The proposed research will attempt to employ cluster analysis techniques to answer questions related to traffic operations such as:

• Can we discriminate between recurrent congestion and outliers?
• For a given corridor with recurrent incidents, can we identify a typical day for operations planning purposes?
• How do we define a corridor-specific typical peak period while considering a variety of demand management initiatives, peak hour spreading, dynamic vehicle routing, connected vehicles and variable weather conditions?
• Can we identify how many time periods VDOT needs to plan for in terms of resource and congestion management?

  In addition, we will establish a methodology to identify the most relevant variables for VDOT and to identify the most suitable clustering algorithm for VDOT needs. The established methodology will be tested on a pilot corridor and the results will be documented.