Yang Zou

Robotic assembly in industrialised construction is increasingly relevant to structured off-site and prefabrication settings, where BIM-derived geometry defines nominal assembly sequences but geometric deviations, tolerance conflicts, or local obstructions can invalidate individual task steps. Existing BIM, ontology, knowledge-graph and symbolic-planning approaches represent components or generate action sequences, but rarely make recovery decisions traceable to geometry, evidence, and graph-level admissibility. This paper addresses task-level failure recovery in such settings — how an executable task graph can be locally rewritten when a nominal action becomes infeasible — through an Action Knowledge Graph (AKG) and active task-graph rewriting. Action primitives and recovery edges are grounded in embodied manipulation datasets, and each execution instance is an active task graph with bypassed-edge audit history. A candidate rewrite is committed only through a grounded recovery edge and only if postcondition, dependency, acyclicity, path-existence, and resource conditions jointly hold.The framework is evaluated in two controlled task-level scenarios: an IFC-derived door-articulation case and a parametric bolted panel-insertion case. A ten-case admissibility matrix separates unrecoverable, diagnosed-but-ungrounded, and admissibly recoverable failures, and edge ablation confirms that recovery and precedence edges are load-bearing. A capability-complete STRIPS-style baseline matches the AKG executor on all cases, clarifying that the added value lies not in exclusive recovery capability but in evidence traceability, local graph deltas, and persistent admissibility audit. Instrumented reasoning-layer timing places one recovery cycle at tens of microseconds, excluding perception, communication, and robot control. The study is a methodological validation at the task-reasoning level, not a benchmark of physical deployment.

1. Production Planning and Control (PP&C) in construction is the process of managing and coordinating all activities related to construction production among stakeholders; it has been recognised as...

Purpose: Exposure to extreme heat poses significant health risks for construction workers due to multiple reasons. However, research on heat-related illnesses (HRIs) in this sector has often relied on statistical methods, limiting insights into causal mechanisms and feature interrelations. This study addresses this gap by employing machine learning (ML) techniques to predict the severity of HRIs (fatal vs non-fatal) and identify critical contributing factors. Design/methodology/approach: A total of 370 Occupational Safety and Health (OSHA) HRI reports are analyzed using natural language processing (NLP) to extract key features. These features are combined with climatic and non-climatic variables. Four ML algorithms – random forest, extreme gradient boosting, k-nearest neighbors and logistic regression – are applied to classify HRIs as fatal and non-fatal. Feature importance and dependencies are explored using SHAP (SHapley Additive exPlanations). Findings: The random forest model achieves the highest accuracy at 70%. Employers' negligence and workers aged over 35 years are key human-centric factors in fatal HRIs. Humid subtropical and hot-summer Mediterranean climate zones are the most hazardous environmental factors. The results suggest that small-scale projects with limited budgets, renovation projects and building projects (residential/commercial) are the key contributing project-related features in the severity of HRIs. Originality/value: This study is among the first attempts to investigate HRI reports specifically in the construction industry by using artificial intelligence techniques. The findings emphasize the need for tailored heat stress prevention programs in construction projects based on their characteristics.

Established in 1975, the Australasian Universities Building Education Association (AUBEA) is a leading non-profit association of universities and institutes of technology in Oceania focused on teaching and research in the built environment. The association organises an annual conference for academics, researchers, and industry professionals to discuss advancements in education and research in the built environment, construction management, and building technologies. The AUBEA 2023 conference was held in Auckland, hosted by Massey University from 26–28 November 2023. The theme of the conference was Creating Capacity and Capability: Embracing Advanced Technologies for a Sustainable Future in Building Education and Practice.

This study presents Adaptive Dual-OPtimization with Tree learning Genetic Programming (ADOPT-GP), a dual-loop evolutionary framework that simultaneously discovers symbolic rule structures and calibrates parameters. ADOPT-GP couples adaptive genetic programming with a two-stage parameter tuning process: rapid logistic-regression initialization followed by evolutionary calibration. Across runs, fitness improves by 20%–40% on average. Against a bilevel sequential baseline, ADOPT-GP delivers similar or better accuracy while reducing runtime by over 85%, demonstrating scalability. In a university library evacuation case, it yields diverse, interpretable rules that expose tensions between group cohesion and spatial constraints, supporting context-sensitive behaviors. The approach can advance inverse generative social science (IGSS) by linking behavioral theory with computation and offers practical tools for emergency planning.

Accurate extraction of the arch profile, the key spatial geometric parameter of the core load-bearing component in arch bridges, is crucial for construction process control and for achieving the designed final bridge configuration. To overcome the limitations of existing methods—geometric information loss, sensitivity to noise, and inefficiency—when extracting continuous, precise profiles from point clouds of complex spatially curved arch ribs, this paper proposes a multi-step point cloud processing workflow. The approach integrates geometric feature constraints specific to arch bridges to enable automated, high-precision extraction of the arch profile during construction. The approach comprises three steps. First, arch point cloud subset partitioning: the primitive arch point cloud is efficiently divided using parameters from down-sampling arch point cloud data. Second, component segmentation: a Random Sample Consensus (RANSAC) algorithm, optimized with cylindrical geometric constraints, is then employed to precisely segment the point cloud of individual arch tube components from each subset point cloud. Third, arch profile extraction: the geometric invariance of the bottom edge of each arch tube is leveraged to identify feature points via local coordinate system transformation and longitudinal constraints. These feature points are then spliced together to reconstruct the complete arch profile. The proposed method is employed in multiple construction stages of a concrete-filled steel tubular (CFST) arch bridge and quantifies the vertical deformation between adjacent stages. Compared with Total Station (TS) measurements, the average error ranged from 0.24 mm to 4.13 mm, with an overall average error of 2.105 mm, demonstrating accuracy and reliability.