Evaluating the effectiveness of an adaptive traffic control system is extremely difficult due to the stochastic and dynamic properties of traffic, and the complexity of an adaptive system. Many adaptive control systems show promise that significant improvements in traffic control performance are possible. However, they often fail to provide reproducible and statistically valid documentation of reported performance improvements. This paper proposes formal evaluation procedures, model requirements, measures of effectiveness, and an assessment framework to systematically evaluate adaptive control methods.
Samenvatting