Towards a software tool for general meal optimisation

doi:10.1007/s10489-022-03935-0

2022-07-27

PRE

James Izzard

Fabio Caraffini *

Francisco Chiclana

DOI：10.1007/s10489-022-03935-0

原文链接

原文求助

摘要

En 中文

The following work presents a software solution capable of designing general meal plans which approach an optimal match of nutritional characteristics submitted by the user. A thorough review of existing literature indicates the absence of a software solution to this problem in its most general form. Existing solutions tend to address particular forms of the problem in specific contexts, for example, optimising culturally typical diets in response to specific medical conditions. Conversely, this work focuses on developing a nutritional software model with sufficient flexibility to be described as general, paired with a simple, specifically designed optimisation algorithm for working with the proposed prototype system. The resulting software tool can express the following characteristics: arbitrary nutritional content; economic characteristics; binary food-type classifications (e.g. vegetarian); and, because of the optimisation framework, can capture goals for any number of meals; meal composition (combinations of recipes for a given meal at a particular time); a maximum economic cost per meal; and nutritional content within each meal. The work outlines a prototype user interface to enable nutritional data and user goals to be entered and validated. Finally, based on ten specific test problems containing varied dietary goals, a basic algorithm tuning approach is described. The results suggest that the proposed prototype system can address the general meal optimisation problem. There is a discussion of several future developments to improve system capabilities and usability further.

Keyword:

Meal planning

Nutritional model

Optimisation

Genetic algorithm

Software design

Data model