Reducing household food waste with a meal plan generating algorithm

Sammanfattning: This thesis aims to see If an algorithm could automatically generate a meal plan that is optimized for low food waste. Design and creation is the method used and the algorithm is developed over a few iterations. The idea is that the algorithm could be used in a mobile application that allows users to search for recipes, mark the ones they like, add their own, and generate a meal plan. The resulting algorithm consists of four parts; an input stage, a stage that decides when to cook each portion, an improvement stage in which the algorithm searches for a better combination of recipes, and an output stage. An adaptation of a neighborhood search is used for the search algorithm, in addition, two heuristic evaluation functions are used to evaluate the waste and other factors of the meal plan. The test results show that the algorithm can generate meal plans with no leftover items (that last shorter than 60 days) for a lot of different test cases. More recipes produce better results up to a certain point, but not so many recipes are needed to have a chance of a successful meal plan generation. Future testing and research in this area could lead to a lot of benefits for people and the climate.