How incidents impact congestion on roadways: A queuing network approach

Motivated by the need for transportation infrastructure and incident management planning, we study traffic density under non-recurrent congestion. This paper provides an analytical solution approximating stationary distribution of in roadways where deterioration service occurs unpredictably. The proposed generalizes a queuing model discussed literature to long segments that are not space-homogeneous. We compare single tandem approaches different lengths verify whether each is appropriate. A single-queue approach works sufficiently well with similar behavior across space. In contrast, tandem-queue more appropriately describes sections having distinct characteristics. These models have comparable fit ones generated using lognormal distribution. However, they also interpretable parameters, directly connecting congestion dynamics roadway behavior. general, adaptable, tractable, thus being instrumental management.