A Path Enumeration Algorithm for Strategic System Optimal Dynamic Traffic Assignment

Accounting for sources of uncertainty in the Dynamic Traffic Assignment (DTA) problem makes an already complex problem even more difficult. This work examines a Strategic System Optimal DTA (StrSODTA) model where optimal path proportions are assigned to minimize the Total System Travel Time (TSTT) across a range of stochastic demand scenarios. However, the LP proposed by Waller et al. (2013) faces significant challenges in terms of computational complexity. This works aims to propose a novel solution algorithm to improve the computational efficiency of the LP formulation for the StrSODTA problem. Namely, we develop a tailored path enumeration strategy to reduce the number of constraints in the aforementioned LP formulation and derive a lower bound on the optimal expected TSTT to measure the progress of the solution algorithm. Results on two benchmark traffic assignment networks are examined and we report that the proposed solution algorithm is able to significantly improve the tractability of the StrSODTA problem.