Agent embedded simulation modeling framework for construction engineering and management applications

The presented research addresses concerns regarding the precision and proficiency of how current construction projects are planned, estimated, and executed. A new approach for construction engineering and management problem-solving is presented. This new problem-solving approach makes use of a simulation modeling environment with embedded intelligent agents to both capture the beneficial properties of simulation modeling with respect to the precision in which construction problems are represented and increase the proficiency with which those problems are solved.

The research focuses on the development of an agent structure that can be deployed within the simulation environment. The agent structure developed provides a means through which autonomous intelligent objects can be employed within the simulation modeling environment to assume roles and possess knowledge otherwise demanded by simulation modelers. The culmination of this development effort has resulted in a generic agent structure that can be instructed to assume a variety of roles and responsibilities during simulation runtime. The agent structure also possesses the flexibility to make use of varying intelligence facilities such as belief networks, feed-forward back propagation neural networks, algorithms, and rule bases. The second component of this research focuses on the development of an agent framework that allows several agents to be employed within the same simulation modeling environment with sufficient organization and governance to ensure that the total influence by the agents on the simulation model follows a common good.