Implementasi Genetic Algorithm untuk Penentuan Menu Diet Sehat Bagi Penderita Penyakit Kronis dan Tidak Menular

Wulandari, Silfia Mei (2026) Implementasi Genetic Algorithm untuk Penentuan Menu Diet Sehat Bagi Penderita Penyakit Kronis dan Tidak Menular. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221073-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Penyakit tidak menular (PTM) seperti diabetes melitus tipe 2, hipertensi, penyakit kardiovaskular, hiperkolesterolemia, dan penyakit ginjal kronis kerap muncul secara komorbid, sehingga penyusunan menu diet manual menjadi sulit karena batasan nutrisi antar penyakit dapat saling bertentangan. Tugas akhir ini mengusulkan sistem pendukung keputusan berbasis Genetic Algorithm dipadukan Local Search untuk menghasilkan rekomendasi menu harian yang memenuhi kebutuhan energi dan batasan nutrisi klinis bagi pengguna dengan satu hingga tiga kondisi penyakit sekaligus. Sistem menggunakan dataset USDA FoodData Central (SR Legacy) yang telah dibersihkan, dipadukan profil kebutuhan gizi berbasis Dietary Reference Intakes (DRI). Algoritma hybrid menjalankan GA untuk pencarian solusi awal, dilanjutkan Local Search yang memperbaiki solusi dengan menargetkan nutrisi berdeviasi terbesar melalui substitusi item makanan, dengan hyperparameter GA ditentukan lewat tuning Optuna (120 populasi, 90 generasi, 0,1 elitism, 0,55 mutation, 45 iterasi local search). Evaluasi dilakukan terhadap 26 skenario profil penyakit menggunakan metrik CSR, deviasi nutrisi, konvergensi, waktu komputasi, dan validasi ahli gizi, dibandingkan dengan Greedy Algorithm. Hasil menunjukkan rata-rata CSR 71,83% (Hybrid GA-LS) berbanding 48,54% (Greedy), dengan rata-rata deviasi nutrisi masing-masing 42,41% dan 48,32%. Namun, validasi ahli gizi memberikan skor rata-rata setara untuk kedua algoritma (3,26 dari 5), mengindikasikan keunggulan kuantitatif belum tentu sejalan dengan persepsi kualitas menu klinis-praktis. Hybrid Genetic Algorithm dan Local Search terbukti lebih unggul dalam pemenuhan batasan nutrisi klinis secara kuantitatif, meski waktu komputasi lebih besar dibanding Greedy Algorithm.
======================================================================================================================================
Non-communicable diseases (NCDs) such as type 2 diabetes mellitus, hypertension, cardiovascular disease, hypercholesterolemia, and chronic kidney disease frequently occur comorbidly, making manual diet planning difficult due to conflicting nutritional constraints across diseases. This thesis proposes a Genetic Algorithm, based decision support system combined with Local Search to generate daily menu recommendations satisfying energy requirements and clinical constraints for users with one to three concurrent disease conditions. The system uses the cleaned USDA FoodData Central (SR Legacy) dataset combined with nutritional profiles based on Dietary Reference Intakes (DRI). The hybrid algorithm runs GA for initial solutions, followed by Local Search which repairs solutions by targeting the nutrient with the largest deviation through food substitution. GA hyperparameters were set via Optuna tuning (population size 120, 90 generations, 0.1 elitism, 0.55 mutation rate, 45 local search iterations). The system was evaluated across 26 disease profile scenarios using CSR, nutritional deviation, convergence, computation time, and nutritionist validation, compared against a Greedy Algorithm. Results show an average CSR of 71.83% for Hybrid GA-LS versus 48.54% for Greedy, with average nutritional deviations of 42.41% and 48.32% respectively, though expert validation yielded comparable scores for both (3.26 out of 5), indicating quantitative superiority does not necessarily align with perceived clinical-practical quality. The Hybrid GA-LS proved superior in satisfying clinical nutritional constraints quantitatively, albeit at a higher computational cost than the Greedy Algorithm.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Genetic Algorithm, Local Search, Sistem Pendukung Keputusan, Menu Diet, Penyakit Kronis dan Tidak Menular, Genetic Algorithm, Local Search, Decision Support System, Diet Menu, Chronic and Non-Communicable Disease
Subjects:	Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA76.9.I58 Recommender systems (Information filtering) Q Science > QA Mathematics > QA9.58 Algorithms
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Silfia Mei Wulandari
Date Deposited:	28 Jul 2026 01:34
Last Modified:	28 Jul 2026 01:34
URI:	http://repository.its.ac.id/id/eprint/138180

Actions (login required)

View Item