrics, statistical decision theory, and inventory management under risk.
