SCHOOLTHY: Automatic Menu Planner for Healthy and Balanced School Meals is a decision support tool that addresses the multi-objective menu planning problem in order to automatically produce meal plans for school canteens. Malnutrition is a widespread problem nowadays and is particularly serious when it affects children. In our environment, nutrition experts design healthy and balanced meal plans for children manually, which leaves significant room for improvement in terms of convenience and efficiency. SCHOOLTHY is presented herein as a proposal to improve and facilitate the work of these professionals. We focus on offering healthy and balanced meal plans that not only satisfy the recommended energy and nutrient intakes, but that also have a minimum cost and maximum variety of courses and food groups. Quantitative analyses that compare the meal plans yielded by SCHOOLTHY for meal plans designed by experts at hand and served in regional schools demonstrate the suitability of the proposal. Finally, we note that, thanks to its flexibility, SCHOOLTHY might be easily adapted to deal with other environments, such as hospitals, prisons and retirement homes, among others.

A multi-objective formulation of the Menu Planning Problem, which is termed the Multi-objective Menu Planning Problem, is presented herein. Menu planning is of great interest in the health field due to the importance of proper nutrition in today’s society, and particularly, in school canteens. In addition to considering the cost of the meal plan as the classic objective to be minimized, we also introduce a second objective aimed at minimizing the degree of repetition of courses and food groups that a particular meal plan consists of. The motivation behind this particular multi-objective formulation is to offer a meal plan that is not only affordable but also varied and balanced from a nutritional standpoint. The plan is designed for a given number of days and ensures that the specific nutritional requirements of school-age children are satisfied. The main goal of the current work is to demonstrate the multi-objective nature of the said formulation, through a comprehensive experimental assessment carried out over a set of multi-objective evolutionary algorithms applied to different instances. At the same time, we are also interested in validating the multi-objective formulation by performing quantitative and qualitative analyses of the solutions attained when solving it. Computational results show the multi-objective nature of the said formulation, as well as that it allows suitable meal plans to be obtained.