An Artificial Intelligence Framework for Project Planning and Control using Decision Tree Analysis and Artificial Neural Network

The challenges of traditional project planning tools, leads to pitiable performance of a variety of projects and has great economic impacts. The aim of this study was to provide a framework for the integration of Decision Tree Analysis (DTA) and Artificial Neural Network (ANN) into project planning and control. Random forest decision tree identified important project variables that served as input variables to feed-forward back-propagation neural network machine learning capabilities. Data were gathered from six projects in Lagos and Ogun States south west Nigeria. Most significant project variables for ANN input, were Number of Resource-Constrained (NRes), Project Size, Critical Path Method (CPM) value and Project Direct Cost (DCOST). Project Maximum Remaining Path Length (MPRL), Latest Finish Time (LFT) and Project Total Cost (TCOST) were output variables. Results showed that the three resource-constrained project scheduling displayed best ANN training performance. Percentages of project time overrun were 58.083% for MRPL and 55.988% for LFT while percentage project cost overrun was 20%. Evaluation showed that artificial intelligence tools were more efficient for project planning and control when compared to traditional tools.