The Control Complexity of $r$-Approval: from the Single-Peaked Case to the General Case

Abstract: We investigate the complexity of $r$-Approval control problems in $k$-peaked elections, where at most $k$ peaks are allowed in each vote with respect to an order of the candidates. We show that most NP-hardness results in general elections also hold in k-peaked elections even for $k=2,3$. On the other hand, we derive polynomial-time algorithms for some problems for $k=2$. All our NP-hardness results apply to Approval and sincere-strategy preference-based Approval as well. Our study leads to many dichotomy results for the problems considered in this paper, with respect to the values of $k$ and $r$. In addition, we study $r$-Approval control problems from the viewpoint of parameterized complexity and achieve both fixed-parameter tractability results and W-hardness results, with respect to the solution size. Along the way exploring the complexity of control problems, we obtain two byproducts which are of independent interest. First, we prove that every graph of maximum degree 3 admits a specific 2-interval representation where every 2-interval corresponding to a vertex contains a trivial interval and, moreover, 2-intervals may only intersect at the endpoints of the intervals. Second, we develop a fixed-parameter tractable algorithm for a generalized $r$-Set Packing problem with respect to the solution size, where each element in the given universal set is allowed to occur in more than one r-subset in the solution.