A Systems Approach for Modelling Warehouse Operations: Human-Machine Interaction Within Industry 4.0 Framework

In today's market, the warehousing industry is not merely for storing items but is a central hub supporting and driving an organisation's strategic objectives. The persistent increase in e-commerce and customer expectations has pressured logistics systems, particularly warehouses, to process orders faster and more accurately. The warehousing industry sector is undergoing a significant transformation driven by Industry 4.0. This evolution has introduced cutting-edge technologies to warehouse processes and activities to reduce cost, reshape operational paradigms, enable real-time decision-making, and optimise systems across supply chains. Industry 4.0 technologies such as automation, artificial intelligence, and Internet of Things (IoT) have reshaped how humans and machines interact within logistics systems. While automation enhances efficiency and reliability, it redefines human roles, skills, and cognitive demands. At the same time, the emerging Industry 5.0 vision focuses on a human-centric, resilient, and sustainable approach, emphasising collaboration between people and technology, rather than pure automation.

This thesis positions itself at the intersection of these two industrial revolution paradigms. It aims to bridge the current Industry 4.0 focus on technology optimisation with Industry 5.0 emphasis on human-centric, using system dynamics (SD) as a methodological framework to understand and stimulate the feedback structure, interaction between human and machine in warehousing systems. Specifically, the study's objectives are: (1) to understand the impact of human factors on warehouse operations, and (2) to analyse the dynamic interactions between human factors and autonomy levels in the warehouse process. This research utilised SD modelling to capture qualitative and quantitative dynamics that shape the system performance over time. Qualitatively, this study performed a systematic literature review to identify existing variables in the literature to be validated with experts. A causal loop diagram (CLD) is developed based on identified variables that underwent a Delphi-expert approach, and which held a final round with independent experts. The variable relationships between 'Cause' and 'Effect' visualised in CLD were modified, validated, and improved through these rounds. Seven balancing loops and four reinforcing loops represent the complex human factors that impact warehouse operations with Industry 4.0. In order to keep the system in balance, the balancing loops serve as stabilising forces that control important factors, including inventory levels, workforce allocation, machine maintenance, and staff workload. On the other hand, the reinforcing loops intensify changes, such as higher productivity and job satisfaction, which increase errors, fatigue and workloads. Overall, the dominance of balancing loops shows that the system in understanding human factors in warehousing is self-regulating and adaptive despite changes. However, the reinforcing loops reveal areas where performance can be improved or may deteriorate if not properly managed especially when workers are under high workload demand and stress.

Quantitatively, the CLD is then transformed into a stock and flow diagram (SFD), deriving suitable equations to analyse the system behaviour over time. The SD model of this study is focused on the warehouse inbound process to test the impact of human factors, considering the autonomy level. The SFD model addressed the question: What is the optimal number of humans and machines needed to optimise the operations, considering the impact of physical, mental, and psychosocial elements on warehouse operations? The simulation ran onto a baseline case to mimic the observed secondary warehouse performance to be compared across different sensitivity analyses. Ten what-if scenarios were developed to evaluate varying degrees of three different dimensions of collaboration levels, human factor effects, and operational demand implication. The scenario analyses indicate the complex interplay between human factor effects under dynamic warehouse conditions. Based on the scenario outcomes, this study proposed two-dimensional managerial interventions between technology and workforce. Technology intervention focused on autonomous warehouse with maintenance disruptions and large throughputs. Meanwhile, workforce intervention focused on labour resources, recovery rate, and well-being programs. The findings demonstrate that the most optimal balance of human and machine is when psychosocial and physical aspects are being considered in improving workers’ well-being and operational efficacy.

In conclusion, this study advances the theory of socio-technical systems modelling within logistics research and expands the application of SD to human-machine interactions. Both qualitative and quantitative developed SD in this study offer decision-making tools to industry stakeholders, engineers, and researchers to optimise the warehousing operations considering human elements in the systems. This study provides conceptual and practical insights into the evolution of warehousing systems towards achieving operational efficiency with human well-being in the context of Industry 5.0.