Short communicationA note on maximizing the minimum voter satisfaction on spanning trees
Introduction
The topic of finding a minimum spanning tree is a classical problem in Combinatorial Optimization. Having determined a minimum spanning tree, the natural question may arise how to “fairly” assign the costs of such a tree to individuals. This issue has been widely studied since Claus and Kleitman (1973), e.g., by Bird (1976), Bogomolnaia and Moulin (2009) and Dutta and Kar (2004).
In this paper, however, we focus on another aspect of fairness in connection with spanning trees. In our framework monetary costs are not taken into consideration. Instead, our approach is based on individuals’ preferences over the edges of a graph. Given such preferences, in Darmann et al. (2008) the quality of methods that fairly, i.e., socially acceptably, construct a spanning tree is analyzed. In a companion paper of Darmann et al. (2009), the focus is laid on the computational complexity of methods to fairly (in a maximin sense) construct a spanning tree. In this paper, we study the computational complexity of another method to fairly construct a spanning tree and therewith answer an open question stated in Darmann et al. (2009).
Our framework covers many applications in which a network needs to be installed, e.g., when countries need to agree on transnational traffic systems or oil pipelines, or when homeowners have diverging opinions about the specific links in a sewage or water network that needs to be constructed.
In the Combinatorial Optimization literature maximin fairness can be found in terms of maximizing the minimum of concepts such as utility, costs, time, etc. However, in Social Choice Theory the maximin approach is a well-known concept of formalizing fairness, originally discussed by Rawls (1972); for extensive studies of social choice rules see, e.g., Brams and Fishburn (2002), Nurmi (2007) and Saari (1995). In this paper, we follow the maximin approach based on different social choice rules proposed in Darmann et al. (2009) and show that the maximin voter satisfaction problem (MMVS) under choose- elections is -hard for each fixed .
Section snippets
Problem formulation
In this work denotes an undirected graph, with a set of vertices and a set of edges. Let be the number of vertices and be the number of edges of . We call a subset a spanning tree of if the graph is connected and acyclic. Let be the set of spanning trees of . A binary relation on a set is called
- •
complete if , or , and
- •
transitive if , .
A relation is called weak order if it is complete and transitive.
In our framework we are
MMVS under choose- elections
Theorem 1 MMVS under choose-2 elections is strongly -complete.
Proof Let be an arbitrary instance of 3-SAT with a set of variables and a set of clauses over . Accordingly, the set of literals is . From we create an instance of MMVS under a choose-2 election by building a graph (see Fig. 1) and a voter preference profile (see Table 1) as follows. In order to create , for every variable we introduce a triangle of edges and an edge
References (14)
- et al.
Voting procedures
- et al.
Maximizing the minimum voter satisfaction on spanning trees
Mathematical Social Sciences
(2009) - et al.
Cost monotonicity, consistency and minimum cost spanning tree games
Games and Economic Behavior
(2004) - et al.
Approximating min–max (regret) versions of some polynomial problems
On cost allocation for a spanning tree: a game theoretic approach
Networks
(1976)- et al.
Sharing the cost of a minimal cost spanning tree: beyond the folk solution
Games and Economic Behavior
(2009) - et al.
Approval Voting
(1983)
Cited by (5)
How hard is it to tell which is a Condorcet committee?
2013, Mathematical Social SciencesCommittee selection under weight constraints
2012, Mathematical Social SciencesCitation Excerpt :In this literature complexity analysis and solution methods discussed in Operations Research and Computer Science are linked with aspects from Social Choice Theory such as the determination of (social) preferences or of (a set of) candidates based on rankings of candidates. Efforts along these lines were e.g. made in Klamler and Pferschy (2007), Darmann et al. (2011, 2009, 2010) and Lu and Boutilier (2010). Second, in political science various approaches have been developed to determine a set of winners for the election of multiple winners in legislative bodies.
Popular spanning trees
2013, International Journal of Foundations of Computer ScienceFair solutions for some multiagent optimization problems
2013, Autonomous Agents and Multi-Agent SystemsFinding socially best spanning trees
2011, Theory and Decision