TIERS Information Technology Journal

A Comparative Study of Three Decision Support Methods: Proving Consistency in Decision-Making with Identical Inputs

Asyahri Hadi Nasyuha — 2025-06-12

Decision-making in complex environments often requires evaluating multiple alternatives against various criteria, which can sometimes result in inconsistent outcomes when different decision support methods are employed. Such inconsistencies pose significant challenges for decision-makers in determining the most reliable methodology. To address this gap, the present study examines whether three widely adopted decision support methods, Simple Additive Weighting (SAW), Simple Multi-Attribute Rating Technique (SMART), and Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), produce consistent results when applied to identical input values, criteria, and alternatives. The primary aim is to explicitly assess the consistency of decision-making outcomes across these methods under controlled conditions. The evaluation was conducted using a set of alternatives, with A1 consistently emerging as the top choice. Specifically, the SAW method produced a final score of 0.8998 for A5, the SMART method assigned a value of 0, and the TOPSIS method yielded a closeness coefficient of 0.826 for the same alternative. The unique contribution of this study lies in its systematic, side-by-side comparison of SAW, SMART, and TOPSIS using precisely the same dataset, an approach seldom addressed in prior research. By empirically demonstrating that these methods generate identical rankings under strictly controlled scenarios, this research provides new evidence supporting the methodological robustness and practical interchangeability of these widely used decision support techniques. The findings underscore the reliability of these methods in facilitating objective decision-making and offer valuable guidance for researchers and practitioners in selecting the most suitable DSS method without concern for inconsistent results.

Optimizing Socioeconomic Features for Poverty Prediction in South Sumatera

Terttiaavini Terttiaavini — 2025-06-12

Poverty in South Sumatera remains a complex challenge influenced by socioeconomic factors. Traditional methods often fail to capture nonlinear relationships critical for accurate prediction. This study enhances poverty prediction by optimizing feature engineering using 32-variable socioeconomic data from South Sumatra for the years 2019 to 2023. Data preprocessing included cleaning, imputation, normalization, and outlier handling. Feature aggregation created composite indices: Education Index (P1, P2, P3), Health Index (AH1–AH4), Economic Index (IE, GR, AI, EG), and Healthcare Workforce Index (HW1–HW9). Feature interaction derived ratios such as Income vs. Economy (AN/Education Index), Infrastructure vs. Health (road length/Healthcare Workforce Index), and Unemployment vs. Workforce (HI/AT), highlighting interdependencies. Dimensionality reduction (PCA) and Lasso Regression selected eight key predictors, including Year and Poverty Level. Among tested models, Random Forest performed best (R²=0.7244, MAE=0.2489). SHAP analysis identified Education and Economic Indices as top predictors. Optimized feature engineering improved model accuracy and interpretability, supporting targeted poverty reduction strategies in South Sumatera.

Ambidextrous AI Governance Design Based on COBIT 2019 Traditional and DevOps for TelCo’s Digital Transformation

Nasywah Nabilah Putri — 2025-06-30

Artificial intelligence (AI) is a key enabler of digital transformation in telecommunications, improving operational efficiency and customer experience. However, telecom companies face governance challenges such as regulatory compliance risks, security vulnerabilities, and limited risk management capabilities, hindering effective AI adoption. This case study aims to address AI adoption challenges by developing an ambidextrous AI governance framework based on COBIT 2019 and DevOps. Using design science research methodology, the framework was designed and evaluated through semi-structured interviews with key TelCo stakeholders and validated with internal documents until data saturation was reached. The analysis applied the ambidextrous COBIT 2019 framework across seven governance components. Governance and Management Objectives (GMOs) were prioritized based on design factors, devops, national regulations (SOE Minister No.PER-2/MBU/03/2023 and ICT Minister No.5/2021), and literature. As a result, APO12 (Managed Risk) was selected as the key objective. Recommendations include formalizing AI governance roles, enhancing AI-related risk training, and implementing advanced GRC and automation tools. This improvement will increase the APO12 maturity level from 3.83 to 4.66. This improvement will enhance TelCo’s capabilities in risk management, compliance, and innovation, offering practical insights for practitioners and contributing to the academic discourse on ambidextrous AI governance for sustainable digital transformation.

Hydrogen Supply Chain Network Optimization for Supporting Urban Hydrogen Vehicle Infrastructure Development

Rahmad Fajri Anasrul — 2025-06-30

This study addressed the rising concerns regarding greenhouse gas emissions and the depletion of fossil fuel resources by exploring hydrogen as a clean energy alternative. The Indonesian government established a national roadmap that prioritized the transportation sector as a starting point for hydrogen deployment. The objective of this research was to design and optimize a hydrogen supply chain network in Jakarta, a densely populated urban area considered strategic for early adoption. The study applied a two-stage approach. First, potential locations for Hydrogen Refueling Stations (HRS) were pre-selected based on spatial and demographic scoring using a modified gravity model. Then the second, the optimal placement of HRS and hydrogen suppliers was determined through a Mixed-Integer Linear Programming (MILP) method. The entire modeling and optimization process was implemented in Python, with MILP solved using the Gurobi optimizer. A total of 216 existing gas stations were assessed and grouped into five priority levels. The optimization was conducted for three planning periods: 2026-2030, 2031-2035, and 2036-2040. The results showed that integrating new HRS into existing infrastructure reduced land use and investment costs. Sensitivity analysis indicated that daily HRS capacity, hydrogen demand, and capital cost were the most influential factors. The study concluded that this integrated approach provides an efficient, flexible, and sustainable foundation for future hydrogen infrastructure development in urban regions.

Solving an Optimization Problem of Image View Layout with Priority using Heuristic Approach

Lely Hiryanto — 2025-06-30

The image view layout with priority (IVLP) problem focuses on efficiently arranging picture cards of uniform height but varying widths into the minimum number of 2D frames or display sets and prioritizing images with higher priority to be placed at the earlier displays. We mathematically modeled IVLP using integer linear programming. To approximate IVLP solutions, we introduce a greedy-based heuristic, Best-Fit-IVLP (BFI), and a swarm optimization algorithm, Ant Colony Optimization (ACO). BFI allocates picture cards in descending order of priority and width for each display line, seeking another card that can optimally fill the remaining space on each line. In contrast, ACO randomly arranges cards from high to low priority within every line. Experimental results using different numbers of SVG images indicate that BFI and ACO generate solutions close to optimal. BFI demonstrates superior practicality, executing significantly faster than ACO; for 160 images, BFI runs in 0.00044 seconds compared to ACO's 117.93 seconds. Both BFI and ACO achieve space utility rates ranging from 0.578 to 0.8. While BFI consistently produces the same card arrangement, ACO offers diverse arrangements for identical optimal display set counts and space utilization.

Comparative Analysis of YOLOv5n and YOLOv8n Deep Learning Models for Precision Detection of Klowong Defects in Batik Fabric

Rifqi Restu Hamidi — 2025-06-30

This study presents a comparative analysis of two deep learning object detection models, YOLOv5n and YOLOv8n, for the precies identification of Klowong defects in batik fabric. The evaluation was carried out using a custom dataset consisting of 3,138 annotated images, with 921 allocated for testing and containing 1,295 defect instances across nine defect classes. The main findings show that YOLOv8n outperforms YOLOv5n in both speed and accuracy. YOLOv8n achieved a higher F1-score of 0.87 at a lower confidence threshold (0.297), compared to YOLOv5n’s F1-score of 0.86 at a higher threshold (0.46). In addition, YOLOv8n reduced training time significantly (0.320 hours vs. 0.868 hours) and delivered much faster inference speed (2.9 ms/image), nearly three times quicker than YOLOv5n. Although both models performed well in detecting common defects, YOLOv8n showed more stable results on complex defect types. These improvements make YOLOv8n more suitable for real-time applications in batik production environments. Its efficiency and accuracy support the development of fast and reliable automated quality control systems in traditional textile industries. This research emphasizes the importance of using modern lightweight architectures like YOLOv8n to enhance defect detection performance in practical manufacturing settings.

Performance Analysis and Traffic Flow Simulation of Tukad Pakerisan Road Segments Using VISSIM in South Denpasar

I Gede Fery Surya Tapa — 2025-06-30

The increasing ownership of motor vehicles has significantly contributed to heightened levels of traffic congestion. This study was conducted on Tukad Pakerisan Road in Denpasar City and aims to evaluate the current performance of the road segment by employing traffic modeling through Vissim software. The research adopts the Indonesian Highway Capacity Guidelines (PKJI) as the methodological framework. Data collection was carried out via a 12-hour on-site survey across two road segments of Tukad Pakerisan. The analysis revealed that the traffic flow volume reached 2025.45 Passenger Car Station/hour on Segment A and 1865.65 Passenger Car Station/hour on Segment B. The respective road capacities were 1877.669 Passenger Car Station/hour and 1671.583 Passenger Car Station/hour. The degree of saturation was found to be 1.08 on Segment A and 1.12 on Segment B, indicating Level of Service (LOS) F—characterized by severe traffic congestion. The simulation indicates significant future congestion, with projected saturation levels exceeding 1.5, underscoring the need for integrated mitigation strategies such as adaptive signal control and vehicle restriction policies. A five-year performance projection further suggests a continual increase in the degree of saturation, surpassing the acceptable limit of 0.85 as stipulated in PKJI 2023. These findings underscore the urgent need for capacity enhancement on Tukad Pakerisan Road. The study recommends the installation of additional traffic signage and the implementation of traffic engineering strategies to mitigate congestion along this critical corridor in South Denpasar.

A Zero-Shot Aspect-Based Sentiment Analysis of Public Perception Toward AI Chatbots

Naufal Andila Fauzan — 2025-06-30

The rapid development of AI chatbots has sparked public discussion on social media regarding their performance, ethical implications, and related concerns. While past studies primarily focused on individual chatbot model using traditional sentiment analysis, this study implements a novel application of Zero-Shot Aspect-Based Sentiment Analysis (ABSA) on 17,562 tweets mentioning AI chatbots such as ChatGPT, Bard (now Gemini), and DeepSeek, utilizing an efficient sentiment extraction method without supervised training. Six aspects were analyzed to understand the sentiment pattern and the results show the discussion was dominated by negative sentiment, with Bard receiving the most positive sentiment, potentially shaped by brand trust and user familiarity. On the other hand, DeepSeek and ChatGPT attracted more criticism, especially related to performance and bias aspects. This study offers data-driven suggestions for developers, including improving response accuracy to shape user trust, reducing biased output, and developing real-time discourse analysis. Future work should incorporate multiple platforms to avoid bias, analyze more AI chatbot models, and include temporal sentiment for broader insights.