<![CDATA[The Indonesian Makan Bergizi Gratis (MBG) program allocates a fixed budget per meal portion, creating the need for an efficient computational approach to menu planning. This study formulates MBG menu selection as a Continuous Knapsack Problem with a cost-minimization objective while satisfying nutritional requirements and compares three algorithmic approaches: the Greedy Algorithm, Grover's Algorithm, and Quadratic Unconstrained Binary Optimization (QUBO). The research was conducted through three iterative experimental stages. The initial stage, without realism constraints, produced mathematically feasible but impractical menus dominated by cooking oil. The second stage introduced ingredient upper bounds, Acceptable Macronutrient Distribution Range (AMDR), and food-category constraints, resulting in nutritionally feasible but low-variety menus. The final stage expanded the ingredient database from 8 to 18 food items, enabling more realistic menu compositions while revealing the scalability limitations of exhaustive search methods. The findings demonstrate that minimum nutritional requirements can be achieved well below the allocated MBG budget, indicating substantial opportunities for cost optimization. However, increasing the number of food alternatives significantly enlarges the search space, reducing the efficiency of exhaustive approaches. This study highlights the potential of optimization algorithms for supporting evidence-based menu planning in large-scale public nutrition programs while emphasizing the importance of balancing computational efficiency, nutritional adequacy, and menu diversity.]]>
Copyrights © 2026
