Peramalan Traffic Load Pada Jembatan Timbang Dengan Model Hybrid Temporal Convolutional Network dan Bidirectional Long Short-Term Memory

Rendiga, Naifa Mumtazah Rendiga (2026) Peramalan Traffic Load Pada Jembatan Timbang Dengan Model Hybrid Temporal Convolutional Network dan Bidirectional Long Short-Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221114-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (13MB) | Request a copy

Abstract

Jembatan timbang merupakan fasilitas penting dalam operasional logistik yang berperan dalam mendukung pengawasan muatan kendaraan dan kelancaran proses distribusi. Namun, karakteristik layanan yang bersifat open access menyebabkan volume arus truk berfluktuasi dan sulit diprediksi secara konsisten. Kondisi tersebut menjadi tantangan bagi PT XYZ dalam menentukan waktu maintenance, karena pengambilan keputusan masih didasarkan pada asumsi manual tanpa dukungan analisis data kuantitatif. Penelitian ini bertujuan untuk membangun model peramalan volume arus truk harian pada jembatan timbang menggunakan pendekatan Hybrid Temporal Convolutional Network dan Bidirectional Long Short-Term Memory (TCN-BiLSTM). Data yang digunakan berupa jumlah truk harian pada periode Januari 2024 hingga Desember 2025. Tahapan penelitian meliputi pengumpulan data, preprocessing, transformasi differencing, penskalaan data, pembentukan sequence, pelatihan dan pengujian model, serta evaluasi menggunakan Mean Absolute Percentage Error (MAPE). Pengujian dilakukan pada tiga proporsi data, yaitu 60:40, 70:30, dan 80:20, serta dibandingkan dengan model TCN tunggal dan BiLSTM tunggal. Hasil pengujian menunjukkan bahwa model Hybrid TCN-BiLSTM memperoleh performa terbaik pada proporsi data 80:20 dengan nilai MAPE sebesar 35,94%. Hasil peramalan dapat digunakan untuk membantu mengidentifikasi periode dengan intensitas lalu lintas rendah sebagai dasar pendukung dalam pelaksanaan maintenance jembatan timbang.
=======================================================================================================================================
Weighbridges are vital facilities in logistics operations that help ensure vehicle load compliance and the smooth flow of distribution processes. However, the open-access nature of these facilities causes truck traffic volumes to fluctuate and makes them difficult to predict consistently. This situation poses a challenge in determining the most appropriate timing for maintenance, as planned maintenance schedules are not fully supported by analyses of actual traffic loads. Conducting maintenance during peak traffic hours risks disrupting service, while delayed inspections can increase the risk of damage to weighbridge components. This thesis proposes a time-series forecasting approach to predict daily truck traffic volume using a Hybrid Temporal Convolutional Network (TCN) and Bidirectional Long Short-Term Memory (BiLSTM) model. The data used consists of truck weighing transaction data from January 2024 to December 2025, aggregated into daily truck counts. The stages of this thesis include data collection, preprocessing and time-series data preparation, design of the Hybrid TCN-BiLSTM model, model training and testing, and evaluation using the Mean Absolute Percentage Error (MAPE). Testing was conducted on data splits of 60:40, 70:30, and 80:20, and the results were compared with those of standalone TCN and BiLSTM models. The test results show that the Hybrid TCN-BiLSTM model achieved the best performance on the 80:20 data split, with a MAPE value of 35.94%. The forecasting results were implemented into a web application to display historical data, forecasting results, and maintenance date recommendations to support data-driven operational decision-making.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bidirectional Long Short-Term Memory, Jembatan Timbang, Peramalan, Temporal Convolutional Network, Volume Arus Truk, Bidirectional Long Short-Term Memory, Forecasting, Temporal Convolutional Network, Truck Traffic Volume, Weighbridge.
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD30.27 Business forecasting Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Naifa Mumtazah Rendiga
Date Deposited:	28 Jul 2026 07:09
Last Modified:	28 Jul 2026 07:09
URI:	http://repository.its.ac.id/id/eprint/138271

Actions (login required)

View Item