Development of a solution algorithm to solve the user equilibrium traffic assignment problem with soft link capacity constraints

2021 ◽  
Vol 23 (5) ◽  
pp. 81-88
Author(s):  
Seungkyu Ryu
Keyword(s):  
Assignment Problem ◽  
Traffic Assignment ◽  
User Equilibrium ◽  
Solution Algorithm ◽  
Capacity Constraints ◽  
Link Capacity ◽  
Traffic Assignment Problem

scholarly journals Traffic Assignment Problem with Strict Capacity Constraints

2020 ◽  
Author(s):  
Ghazaleh Azimi
Keyword(s):  
Assignment Problem ◽  
Flow Problem ◽  
Traffic Assignment ◽  
Transportation Network ◽  
Capacity Constraints ◽  
Movement Direction ◽  
Assignment Problems ◽  
Link Capacity ◽  
Traffic Assignment Problem ◽  
Cost Flow

To improve the performance of the transportation network operation, traffic assignment problems require more reliable results. Disregard the solution method, one factor that leads to more accurate results in traffic assignment problems is considering the link capacity constraints. In traffic assignment problems, techniques such as the Frank-Wolfe (FW) algorithm has extensive applications, but not considering link capacity in these methods results in inaccurate outputs. In every iteration of FW, move direction is obtained by solving the minimum cost flow problem for the network without considering the capacity constraints, which is named the all or nothing assignment algorithm. On the contrary, there is another algorithm called the Out of kilter (OFK) algorithm that solves the traffic assignment problem by minimizing the cost flow problem in networks with link capacity constraints. In this dissertation, the movement direction of the FW algorithm is obtained by the OFK algorithm. Therefore, in every iteration, link flows are lower than or equal to their capacities. In this method, we assume that the network has a circulation form, meaning that all origins are pure origin, and all destinations are the pure destination. Moreover, the results of the proposed algorithm for the Hearn and Sioux Falls transportation networks are presented and compared to the condition that link capacities are considered infinite.

Sign in / Sign up

Export Citation Format

Share Document