Constraint-Aware Automated Daily Meal Planning: A Controlled Comparison of Genetic Algorithm and Particle Swarm Optimization

This study presents a controlled proof-of-concept comparison of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) for automated daily meal-plan generation. The evaluation is deliberately restricted to the data available in the current prototype: a 24-item Japanese-style food database, one nutritional profile, one random seed (42), 100 iterations, and an equal population/swarm size of 60. Both optimizers use the same scalarized fitness function that combines nutrient-target adherence, penalties for constraint violations, and a food-diversity bonus. In the representative run, GA achieved a final fitness of 5.15 compared with 4.92 for PSO and reached its plateau by generation 14, whereas PSO reached its final value near iteration 75. The GA solution had a daily cost of ¥980, while the PSO solution had a lower daily cost of ¥650. Both solutions remained within the specified macronutrient tolerance bounds. These results reveal a configuration-specific trade-off: GA produced a higher scalar fitness and earlier convergence, whereas PSO produced a lower-cost plan and smaller deviations for several macronutrients. Because the experiment uses a single profile and a single random seed, the findings are descriptive and should not be interpreted as statistical evidence of general algorithmic superiority.

@article{Arai2026,
  title     = {Constraint-Aware Automated Daily Meal Planning: A Controlled Comparison of Genetic Algorithm and Particle Swarm Optimization},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {17},
  number    = {9},
  year      = {2026},
  publisher = {The Science and Information Organization},
  author    = {Kohei Arai and Diva Kurnianingtyas and Nathan Daud},
  doi       = {10.14569/IJACSA.2026.0170915},
  url       = {https://doi.org/10.14569/IJACSA.2026.0170915}
}