Design and Development of an AI-Based Interactive Simulation to Support Human Factors Mixed-Methods Analysis

Yusdin, Hammam Dhiyaurrahman (2026) Design and Development of an AI-Based Interactive Simulation to Support Human Factors Mixed-Methods Analysis. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010232017-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (34MB) | Request a copy

Abstract

Praktik Human Factors (HF) dan teknik industri memerlukan analisis mixedmethods yang bergantung pada kemampuan integrative reasoning untuk menggabungkan data kuantitatif dan narasi kualitatif pekerja. Namun, pembelajaran teknik masih didominasi studi kasus statis sehingga kurang merepresentasikan kompleksitas praktik lapangan. Penelitian ini mengembangkan dan mengevaluasi Multi-Agent Integrated Reasoning Assistant (MIRA), aplikasi simulasi berbasis Large Language Model dengan Virtual Simulated Person (VSP) dan pedagogical agent yang memberikan cognitive scaffolding secara real-time. Melalui workshop kuasi-eksperimental, MIRA dibandingkan dengan pembelajaran berbasis kertas menggunakan evaluasi kinerja sistem, System Usability Scale (SUS), NASA Task Load Index (NASA-TLX), Expectancy-Value-Cost (EVC), dan normalized learning gain (g). MIRA menunjukkan keselarasan interaksi yang tinggi (Fslot = 96,76%; Hturn = 98,38%) tanpa leakage maupun drift, serta mendukung extraneous cognitive load yang rendah (SUS = 77,36), linguistic cost yang lebih rendah (EVC-C = 2,83), dan productive struggle (NASA-TLX = 4,00; EVC-V = 4,32). Meskipun ceiling effect akibat skor pre-test yang tinggi membatasi peningkatan keseluruhan (gtotal = 0,1917), kelompok MIRA tetap menunjukkan peningkatan integrative reasoning (g = 0,1064), sedangkan kelompok berbasis kertas hampir tidak meningkat (g = 0,04). Temuan ini menunjukkan bahwa cognitive scaffolding berbasis AI, didukung Cognitive Load Theory dan Self-Regulated Learning, berpotensi meningkatkan integrative reasoning dan ergonomic diagnostic judgment.
==================================================================================================================================
Human Factors (HF) and industrial engineering practice requires mixedmethods analysis, an analytical task that depends on integrative reasoning, the ability to synthesize objective quantitative risk data with subjective qualitative worker narratives. Although international standards (ISO 6385:2016) emphasize this approach for effective work system design, traditional engineering pedagogy still relies heavily on static, rule-based case studies, leaving novice students ill-equipped for the ambiguity of real-world fieldwork. To address this gap, this study presents the design, development, and empirical evaluation of the Multi-Agent Integrated Reasoning Assistant (MIRA). MIRA is an AI-based interactive simulation utilizing a dual-agent architecture: a Virtual Simulated Person (VSP) that replicates stakeholder interactions and an intelligent pedagogical agent that provides real-time cognitive scaffolding. A quasi-experimental workshop compared MIRA with a traditional paper-based approach, evaluating system performance, usability (SUS), subjective workload (NASA-TLX), motivation (Expectancy-Value-Cost), and normalized learning gains (g). MIRA demonstrated high interactional alignment (Fslot = 96.76%, Hturn = 98.38%) and absolute boundary compliance (zero leakage or drift). This stable interaction reduced extraneous load (SUS = 77.36) and lowered the linguistic cost (EVC-C = 2.83) for non-native English speakers while promoting productive struggle (NASA-TLX = 4.00; EVC-V = 4.32). Although a high pre-test baseline (180.88 ± 41.86) produced a ceiling effect that constrained overall improvement (gtotal = 0.1917), the MIRA cohort achieved positive gains in integrative reasoning (g = 0.1064), whereas the paper-based group showed minimal improvement (g = 0.04). These findings suggest that AI-mediated cognitive scaffolding can support the development of field-ready integrative reasoning and ergonomic diagnostic judgment.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Cognitive Load Theory, Human Factors, Large-Language-Model, SelfRegulated Learning, Virtual Simulated Person (VSP)
Subjects:	B Philosophy. Psychology. Religion > BF Psychology > BF318 Learning, Psychology of (Deep learning) B Philosophy. Psychology. Religion > BF Psychology > BF441 Critical thinking L Education > LB Theory and practice of education > LB1029.S53 Educational games. Simulation methods L Education > LB Theory and practice of education > LB1044.87 Internet in education (e-learning). Virtual reality in education. L Education > LB Theory and practice of education > LB2300 Higher Education
Divisions:	Faculty of Industrial Technology > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Hammam Dhiyaurrahman Yusdin
Date Deposited:	02 Aug 2026 19:33
Last Modified:	02 Aug 2026 19:33
URI:	http://repository.its.ac.id/id/eprint/141749

Actions (login required)

View Item