Applying distributed artificial intelligence to the prequalification of construction contractors

The construction industry has been criticized, to a large extent, for its slow acceptance and use of technological improvements to plan and execute projects. The changing global environment and the increasing complexity of the industry has created a need for adopting advanced technologies. Computer-based technologies such as artificial intelligence techniques are generating interest as potential aids for decision making in different engineering and management decision domains. Knowledge-based systems have steadily been introduced for different applications in the construction industry. Most of the knowledge-based system applications currently available for the construction industry can be described as a single agent system in which a single body of knowledge is used to solve the entire problem. Such systems require intensive software development and maintenance in addition to large allocated memory. They also lack the ability to learn by themselves, generalize solutions and adequately respond to highly correlated, noisy or previously unseen data. Moreover, single agent systems are not adequate for solving human problems that usually requires the involvement of multiple decision makers. This dissertation illustrates the capabilities of distributed artificial intelligence in representing and using knowledge in the construction industry. The objective of this study is two-fold: (1) to introduce a methodology, that uses the distributed artificial intelligence capabilities, to develop adaptive DSS for solving construction industry problems and (2) to develop an adaptive computerized tool for performing owner-contractor prequalification. These objectives were achieved through the development of the Contractor Selector Decision Support System, CONSEL, an adaptive DSS for prequalifying construction contractors. The problem solving strategy of the developed DSS was distributed among eight problem solvers to mimic the actual prequalification procedure that involves several tasks. Different problem solvers of the system were developed by using the learning capabilities of neural networks and production rules. The distributed artificial intelligence architecture suits the area of contractor prequalification in which several tasks are examined and integrated to arrive at the final decision. A learning subsystem is included to modify its problem solving knowledge through experience. These learning capabilities will constantly improve the system's performance.