Stochastic Optimisation for Tomato Paste Production Under Rainfall Uncertainty: A Case Study of Northern Nigeria

This paper addresses the critical planning challenge faced by tomato processors in Northern Nigeria, where production decisions must be made months before the harvest, under profound uncertainty driven by rainfall variability. Despite being a major tomato producer, Nigeria relies heavily on tomato paste imports due to a disconnect between seasonal production and processing capacity, leading to massive post-harvest losses in high-yield years and idle capacity in low-yield years. We develop a two-stage stochastic programming model to optimise how a processor should reserve capacity across three product streams, fresh market, paste, and dried tomatoes, to maximise expected profit under uncertain, rainfall-driven yields. The model is calibrated with a decade of historical rainfall, yield, and price data from Kano State, Nigeria. Our key findings demonstrate that the stochastic model yields a Value of Stochastic Solution (VSS) of ₦8.4 million (18.6%) compared to a deterministic approach based on average expectations. Furthermore, it reduces expected spoilage by 55% , from 412 to 187 tonnes. The optimal strategy shifts capacity towards paste production, which acts as a critical buffer against yield volatility. To bridge the gap between academic modelling and practical application, we developed and validated TomatoPro, an interactive decision-support tool. In validation workshops with 15 processors and extension agents, TomatoPro achieved an excellent System Usability Scale (SUS) score of 82.4 and significantly improved user decision-confidence. This research provides the first stochastic optimisation model tailored to West Africa's tomato value chain and offers a replicable framework for improving climate resilience and reducing food loss in perishable agricultural supply chains across sub-Saharan Africa.