Simulation based approach for inventory management of packaged milk at an unorganized Indian retailer: A case study

Managing inventory for perishable goods such as milk is a complex task, particularly for small unorganized retailers who handle multiple products and lack the resources to closely monitor each item or employ costly optimization tools. Such retailers typically relay on simple thumb rules or fixed-order policies to manage stock. This study develops a Discrete-Event Simulation (DES) model to replicate the daily operations of such a retailer managing multiple milk SKUs under demand uncertainty and product perishability. A Genetic Algorithm (GA) is then applied to optimize the parameters of a base-stock policy, aiming to improve ordering decisions. Results indicate that the GA-optimized policy yields marginally higher profitability and reduced lost sales compared to the traditional fixed-order approach. While the improvement is statistically significant, its magnitude remains modest, suggesting that the simple heuristics used by small retailers are relatively effective and continue to be practical for real-world supply chain operations.