Genetic Algorithm Meal Optimizer

A nature-inspired optimization project that designs healthy meals meeting multiple nutritional targets using a Genetic Algorithm, and benchmarks it against Cuckoo Search and Ant Colony Optimization.

Originally developed as an academic coursework project.

Overview

Given a dataset of foods — each described by its calorie, vitamin, and mineral content — the system searches for meal combinations that best satisfy a set of nutritional criteria. The core solver is a Genetic Algorithm built with the DEAP framework. Two further nature-inspired metaheuristics, Cuckoo Search and Ant Colony Optimization, are implemented as comparison baselines, and their performance is compared using statistical tests.

The optimization problem

Each candidate meal is scored by a fitness function that rewards meals falling within healthy ranges (thresholds informed by common WHO / NHS / USDA guidance):

Criterion	Target range
Calories	500–800 kcal
Water-soluble vitamins	≥ 50 mg
Fat-soluble vitamins	5–25 mg
Beneficial minerals (iron, magnesium, potassium)	≥ 300 mg
Potentially unsafe minerals (sodium)	≤ 200 mg

Approach

• Representation: linear chromosome, where each gene is one food item
• Selection: roulette-wheel selection
• Termination: 100 generations, or no fitness improvement for 10 generations
• Genetic Algorithm implemented with DEAP, including a custom ("bespoke") mutation operator
• Parameter tuning and population-diversity analysis
• Comparison against Cuckoo Search (CSA) and Ant Colony Optimization (ACO), with statistical significance testing via SciPy

Repository structure

File	Description
task1_data.py	Dataset, Food / Meal classes, train/test split
task2_algorithm_setup.py	Fitness function, chromosome encoding, CSV export
task3_genetic_algorithm.py	Core Genetic Algorithm (DEAP)
task4_csa.py	Cuckoo Search comparison
task5_parameter_tuning.py	GA parameter tuning
task6_diversity_analysis.py	Population diversity and summary statistics
task7_operator_enhancement.py	Enhanced GA with custom mutation
task10.py	Ant Colony Optimization comparison
*.csv	Result logs and comparison outputs

Requirements

• Python 3.9+
• deap, numpy, scipy

Install dependencies:

pip install -r requirements.txt

Running

Run any task from inside the project folder, for example:

python task3_genetic_algorithm.py

The later tasks import from the earlier ones (e.g. the fitness function and dataset), so run them from within the project directory.

Tech stack

Python · DEAP · NumPy · SciPy

Author

Mariam — @mariamharoon17