Method for Scenario Selection and Probability Adjustment for Reliability and Active Traffic Management Analysis in a Highway Capacity Manual Context

This paper presents an optimization-based probability adjustment approach that enables an analyst to minimize the error and bias in estimating freeway reliability performance measures by using a small sample of reliability scenarios. The freeway facilities travel time reliability methodology proposed for the Highway Capacity Manual 2010 can produce up to 24,000 scenarios to be evaluated for estimating the travel time distribution. The methodology uses a deterministic scenario generation approach to account for any possible operational condition of freeway facilities. In addition to processing time considerations, the large number of scenarios poses analytical challenges because (a) a detailed assessment and scrutiny of individual scenarios cannot be performed and (b) the customized selection of active traffic management is infeasible. The proposed method allows the analyst to adjust the selected scenario probabilities to estimate the real-world freeway performance measure better with a manual biased sample. The biased sampling probability adjustment method is applied to two real-world case study examples. These illustrate that the population travel time distribution can be approximated adequately through the probability adjusted sample. Findings from this research have implications for proposed reliability and active traffic management methodologies in the Highway Capacity Manual 2010. The number of scenarios can be greatly reduced from several thousand to less than 100 while maintaining the shape of the reliability distribution and key performance measures of the scenario population.