Proactive planning and control of construction productivity and cost using time series data

Effective planning and control of productivity and cost is a significant factor to achieving successful management of construction projects. It is, however, challenging for the uncertainties and dynamics involved in construction processes. Although the construction industry has experienced such a challenging situation for a long time, improvements have been slow in solving the challenges. Construction projects are subject to changes and so are the processes of project activities. Accordingly, construction productivity and cost of an on-going activity evolves dynamically and stochastically over its progress. In addition, construction cost changes over time and is often affected by market conditions, which result in variations in its change. The construction industry, however, has experienced difficulties in dealing with uncertainties and dynamics. In order to analyze the dynamic changes in productivity and cost, practitioners inevitably need to update existing estimates and plans repeatedly for the planning and control of the future processes. Overall, it is essential to monitor and evaluate productivity frequently and regularly so as to update existing plans and implement the updated plans in a timely manner. Current approaches lack capabilities to deal with such a dynamic and stochastic change in a timely manner for the planning and control of productivity and cost. To improve the current state-of-art of planning and control, this research developed methods for continual forecasting and updating construction productivity and cost. The proposed method predicts productivity of an on-gong activity on a weekly basis using contemporary productivity data from the activity. Using univariate time series modeling, this method provides a more appropriate and effective methodology for the development of forecasting models and making the best use of such time series data. The proposed method was validated using multiple real cases within the context of reinforced concrete work. Validation proved that the models are effective in predicting short-term productivity trend. In addition to productivity prediction, this research exploits construction cost. Noting that the construction cost index provides a good means for forecasting the future construction costs, this research developed and validated cost forecasting models for the prediction of cost index on a monthly basis. Validation revealed that the proposed models produce reliable results for both short-term and long-term prediction. The proposed univariate time series model and dynamic regression models outperform existing models in accuracy. Overall, the proposed models for both productivity and cost proved to be feasible and practical for their reasonable accuracy and practical implementation. Using the proposed models, practitioners can objectively and accurately predict construction productivity and cost in a timely manner. Thereby, practitioners can make more reliable decisions and implement the decisions in a more effective way while managing projects. Overall, it is envisioned that the new approach can help practitioners to be more proactive so that they can reduce potential risk in the future construction processes, improve efficiency of collaboration and coordination among participants involved in a project, enhance updating on-site plans and schedules, and optimize resource use. It is also conceivable that the proposed method can be integrated with supportive technologies for automatic data collection and analysis, so it will potentially support the real-time planning and control of construction processes in the future.