Improving school lunch menus with multi-objective optimisation: nutrition, cost, consumption and environmental impacts

OBJECTIVE: To support school foods programmes by evaluating the relationship between nutritional quality, cost, student consumption and the environmental impacts of menus. DESIGN: Using linear programming and data from previously served menu items, the relationships between the nutritional quality, cost, student consumption and the environmental impacts of lunch menus were investigated. Optimised lunch menus with the maximum potential student consumption and nutritional quality and lowest costs and environmental impacts were developed and compared with previously served menus (baseline). SETTING: Boston Public Schools (BPS), Boston Massachusetts, USA. PARTICIPANTS: Menu items served on the 2018–2019 BPS lunch menu (n 142). RESULTS: Using single-objective models, trade-offs were observed between most interests, but the use of multi-objective models minimised these trade-offs. Compared with the current weekly menus offered, multi-objective models increased potential caloric intake by up to 27 % and Healthy Eating Index scores by up to 19 % and reduced costs and environmental impacts by up to 13 % and 71 %, respectively. Improvements were made by reducing the frequency of beef and cheese entrées and increasing the frequency of fish and legume entrées on weekly menus. CONCLUSIONS: This work can be extrapolated to monthly menus to provide further direction for school districts, and the methods can be employed with different recipes and constraints. Future research should test the implementation of optimised menus in schools and consider the broader implications of implementation.