Yutaro Yamaguchi

Packing non-zero A-paths via matroid matching

A -labeled graph is a directed graph G in which each edge is associated with an element of a group by a label function :E(G). For a vertex subset AV(G), a path (in the underlying undirected graph) is called an A-path if its start and end vertices belong to A and does not intersect A in between, and an A-path is called non-zero if the ordered product of the labels along the path is not equal to the identity of . Chudnovsky etal. (2006) introduced the problem of packing non-zero A-paths and gave a minmax formula for characterizing the maximum number of vertex-disjoint non-zero A-paths. In this paper, we show that the problem of packing non-zero A-paths can be reduced to the matroid matching problem on a certain combinatorial matroid, and discuss how to derive the minmax formula based on Lovsz idea of reducing Maders S-paths problem to matroid matching.

Cyber security analysis of power networks by hypergraph cut algorithms

This paper presents exact solution methods for analyzing vulnerability of electric power networks to a certain kind of undetectable attacks known as false data injection attacks. We show that the problems of finding the minimum number of measurement points to be attacked undetectably reduce to minimum cut problems on hypergraphs, which admit efficient combinatorial algorithms. Experimental results indicate that our exact solution methods run as fast as the previous methods, most of which provide only approximate solutions. We also present the outline of an algorithm for enumerating all small cuts in a hypergraph, which can be used for finding vulnerable sets of measurement points.

Packing

Maders disjoint S-paths problem is a common generalization of non-bipartite matching and Mengers disjoint paths problems. Lovasz (1980) suggested a polynomial-time algorithm for this problem through a reduction to matroid matching. A more direct reduction to the linear matroid parity problem was given later by Schrijver (2003), which leads to faster algorithms. As a generalization of Maders problem, Chudnovsky, Geelen, Gerards, Goddyn, Lohman, and Seymour (2006) introduced a framework of packing non-zero A-paths in group-labelled graphs, and proved a min-max theorem. Chudnovsky, Cunningham, and Geelen (2008) provided an efficient combinatorial algorithm for this generalized problem. On the other hand, Pap (2007) introduced a framework of packing non-returning A-paths as a further genaralization. In this paper, we discuss a possible extension of Schrijvers reduction technique to another framework introduced by Pap (2006), under the name of the subgroup model, which apparently generalizes but in fact is equivalent to packing non-returning A-paths. We provide a necessary and sufficient condition for the groups in question to admit a reduction to the linear matroid parity problem. As a consequence, we give faster algorithms for important special cases of packing non-zero A-paths such as odd-length A-paths. In addition, it turns out that packing non-returning A-paths admits a reduction to the linear matroid parity problem, which leads to the quite efficient solvability, if and only if the size of the input label set is at most four.

A

Influence maximization is a well-studied problem of finding a small set of highly influential individuals in a social network such that the spread of influence under a certain diffusion model is maximized. We propose new diffusion models that incorporate the time-decaying phenomenon by which the power of influence decreases with elapsed time. In standard diffusion models such as the independent cascade and linear threshold models, each edge in a network has a fixed power of influence over time. However, in practical settings, such as rumor spreading, it is natural for the power of influence to depend on the time influenced. We generalize the independent cascade and linear threshold models with time-decaying effects. Moreover, we show that by using an analysis framework based on submodular functions, a natural greedy strategy obtains a solution that is provably within

Maximizing Time-Decaying Influence in Social Networks

1-1/e

Finding a Path in Group-Labeled Graphs with Two Labels Forbidden

of optimal. In addition, we propose theoretically and practically fast algorithms for the proposed models. Experimental results show that the proposed algorithms are scalable to graphs with millions of edges and outperform baseline algorithms based on a state-of-the-art algorithm.

Making Bipartite Graphs DM-Irreducible

This paper presents exact solution methods for analyzing vulnerability of electric power networks to a certain kind of undetectable attacks known as false data injection attacks. We show that the problems of finding the minimum number of measurement points to be attacked undetectably reduce to minimum cut problems on hypergraphs, which admit efficient combinatorial algorithms. Experimental results indicate that our exact solution methods run as fast as the previous methods, most of which provide only approximate solutions. We also present an algorithm for enumerating all small cuts in a hypergraph, which can be used for finding vulnerable sets of measurement points.

Antimatroids induced by matchings

The parity of the length of paths and cycles is a classical and well-studied topic in graph theory and theoretical computer science. The parity constraints can be extended to the label constraints in a group-labeled graph, which is a directed graph with a group label on each arc. Recently, paths and cycles in group-labeled graphs have been investigated, such as finding non-zero disjoint paths and cycles.