<p>The oil refining process, situated at the downstream phase of the oil and gas value chain, converts crude oil into fuels and petrochemical feedstocks. Planning this process involves selecting crude oil types, determining purchase volumes, and setting operational production levels. This work addresses the oil refining planning problem under demand uncertainty, a critical challenge in competitive markets, using a two-stage stochastic optimization framework for both profit maximization and cost minimization objectives. We investigate two solution methods: (i) a scenario-based approximation that models uncertainty via a finite set of realizations, and (ii) a linear decision rule (LDR)-based method that models adaptive decisions as affine functions of uncertain parameters. The solution quality is evaluated through simulation-based testing, and the value of incorporating uncertainty is quantified using the Value of Stochastic Solution (VSS) metric. Numerical experiments demonstrate significant VSS values achieved either by substantial crude selection changes or minimal refined purchase adjustments. Further, the qualitative benefit of the stochastic solution is emphasized, particularly in mitigating future demand shortfalls for the cost minimization objective. The LDR method is shown to be computationally efficient and independent of instance-specific parameters like the number of scenarios. We then highlight the role of uncertainty definition—degrees of freedom– in determining the flexibility of LDR structure. Across all numerical experiments, simulation-based evaluation provided enhanced insights into solution quality, underscoring its importance in assessing stochastic approximations.</p>