Utilizing Soft Computing Methods in Analyzing Build-Operate-Transfer (BOT) Contracts

An objective Build-Operate-Transfer (BOT) contract evaluation at the conceptual stage, in countries facing budget constraints, will lead to undertaking projects which are anticipated to be viable in the future. An objective analysis of various risk variables and their impact on a BOT project’s future outcome requires study and integration of many likely scenarios into the contract terms, which is complicated and time-consuming. If the process of examining the financial parameters and uncertainties of a BOT project could be automated, this would be a milestone in objective decision-making from various stakeholders’ points of view. A soft computing model would let the user analyze many probable scenarios more accurately. In this study two soft computing methods, artificial neural network (ANN) and gene expression programming (GEP) are applied onto two distinct BOT case studies to illustrate automation of their assessment processes. First a case study of BOT model on dormitory projects in Cyprus is analyzed. An ANN model with correlation coefficient of 0.9064 is developed to model the relationship between important project parameters and risk variables. Significant factors, used in ANN model development, were extracted from sensitivity analysis and Monte Carlo simulation results obtained from conventional spreadsheet data. The resulting consensus based on this model would yield to fair contractual agreements for both the government and the concession company. iv Second financial viability of undertaking a BOT contract for sewer and water projects in California, USA is analyzed. Furthermore by aid of sensitivity analysis, risk parameters are identified. Sensitivity analysis results demonstrated that project construction cost factor determines the financial viability of undertaking a BOT contract. Therefore, reliable construction cost prediction, based on limited information, at early stages of the project planning phase is crucial for development of an objective BOT agreement. This study utilized gene expression programming (GEP) which is a derivative of genetic algorithm (GA) and genetic programming (GP), and developed a prediction model with correlation coefficient of 0.8467 for estimating the construction cost of water and sewer rehabilitation/replacement projects. Contribution of this thesis to knowledge is by exploiting ANN model’s capability to incorporate many scenarios, we developed an automated tool to define concession terms considering potential risks; and by utilizing GEP model ‘s ability to create an explicit equation, we developed a formula for a project construction cost prediction to help improve objective financial appraisal of a BOT project. Author keywords: Public-Private-Partnership; Build-Operate-Transfer; Monte Carlo simulation; Contracts; Cost Estimation; Artificial Neural Network; Gene Expression programming; Dormitory Projects; Water and Sewer Replacement/Rehabilitation Projects.