Mariyadi, Budiyan
Unknown Affiliation

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

PENDAMPINGAN DESAIN TATA LETAK GUDANG EFISIEN MENGGUNAKAN METODE ARC–ARD BERBASIS SIMULASI FLEXSIM DI PT MIKRON PRESISI INDONESIA Raka Fahrezi Putrapriatna; Mariyadi, Budiyan; Inten Tejaasih
Jurnal Pengabdian Indonesia (JPI) Vol. 2 No. 1 (2026): Vol. 2 No. 1 Edisi Januari 2026
Publisher : PT. Jurnal Center Indonesia Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62567/jpi.v2i1.1523

Abstract

This study aims to design the warehouse tata letak at PT MIKRON PRESISI INDONESIA that experiences post-relocation irregularities in production facilities, along with adjustments to the area area and available rental costs. The absence of a racking system and clear classification of goods has caused the storage, retrieval, and work safety processes to not run efficiently. The research approach uses a qualitative descriptive method by collecting data through field observations, interviews, and documentation. The tata letak design was carried out by implementing the Activity Relationship Chart (ARC) and Activity Relationship Diagram (ARD) methods, then tested through simulations using FlexSim software with two operator scenarios. The simulation results showed that the proposed alternative tata letak design was able to improve the workflow by grouping goods into seven categories and arranging shelves based on the level of proximity of activities. However, the utilization rate of operators is still relatively low, namely 17.65% and 19.78%, due to the periodic pattern of goods arrival and efficient search times due to a more structured storage system. These findings indicate opportunities for human resource optimization through reducing the number of operators or adjusting workloads. The resulting tata letak design makes a practical contribution to improving warehouse space efficiency and operations, and can be used as a reference for applications for companies with similar characteristics.
STUDI KOMPARATIF MODEL MACHINE LEARNING DALAM MEMPREDIKSI KETERLAMBATAN PEGAWAI: LOGISTIC REGRESSION, SVM, DAN RANDOM FOREST Palupi, Inggrid Nindia Aprila; Mardianto, M Fariz Fadillah; Yuadi, Imam; Mariyadi, Budiyan
J@ti Undip: Jurnal Teknik Industri Vol 21, No 1 (2026): Januari 2026
Publisher : Departemen Teknik Industri, Fakultas Teknik, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/jati.21.1.76-87

Abstract

Keterlambatan karyawan adalah salah satu jenis pelanggaran terhadap disiplin kerja yang dapat berdampak pada produktivitas dan efektivitas organisasi. Penelitian ini bertujuan untuk mengembangkan serta membandingkan performa dari tiga algoritma machine learning Regresi Logistik, SVM, dan Random Forest dalam memprediksi keterlambatan pegawai dengan menggunakan data keterlambatan dan karakteristik individu. Dataset yang digunakan terdiri dari 1902 data, yang dibagi 80% data training dan 20% data testing dengan enam variabel, mencakup usia, lama bekerja, status pernikahan, jarak tempat tinggal ke kantor, jenis kendaraan yang digunakan, dan gaya hidup. Hasil analisis menunjukkan bahwa Random Forest memberikan kinerja prediktif yang paling baik dalam mengenali pegawai yang memiliki potensi untuk terlambat, dengan nilai akurasi tertinggi sebesar 0.82, presisi sebesar 0.93, recall sebesar 0.84, dan F1-score sebesar 0.88. Model ini terbukti dapat menunjukkan kemampuan klasifikasi yang andal dan seimbang. Analisis feature importance mengidentifikasi usia dan masa kerja sebagai faktor paling berpengaruh terhadap prediksi keterlambatan. Temuan ini tidak hanya memberikan wawasan baru dalam pengelolaan kedisiplinan pegawai, tetapi juga membuka peluang implementasi sistem peringatan dini yang dapat diintegrasikan ke dalam sistem kehadiran digital organisasi. Penelitian ini merekomendasikan perluasan variabel untuk studi lanjutan dan pemanfaatan hasil model sebagai dasar penyusunan kebijakan SDM yang lebih adaptif dan berbasis data. Abstract[Comparative Study of Machine Learning Models in Predicting Employee Delay: Logistic Regression, SVM, and Random Forest] Employee tardiness is one type of violation of work discipline that can impact organizational productivity and effectiveness. This study aims to develop and compare the performance of three machine learning algorithms Logistic Regression, SVM, and Random Forest in predicting employee tardiness using tardiness data and individual characteristics. The dataset used consists of 1902 data, which is divided into 80% training data and 20% with six variables, including age, length of service, last education level, marital status, distance from residence to office, type of vehicle used, and lifestyle. The results of the analysis show that Random Forest provides the best predictive performance in identifying employees who have the potential to be late, with the highest accuracy value of 0.82, precision of 0.93, recall of 0.84, and F1-score of 0.88. This model is proven to be able to demonstrate reliable and balanced classification capabilities. Feature importance analysis identifies age and length of service as the most influential factors in predicting tardiness. These findings not only provide new insights into employee discipline management but also open up opportunities for the implementation of an early warning system that can be integrated into the organization's digital attendance system. This study recommends expanding the variables for further studies and utilizing the model results as a basis for formulating more adaptive and data-based HR policies.Keywords: sustainability industry; developing strategy; MCDM