Title
The Power-Compositions Determinant and Its Application to Global Optimization
Abstract
Let C(n,p) be the set of p-compositions of an integer n, i.e., the set of p-tuples $\alpha=(\alpha_1,\ldots,\alpha_p)$ of nonnegative integers such that $\alpha_1+\cdots+\alpha_p=n$. The main result of this paper is an explicit formula for the determinant of the matrix whose entries are $\alpha^{\beta}=\alpha_1^{\beta_1}\cdots\alpha_p^{\beta_p}$ where $\alpha,\beta\in C(n,p)$. The formula shows that the determinant is positive and has a nice factorization. As an application, it is shown that the polynomials $p_\alpha(x)=(\alpha_1x_1+\cdots+\alpha_px_p)^n$ with $\alpha\in C(n,p)$ form a basis of the vector space Hn[x1, . . . ,xp] of homogeneous polynomials of degree n in p variables. The result is of interest in the context of global optimization because it allows an explicit representation of polynomials as a difference of convex functions.
Year
DOI
Venue
2001
10.1137/S0895479800369141
SIAM Journal on Matrix Analysis and Applications
Keywords
Field
DocType
composition,homogeneous polynomial,Vandermonde's determinant,difference of convex functions
Alpha (ethology),Integer,Vector space,Combinatorics,Matrix (mathematics),Mathematical analysis,Convex function,Homogeneous polynomial,Factorization,Vandermonde matrix,Mathematics
Journal
Volume
Issue
ISSN
23
2
0895-4798
Citations 
PageRank 
References 
2
0.55
0
Authors
2
Name
Order
Citations
PageRank
Josep M. Brunat1425.52
Antonio Montes220519.68