Design and implementation of a constraint satisfaction algorithm for meal planning

Abstract: The world’s population is ageing. Due to societal improvements in healthcare, living standards, and socio-economic status, more and more people are living to old age. The proportion of the world's population aged 65 or over is expected to increase from 11% in 1998 to 16% in 2025. This causes a major public health issue, because with increased age there is an increased risk of developing a number of age-related diseases. However, there is increasing scientific evidence that many of the biological changes and risks for chronic disease, which have traditionally been attributed to ageing, are in fact caused by malnutrition (sub-optimal diets and nutrient intakes). This report presents a constraint satisfaction approach to planning meals while taking into account amongst other things nutritional and economic factors. Two models for generating meal plans are presented and their respective strengths and weaknesses discussed. System design, implementation and the main algorithms used are described in more detail. These algorithms include Depth First Branch and Bound and its various improvements for meal plan generation as well as Item-based Collaborative Filtering for user preferences. Our test runs show that the system works well for smaller applications but runs into problems when the number of available recipes grows or a larger number of meals are planned. The tests also show that the two modelling approaches both have useful applications. Based on the test results some suggestions for further improvement of the system conclude the report.