Abstract (eng)
In this Master’s Thesis the bi-objective Ring Star Problem is solved. Different to the
general formulation of the RSP, the assignment problem was extended by User Equilibrium
constraints. In that way the assignment to ring is not just based on the distance
to it, but also on the "service quality", respectively the incoming flow to a node on the
ring. As a solution approach the NSGA-II was implemented, with nested Clarke&Wright
savings algorithm and Frank Wolfe algorithm. The savings algorithm was used for solving
a TSP for each solution. For the approximation of the User Equilibrium the Frank
Wolfe algorithm was applied, after the assignment problem was transferred into a network
flow problem. Since no benchmark solution for such a formulation of a RSP exists,
test instances were generated and enumerated. These were used for the evaluation of the
quality of the implemented solution method.