The shape of the solution set for systems of interval linear equations with dependent coefficients

被引:16
作者
Alefeld, G
Kreinovich, V
Mayer, G
机构
[1] Univ Karlsruhe, Inst Angew Math, D-76128 Karlsruhe, Germany
[2] Univ Texas, Dept Comp Sci, El Paso, TX 79968 USA
[3] Univ Rostock, Fachbereich Math, D-18051 Rostock, Germany
关键词
linear interval equations; solution set; symmetric linear equation systems; systems with dependent coefficients; interval analysis; semialgebraic sets; quantifier elimination;
D O I
10.1002/mana.19981920103
中图分类号
O1 [数学];
学科分类号
0701 ; 070101 ;
摘要
A standard system of interval linear equations is defined by Ax = b, where A is an m x n coefficient matrix with (compact) intervals as entries, and b is an m-dimensional vector whose components are compact intervals. It is known that for systems of interval linear equations the solution set, i.e., the set of all vectors a for which Ax = b for some A is an element of A and b is an element of b, is a polyhedron. In some cases, it makes sense to consider not all possible A is an element of A and b is an element of b, but only those A and b that satisfy certain linear conditions describing dependencies between the coefficients. For example, if we allow only symmetric matrices A (a(ij) = a(ji)), then the corresponding solution set becomes (in general) piecewise-quadratic. In this paper, we show that for general dependencies, we can have arbitrary (semi)algebraic sets as projections of solution sets.
引用
收藏
页码:23 / 36
页数:14
相关论文
共 17 条
[1]  
Alefeld G, 1996, APPL OPTIMIZAT, V3, P61
[2]  
Alefeld G., 1983, INTRO INTERVAL COMPU
[3]  
ALEFELD G, 1997, SIAM J MATRIX ANAL A, V18
[4]  
ALEFELD G, 1995, SIAM J MATRIX ANAL A, V16
[5]  
ARNOLD VI, 1983, GEOMTRICAL METHODS T
[6]  
Barmish B.R., 1994, New Tools for Robustness of Linear Systems
[7]  
BEECK H, 1972, COMPUTING, V10
[8]  
HARTFIEL DJ, 1980, NUMER MATH, V35
[9]  
JANSSON C, 1991, COMPUTING, V46
[10]  
Kearfott R. B., 1996, Applications of Interval Computations