The Minimum Clique Routing Problem on Cycles

Abstract: In the Minimum Clique Routing Problem on Cycles \textsc{MCRPC} we are given a cycle together with a set of demands (weighted origin-destination pairs) and the goal is to route all the pairs minimizing the maximum weighted clique of the intersection graph induced by the routing. The vertices of this graph are the demands with their corresponding weights and two demands are adjacent when their routes share at least one arc. In this work we are not only interested in the \textsc{MCRPC} but also in two natural subproblems. First, we consider the situation where the demands are disjoint, in the sense that every two demands do not share any of their corresponding ends. Second, we analyze the subproblem where the weights of the routes are all equal. We first show that the problem is NP-complete even in the subproblem of disjoint demands. For the case of arbitrary weights, we exhibit a simple combinatorial 2-approximation algorithm and a $\frac{3}{2}$-approximation algorithm based on rounding a solution of a relaxation of an integer linear programming formulation of our problem. Finally, we give a Fixed Parameter Tractable algorithm for the case of uniform weights, whose parameter is related to the maximum degree of the intersection graph induced by any routing.