Penerapan Algoritma Variable Neighborhood Search (VNS) Untuk Sistem Pendukung Keputusan Dalam Pemilihan Produk Skincare Harian Berdasarkan Jenis Kulit Dan Anggaran

Purwiantono, Greta Septy (2026) Penerapan Algoritma Variable Neighborhood Search (VNS) Untuk Sistem Pendukung Keputusan Dalam Pemilihan Produk Skincare Harian Berdasarkan Jenis Kulit Dan Anggaran. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6026241006-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Perkembangan industri skincare di Indonesia membuat konsumen memiliki banyak pilihan produk dengan variasi harga, merek, dan kualitas yang berbeda. Kondisi ini menimbulkan kesulitan bagi konsumen dalam menentukan kombinasi produk skincare harian yang sesuai dengan jenis kulit dan tetap berada dalam batas anggaran. Penelitian ini bertujuan untuk merancang dan membangun sistem pendukung keputusan berbasis website untuk merekomendasikan paket produk skincare harian berdasarkan jenis kulit dan budget pengguna. Permasalahan pemilihan produk diformulasikan ke dalam model Multiple-Choice Knapsack Problem (MCKP), yaitu pemilihan satu produk dari setiap kategori skincare dengan batasan total harga tidak melebihi budget pengguna. Data yang digunakan diperoleh dari website Sociolla melalui proses web scraping, yang terdiri dari kategori facial wash, moisturizer, sunscreen, dan serum. Data tersebut melalui tahap pra-pemrosesan, meliputi pembersihan data, ekstraksi jenis kulit, serta perhitungan Weighted Average Rating (WAR) untuk menilai kualitas produk berdasarkan rating dan jumlah ulasan. Jumlah data bersih yang didapatkan yaitu 1170 data produk skincare. Penelitian ini mengimplementasikan dan membandingkan tiga algoritma, yaitu VNS Basic, Random Search, dan Greedy + VNS. Pengujian dilakukan pada 12 skenario berdasarkan empat jenis kulit, yaitu berminyak, kering, kombinasi, dan sensitif, dengan tiga tingkat budget, yaitu Rp200.000, Rp300.000, dan Rp400.000. Hasil pengujian menunjukkan bahwa ketiga algoritma mampu menghasilkan solusi yang feasible pada seluruh skenario. Random Search memiliki waktu komputasi paling cepat, tetapi nilai objektif dan tingkat kesesuaian produknya cenderung lebih rendah. VNS Basic mampu menghasilkan nilai objektif yang tinggi dan stabil, namun pada beberapa skenario belum mencapai kesesuaian penuh. Sementara itu, Greedy + VNS memberikan hasil terbaik secara keseluruhan karena mampu menghasilkan nilai objektif tinggi dan exact match maksimal pada seluruh skenario. Dengan demikian, Greedy + VNS dapat digunakan sebagai pendekatan yang efektif untuk menghasilkan rekomendasi paket skincare harian yang sesuai dengan jenis kulit, memenuhi batasan budget, dan memiliki kualitas solusi yang baik.
===============================================================================================================================
The development of the skincare industry in Indonesia has provided consumers with a wide variety of product choices in terms of price, brand, and quality. This condition creates difficulties for consumers in determining a combination of daily skincare products that suits their skin type while remaining within their budget. This study aims to design and develop a website-based decision support system to recommend daily skincare product packages based on users’ skin type and budget. The product selection problem is formulated using the Multiple-Choice Knapsack Problem (MCKP) model, in which one product is selected from each skincare category under the constraint that the total price does not exceed the user’s budget. The data used in this study were obtained from the Sociolla website through a web scraping process, consisting of four product categories: facial wash, moisturizer, sunscreen, and serum. The data underwent several preprocessing stages, including data cleaning, skin type extraction, and the calculation of the Weighted Average Rating (WAR) to evaluate product quality based on ratings and the number of reviews. The final cleaned dataset consisted of 1,170 skincare product records. This study implements and compares three algorithms, namely VNS Basic, Random Search, and Greedy + VNS. The evaluation was conducted using 12 scenarios based on four skin types, namely oily, dry, combination, and sensitive skin, with three budget levels: IDR 200,000, IDR 300,000, and IDR 400,000. The evaluation results show that all three algorithms were able to produce feasible solutions in all scenarios. Random Search achieved the fastest computation time; however, its objective value and product suitability level tended to be lower. VNS Basic was able to generate high and stable objective values, although in several scenarios it did not achieve full suitability. Meanwhile, Greedy + VNS provided the best overall performance because it produced high objective values and maximum exact match results in all scenarios. Therefore, Greedy + VNS can be used as an effective approach to generate daily skincare product package recommendations that are suitable for users’ skin type, satisfy budget constraints, and provide good solution quality.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Sistem Pendukung Keputusan, Skincare, Jenis Kulit, Budget, Multiple-Choice Knapsack Problem, Variable Neighborhood Search, Greedy + VNS. Decision Support System,Skincare,Skin Type,Budget,Multiple-Choice Knapsack Problem,Variable Neighborhood Search,Greedy + VNS
Subjects:	T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 59101-(S2) Master Thesis
Depositing User:	Greta Septy Purwiantono
Date Deposited:	29 Jul 2026 01:16
Last Modified:	29 Jul 2026 01:16
URI:	http://repository.its.ac.id/id/eprint/139158

Actions (login required)

View Item