Journal of Construction Engineering, Management & Innovation

Melike Aygün Çakıroğlu Suat Özdemir

Bike-sharing systems rely on accurate short-term demand forecasts to prevent shortages, surpluses, and costly rebalancing operations. Accurate predictions are also essential for enhancing the environmental, operational, and social sustainability of these systems, as improved forecasting may help reduce truck-based redistribution, potentially lowering emissions and supporting more efficient resource usage and more reliable urban mobility services. In this study, we evaluate a wide spectrum of forecasting approaches—ranging from classical time-series models (ARIMA, Prophet) and ensemble learners (Random Forest, XGBoost) to spatio-temporal deep learning models (LSTM variants, ST-GCN, GraphWaveNet, and GNN)—using a local station-level dataset. We incorporate both temporal history and spatial dependencies among stations under a unified evaluation protocol (24-hour look-back, one-hour-ahead prediction). Our findings show that integrating spatial context consistently improves accuracy. The spatio-augmented Random Forest (RF-ST) achieves the best performance, reducing error rates by over 10% compared to its temporal-only counterpart and by more than 35% relative to ARIMA and Prophet. Graph-based neural models (e.g., GNN) deliver comparable accuracy, further confirming the benefits of explicit spatial modeling. These results highlight the potential sustainability implications of spatio-temporal forecasting, suggesting that more accurate station-level predictions may support more sustainable rebalancing strategies, potentially help reduce operational inefficiencies, and strengthen the long-term viability of bike-sharing systems as a sustainable transportation mode.

https://doi.org/10.31462/jcemi.2026.652

Şerife AK

The aim of the current study is to carry out a scientometric analysis of academic research investigating gender-based barriers encountered by women employed in the construction industry and to identify gaps in the literature. In this context, 214 publications indexed in the Web of Science (WoS) database were identified using bibliometric methods and analyzed using scientometric methods. The findings demonstrated a steady increase in publication trends over time. The literature was structured around themes of gender discrimination, sexism, women in construction, culture, and work-life integration. The findings showed that while country-specific studies were common, comparative and multi-contextual analyses remained limited. The results indicate that research has primarily focused on barriers and constraints. Studies on women’s career advancement, resilience, and experiences of success were relatively scarce. Overall, this study maps key research trends, identifies gaps in the existing literature, and offers directions for future research to develop a more comprehensive understanding of gender inequalities in the construction industry.

https://doi.org/10.31462/jcemi.2026.640

Volkan Arslan Şefik Hakan Papila

Despite the ambitious 2053 net-zero targets, the Turkish construction regulatory framework remains predominantly focused on operational energy efficiency, leaving a critical hidden carbon gap regarding embodied emissions. This study evaluates the misalignment between national codes (BEP-TR) and international whole-life carbon principles by analyzing the material-related carbon gap of a typical residential building in Zonguldak. Using verified Bill of Quantities (BoQ) data, a scenario-based embodied carbon assessment was conducted across three trajectories: Pessimistic (Business-as-Usual), Baseline (Standard Practice), and Optimistic (Low-Carbon Transformation). he results reveal that current regulations ignore 249.0 kgCO₂e/m² of upfront embodied carbon in a standard code-compliant building, increasing to 256.0 kgCO₂e/m² when scenario-based end-of-life emissions (C1–C4) are included. The structural system alone accounts for 71.4% of these unregulated upfront emissions. Scenario analysis demonstrates an approximately 48.0% reduction potential between the pessimistic and optimistic supply-chain trajectories, confirming that material selection is as critical as design geometry in reducing embodied carbon. Notably, the Optimistic Scenario indicates that utilizing existing low-carbon technologies could reduce the upfront embodied carbon by 32.2% compared with the baseline without altering architectural practices. However, a readiness paradox exists in which advanced materials produced for EU CBAM compliance are not sufficiently incentivized in the domestic market due to legislative inertia. This study concludes by proposing a phased regulatory roadmap that begins with embodied carbon disclosure and progressively integrates mandatory WLCA reporting, thereby aligning the Turkish construction sector with the EU EPBD and improving economic resilience against future carbon taxes.

https://doi.org/10.31462/jcemi.2026.661

Caner YETİŞ Merve TUNA KAYILI

Medium density fiberboard (MDF) production creates environmental burdens mainly through resin consumption, energy intensive pressing, and material losses. This study developed a Grey Wolf Optimizer (GWO) enhanced machine learning framework to predict global warming potential (GWP) and primary energy demand (PED) using directly controllable MDF process variables. Resin ratio, waste ratio, press temperature, press pressure, and press time were used as inputs, while GWP and PED were defined as outputs. A dataset of 60 production scenarios was used to train four GWO optimized models: CatBoost, Gradient Boosting Machine (GBM), Random Forest (RF), and Extra Trees Regressor (ETR). Model performance was evaluated using R, R², RMSE, RMSRE, RRMSE, MAE, and MAPE. GWO CatBoost achieved the best GWP prediction performance with test values of R² = 0.9467, RMSE = 10.81, MAE = 8.31, and MAPE = 1.46%, while GWO GBM provided the highest PED accuracy with R² = 0.9214, RMSE = 173.92, MAE = 153.23, and MAPE = 2.22%. SHAP analysis identified resin ratio as the dominant predictor for both outputs, followed by press time and press temperature, while waste ratio and press pressure had smaller but meaningful effects. ICE results showed that higher resin content and more energy intensive pressing conditions generally increased GWP and PED. Overall, the proposed framework provides a rapid and explainable decision support tool for environmentally informed MDF process optimization.

https://doi.org/10.31462/jcemi.2026.691

Zheng Ma Mirosław J. Skibniewski Zhen-Song Chen

The architecture, engineering, and construction (AEC) sector has achieved substantial digitalization through building information models, common data environments, digital twins, platforms, and large language model-based assistants. Yet project knowledge remains fragmented across data structures, software systems, organizations, and lifecycle stages, limiting its reuse, governance, and translation into coordinated action. This paper argues that the next bottleneck in AEC is not the lack of digital tools but the lack of an infrastructure capable of transforming fragmented project artifacts into governed and executable knowledge assets. Drawing on a critical synthesis of research on AEC fragmentation, digitalization, knowledge governance, and autonomous agents, the paper introduces the concept of agent-native infrastructure in AEC. It defines such infrastructure as a governable socio-technical architecture that couples semantic knowledge organization, intelligence mediation, tool interaction, and policy-constrained workflow execution. The paper reframes AEC digitalization as an infrastructure problem centered on knowledge executability, explains why existing pathways focused on connection, representation, and visibility have not yet produced governable execution at scale, and outlines a research agenda on technical development, human-AI collaborative systems, organizational knowledge management, and knowledge co-production between research and practice.

https://doi.org/10.31462/jcemi.2026.721

Muhammed Ali Çolak Gürkan Emre Gürcanlı

The construction industry faces growing exposure to uncertainty and volatility, particularly under conditions of armed conflict and geopolitical disruption. While existing research largely concentrates on post-conflict reconstruction, limited attention has been devoted to projects that continue during active warfare, where risk factors remain fluid and continuously evolving. This study examines the distinct risks associated with construction projects executed during wartime, identifying their typologies across recent conflict contexts and developing a structured classification framework to support decision-making under extreme uncertainty. Drawing upon an extensive literature mapping and an in-depth case analysis of the Moscow Headquarters Project—undertaken amid the Russia–Ukraine conflict—the research systematizes wartime risks into six domains: political and security, supply chain, financial and contractual, operational and workforce, technological, and environmental. The framework highlights how contemporary conflicts increasingly extend beyond kinetic warfare to include prolonged economic sanctions, regulatory isolation, and technological embargoes, all of which reshape the landscape of construction risk. By bridging theoretical insights and empirical observation, this study advances the understanding of construction project resilience under contingent circumstances, offering a conceptual basis for adaptive management strategies and risk-informed decision-making in conflict-affected environments.

https://doi.org/10.31462/jcemi.2026.617

Maryam Izadbakhsh Farid Ghannadpour Morteza Bagherpour Mohammad Mahdavi Mazdeh

Many construction organizations struggle with limited resources and tight budgets for project execution. These firms often rely on non-renewable resources of inconsistent quality or perform summary tasks in ways that reduce costs but sacrifice accuracy. Consequently, the industry requires a quality management system that tackles resource selection, task performance, and cost control in a single framework. Prior studies addressed quality management in segmented sections of the project supply chain but rarely treated the entire sequence as an integrated process. This study proposes an optimal model that views each construction project as a continuous supply chain under two scenarios: one with a discount policy applied to selected suppliers and one without. The model covers every phase, from project initiation and resource procurement through task scheduling and final delivery. It aims to improve the quality of summary task performance, select non-renewable resources that meet specified thresholds, and reduce overall logistics costs. In this formulation, the quality metric for each summary task corresponds to the level of technology applied; higher‐quality outcomes require more advanced equipment or methods. The two-objective model is solved using the augmented ε-constraint method, and the results are illustrated with a numerical example. Sensitivity analysis examines how changes in cost coefficients and quality thresholds affect optimal decisions. The model's performance is then validated by varying key parameters. The findings offer decision-makers a clear guide for balancing cost and quality, choosing suppliers, allocating resources, and scheduling tasks to achieve consistent, high-quality project outcomes.

https://doi.org/10.31462/jcemi.2026.677

Fatemeh Mostofi Onur Behzat Tokdemir Vedat Toğan

Earned value management (EVM) is critical for monitoring financial and operational performance in large-scale projects. However, its implementation is limited by the complexity of modern construction workflows. Particularly, modeling the cost performance index (CPI) to evaluate the efficiency of activity delivery plans has led to the development of various relational learning models. This study addresses the inability of current relational models to handle high-resolution construction data with dense WBS-based activity networks, where uncontrolled information flow and class imbalance reduce model expressiveness. In this study, we introduce a CPI-specific activity learning mechanism that forecasts CPI using a gated approach to regulate information propagation in dense spatiotemporal data. To contextualize the proposed information modeling framework, this study draws upon a large-scale dataset derived from 77 construction progress reports collected over an 18-month period, which were cleaned and encoded into 52 unique delivery-week categories for model training. We demonstrate that the proposed gated architectures outperform previously proposed attention-based mechanisms. Across eight timeframes of real-world construction datasets, GatedGNN shows superior performance, improving accuracy by 14% and F1 scores by 13% over GCN and GAT models. These improvements enable more reliable CPI classification (efficient vs. inefficient) and support timely corrective decision-making by producing activity-level risk flags that can guide managerial review of resource allocation, schedule coordination, and WBS-specific cost deviations. By improving fidelity in class distribution and managing relational complexity, gated GNNs enable earlier and more accurate identification of inefficiencies. This facilitates proactive decision-making, allowing engineering managers to intervene before performance deviations escalate.

https://doi.org/10.31462/jcemi.2026.680

Feride Çiğdem KARA DÜLGER MERVE TUNA KAYILI

The growing demand for energy-efficient buildings has led to the widespread adoption of Passive House (PH) standards, emphasizing operational energy reduction through optimized insulation and airtight construction. However, the embodied environmental impacts of building materials remain a critical concern. This study presents a Life Cycle Assessment (LCA) of the Gaziantep Ecological House (GEH), Turkey’s first PH-certified building, focusing on both operational and embodied carbon emissions across its construction phases. The study evaluates the A1-A3 (material extraction, processing, and manufacturing) and A4 (transportation) stages, incorporating various wall construction scenarios to assess their environmental trade-offs. The results reveal that while high-performance insulation materials contribute to energy efficiency, their embodied carbon emissions can significantly impact sustainability. Among alternative wall materials, brick-based scenarios exhibited the highest global warming, acidification, and eutrophication potentials due to high-temperature kiln-firing and fossil fuel dependency. Conversely, adobe (earthen) walls demonstrated the lowest environmental impact across multiple categories, reinforcing their viability as a low-carbon construction material. However, challenges related to moisture resistance and structural performance require further investigation. Additionally, hempcrete, often perceived as an environmentally friendly material, showed higher-than-expected ozone depletion potential due to its cementitious binder content, highlighting the trade-offs between carbon sequestration benefits and secondary environmental burdens. The study also underscores the critical role of transportation emissions (A4 phase), where locally sourced materials such as adobe and autoclaved aerated concrete (AAC) significantly reduced transport-related carbon emissions compared to imported alternatives. Overall, this research emphasizes the need for a balanced approach to sustainable construction, integrating both operational energy efficiency and embodied carbon reduction. Future studies should explore hybrid material strategies, bio-based insulation alternatives, and circular economy principles to further minimize lifecycle environmental impacts. By integrating LCA-driven decision-making into early-stage building design, policymakers, architects, and engineers can optimize passive house construction for long-term environmental sustainability.

https://doi.org/10.31462/jcemi.2026.556