COMPILING A DIET BASED ON A GENETIC ALGORITHM

Abstract

This work is devoted to solving the problem of compiling a diet using a genetic algorithm.
The task of compiling a diet is a combinatorial optimization problem. The main purpose of solving
the problem of compiling a diet is to find a suitable combination of dishes to perform the distribution
in accordance with the special needs of a person. The article provides a statement of compiling
a diet problem and its mathematical model. Since the task of compiling a diet is NP-hard and
the input data may require large computational costs for an accurate algorithm, it is reasonable to
apply a heuristic approach to solving this problem. The article highlights are in detail the main
concepts of the theory of genetic algorithms, the sequence of steps of the developed genetic algorithm
for compiling the diet, the flowchart of the genetic algorithm. To research the genetic algorithm
of compiling a diet there was developed a client-server application running the Android
operating system. The result of the genetic algorithm for compiling a diet is the seven days menu,
which is displayed and stored in the application. The client-server architecture of the application
was chosen in order to save the user's phone resources. The description of the Android-application user interface with the ability to adjust various parameters of the algorithm is given in the article.
Also the analysis of the obtained algorithm efficiency is highlighted: an estimation of the accuracy
and operating time of the developed genetic algorithm with different configurations of the algorithm.
Based on the results of the experiments, it was possible to determine the optimal values of
the configurable parameters of the genetic algorithm (the number of chromosomes, the number of
iterations, the probability of mutation), allowing to obtain good results in an acceptable time.
The characteristic features of the implemented genetic algorithm of compiling a diet is a relatively
short operating time, even in a large input data. In addition, the developed solution has a high
economic value due to the application of the algorithm in practice, for example, in the work of
nutritionists, fitness trainers, as well as for ordinary overweight users.