This paper provides insights into the management of perishable products from a simulation-inventory modeling perspective. By studying a just-in-time (JIT) inventory system under a single-setup multiple-delivery setting, we propose a new simulation-optimization approach for analyzing JIT purchasing agreements between a single supplier and a single retailer when dealing with randomness. Our findings show that using deterministic inventory models in a stochastic environment can be particularly useful for modeling and optimizing complex inventory systems, i.e., when difficulties arise in characterizing randomness through an analytical model.