CONSTRAINT REASONING BASED ON INTERVAL ARITHMETIC - THE TOLERANCE PROPAGATION APPROACH

被引:82
作者
HYVONEN, E
机构
[1] Technical Research Centre of Finland (VTT), Laboratory for Information Processing, 00340 Helsinki
基金
芬兰科学院;
关键词
D O I
10.1016/0004-3702(92)90005-I
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Interval constraint satisfaction (interval labeling) systems have traditionally been based on local Waltz filtering techniques that cannot in general determine global solutions. In contrast, this paper documents a related technique, tolerance propagation (TP), that generalizes the idea of numerical exact value propagation into interval propagation. In TP, consistency techniques based on the topology of the constraint net can be combined with techniques of interval arithmetic in a new fruitful way. In particular, by TP it is possible to determine global solutions for interval constraint satisfaction problems with arbitrary accuracy and without losing all attractions of simple local computations.
引用
收藏
页码:71 / 112
页数:42
相关论文
共 51 条
[1]  
AIBA A, 1988, 5TH P INT C GEN COMP, P263
[2]   ON INTERVAL ARITHMETIC RANGE APPROXIMATION METHODS OF POLYNOMIALS AND RATIONAL FUNCTIONS [J].
ALANDER, J .
COMPUTERS & GRAPHICS, 1985, 9 (04) :365-372
[3]  
Alefeld G., 1983, INTRO INTERVAL COMPU
[4]   MAINTAINING KNOWLEDGE ABOUT TEMPORAL INTERVALS [J].
ALLEN, JF .
COMMUNICATIONS OF THE ACM, 1983, 26 (11) :832-843
[5]  
[Anonymous], 14TH P ACM S PRINC P
[6]  
ASAITHAMBI NS, 1982, COMPUTING, V28, P225, DOI 10.1007/BF02241750
[7]  
Clearly J. G., 1987, Future Computing Systems, V2, P125
[8]   CONSTRAINT PROPAGATION WITH INTERVAL LABELS [J].
DAVIS, E .
ARTIFICIAL INTELLIGENCE, 1987, 32 (03) :281-331
[9]   TREE CLUSTERING FOR CONSTRAINT NETWORKS [J].
DECHTER, R ;
PEARL, J .
ARTIFICIAL INTELLIGENCE, 1989, 38 (03) :353-366
[10]  
DECHTER R., 1988, ARTIF INTELL, V34, P1, DOI DOI 10.1016/0004-3702(87)90002-6