Performance measurement indices for simulated construction operations

This paper describes the development of indices useful in automating the ex perimentation process of a computer simulation. Simulation methodologies ha ve been developed to model construction systems, but most of these systems require the experimentation process to be carried out manually. In achievin g optimum performance, one has to repeat an exhaustive number of experiment s. The indices can be used to automate this process, as they are indicators of bottlenecks in the system that can be tracked through the simulation ou tput. They are based upon user-defined performance guidelines for the resou rces. Where a performance index falls outside the acceptable range, remedia l action may be taken. Belief networks, a probabilistic form of artificial intelligence, were used to automate the analysis of the indices to determin e the most likely causal factor of poor performance.