The food banking system is a possible solution for two global problems, namely, food loss/waste and food insecurity. At present, experienced volunteers manage decision-making operations on food distribution at a focused food bank office. A computational optimisation system can help in sustainable operations. This research aims to develop a computational optimisation model that shows the food bank’s optimal food distribution. The problem solved in this research is a combinatorial optimisation problem. The food bank seeks the best combination of foods they receive to fulfil each welfare organisation’s requirements, prioritising small ones. Two types of vectors, namely, food and request vectors, were set. Cosine similarity was calculated using the two vectors to determine how well the distributed food fulfils each organisation’s needs. Simulated annealing (SA), a well-known metaheuristic algorithm was used for optimisation in this study. The first experiment was conducted to determine the optimal condition of SA for food distribution by comparing 18 different combinations of the two parameters. The second experiment confirmed the performance of the program on other datasets. The last experiment was conducted to determine the effect of the
cutoff concept to save calculation time on a food bank distribution problem. After the experiments, the fundamental parameter conditions of SA were established. Further work can be implemented to find solutions for fulfilling the requests of more organisations with large numbers of users.
View full abstract