L-Infinity Optimization in Tropical Geometry and Phylogenetics
نویسنده
چکیده
We investigate uniqueness issues that arise in l∞-optimization to linear spaces and Bergman fans of matroids. For linear spaces, we give a polyhedral decomposition of R based on the dimension of the set of l∞-nearest neighbors. This implies that the l∞-nearest neighbor in a linear space is unique if and only if the underlying matroid is uniform. For Bergman fans of matroids, we show that the set of l∞-nearest points is a tropical polytope and give an algorithm to compute its tropical vertices. A key ingredient here is a notion of topology that generalizes tree topology. These results have practical implications for distance-based phylogenetic reconstruction using the l∞-metric. We analyze the possible dimensions of the set of l∞-nearest equidistant tree metrics to an arbitrary dissimilarity map and the number of tree topologies represented in this set. For both 3 and 4-leaf trees, we decompose the space of dissimilarity maps relative to the tree topologies represented.
منابع مشابه
THE SPACE OF n POINTS ON A TROPICAL LINE IN d-SPACE
The tropical semiring (R,min,+) has enjoyed a recent renaissance, owing to its connections to mathematical biology as well as optimization and algebraic geometry. In this paper, we investigate the space of n points on a tropical line in d-space, which is interpretable as the space of n-species phylogenetic trees. This is equivalent to the space of n × d matrices of tropical rank two, a simplici...
متن کاملTropical Principal Component Analysis and its Application to Phylogenetics
Principal component analysis is a widely-used method for the dimensionality reduction of a given data set in a high-dimensional Euclidean space. Here we define and analyze two analogues of principal component analysis in the setting of tropical geometry. In one approach, we study the Stiefel tropical linear space of fixed dimension closest to the data points in the tropical projective torus; in...
متن کاملAn approach to rank efficient DMUs in DEA based on combining Manhattan and infinity norms
In many applications, discrimination among decision making units (DMUs) is a problematic technical task procedure to decision makers in data envelopment analysis (DEA). The DEA models unable to discriminate between extremely efficient DMUs. Hence, there is a growing interest in improving discrimination power in DEA yet. The aim of this paper is ranking extreme efficient DMUs in DEA based on exp...
متن کاملTropical Convexity
The notions of convexity and convex polytopes are introduced in the setting of tropical geometry. Combinatorial types of tropical polytopes are shown to be in bijection with regular triangulations of products of two simplices. Applications to phylogenetic trees are discussed.
متن کاملRanking efficient DMUs using the infinity norm and virtual inefficient DMU in DEA
In many applications, ranking of decision making units (DMUs) is a problematic technical task procedure to decision makers in data envelopment analysis (DEA), especially when there are extremely efficient DMUs. In such cases, many DEA models may usually get the same efficiency score for different DMUs. Hence, there is a growing interest in ranking techniques yet. The purpose of this paper is ra...
متن کامل