Claim Missing Document
Check
Articles

Found 4 Documents
Search

Multi-Agent Deep Reinforcement Learning for Handover Management in Massive Industrial Internet of Things Networks Naufan Raharya; Muhammad Suryanegara
Buletin Pos dan Telekomunikasi Vol. 21 No. 1 (2023): November 2023
Publisher : Centre for Research and Development on Resources, Equipment, and Operations of Posts and I

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.17933/bpostel.v21i1.374

Abstract

The industrial internet of things (IioT) is considered one of theIapplications in the fifth generation (5G) networks. In this application, users’ high mobility in a typical industrial scenario needs high reliability. The high mobility creates frequent handover, creating extra control signalling to a new base station (BS). The users’ congestion to the new BSs can lead to an outage. In this paper, we investigate how to manage the handover of users to improve reliability in a high-mobility scenario using deep learning. We first use an offline centralized algorithm to create labels for user association to a BS that is done without adding handover coefficient. Then, we train the neural network and use the trained parameter to make the multi-agent deep reinforcement learning (RL) learns better. This is done to avoid long iterative methods in reinforcement learning. The results show that our method can outperform the offline centralized algorithm by 40% when the handover coefficient increases.
An Implementation of Quasi-Newton Algorithm for Fast-charging Lithium-Ion Battery (LIB) Optimization in Electric Vehicle Application Anjarani, Mahmudda Mitra Anjarani; Raharya, Naufan
International Journal of Electrical, Computer, and Biomedical Engineering Vol. 2 No. 2 (2024)
Publisher : Universitas Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62146/ijecbe.v2i2.54

Abstract

Lithium-Ion Battery (LIB) is still an effective alternative technology in maximizing the efficiency of electric vehicles (EV). The application of EVs has had a significant impact in order to reduce the issue of global problems - reducing carbon gas emissions. The LIB charging mechanism with the fast-charging method is an alternative to the application of EVs on a more massive scale. However, the dynamics of the battery where the battery work function can decrease over time will affect battery performance. In addition, fast-charging efforts at LIB with maximum speed have the impact of increasing the risk of battery temperature and the existence of a larger gap in battery degradation. This paper proposes the application of Limited-Memory-Broyden-Fletcher-Goldfarb-ShannoBound Constrained (L-BFGS-B) algorithm for Lithium-Ion Battery (LIB) fast-charging optimization as an innovative solution approach in dealing with the complex LIB fastcharging dynamics. The results show that this approach is able to improve fast-charging speed and efficiency.
Anomaly Detection in Imbalance Secure Water Treatment Dataset Using LSTM-DC-Wasserstein Generative Adversarial Network with Gradient Penalty Kevin, Jonathan Marshell; Raharya, Naufan
International Journal of Electrical, Computer, and Biomedical Engineering Vol. 2 No. 3 (2024)
Publisher : Universitas Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62146/ijecbe.v2i3.56

Abstract

In modern industrial systems, particularly with the advancement of the Internet of Things (IoT), industry players can record machine and system data for comprehensive analysis. This capability is crucial for detecting anomalies and taking necessary corrective actions.However, it is common for manufacturers to lack recorded anomaly datasets, especially for newly operational systems. In this paper, we develop a model to detect anomalies in an imbalanced dataset from the Secure Water Treatment (SWaT) system. The performance of the proposed model is compared with previous works, demonstrating significant improvements in anomaly detection capabilities where it achieves accuracy of 0.9546, precision of 0.9086, recall of 0.6654, and F1 score of 0.7681
Energy Consumption Optimization for Flexible Job-Shop Scheduling in Manufacturing Industry Using Multi-Agent Reinforcement Learning Satwika Bintang Bahana; Naufan Raharya
International Journal of Electrical, Computer, and Biomedical Engineering Vol. 4 No. 1 (2026)
Publisher : Universitas Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62146/ijecbe.v4i1.227

Abstract

Modern manufacturing industries are increasingly demanding production systems that are both flexible and efficient in handling dynamic operational conditions. Traditional scheduling approaches, such as the job-shop scheduling Problem (JSSP), are limited by the constraint that each operation must be executed on a predefined machine. To address this limitation, the flexible job-shop scheduling Problem (FJSSP) was introduced, allowing alternative machine options for each operation and thereby enhancing system flexibility. This study proposes a scheduling optimization approach based on Multi-Agent Reinforcement Learning (MARL), to support decision-making in complex production environments. Experimental results demonstrate that the proposed method reduces energy consumption by up to 40.62% compared to the First Come First Serve (FCFS) method and by 35.23% compared to the Fastest Available Agent (FAA) method. Moreover, the model shows superior performance in controlling theworst-case makespan and achieves significantly higher success rates in satisfying various production constraints compared to all tested rule-based methods.