000			04234nas|a2200409 i 4500
005			20260818214644.0
008			020801s2000 th uzm rtt 00| a1eng d
035			_a.b11733718
099		9	_aAIT Diss. no. CS-98-1
100	0		_aDuong Tuan Anh
245	1	0	_aSolving incrementally constraint hierarchies over real intervals
260			_aBangkok : _bAsian Institute of Technology, _c1998
300			_a166 p.
490	1		_aDissertation ; _vno. CS-98-1
500			_aA dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering, School of Advanced Technologies
502			_aThesis (Ph.D.) - Asian Institute of Technology, 1998
520			_aCSP over real intervals (ICSP) is an important class of CSP. Several applications involving continuous domains or imprecise data can be formulated as ICSPs. While the key feature of Finite CSPs is the {uFB01}niteness of the domains, the key idea behind solving ICSPS is the "approximation to finiteness" of infinite domains. Although not all basic results from Finite CSPs are applicable to ICSPs, arc-consistency techniques for Finite CSPs can be adapted to ICSPs by some kind of approximation based on Interval Arithmetic. The present research represents an attempt to develop constraint satisfaction algorithms for ICSPs in dynamic environments. For this purpose, we first devise an incremental constraint deletion technique for constraints over real intervals by which recomputation from scratch can be avoided and the time complexity is the same as that of algorithm for constraint addition. This constraint deletion technique helps to bring out the arc-consistency algorithm for dynamic CSPs over real intervals. Then we employ the resultant arc-consistency algorithm as a flat constraint solver in a hierarchy solver for constraint hierarchies over real intervals. The hierarchy solver makes use of dependency information among constraints to detect the possible causes of inconsistencies in the constraint hierarchy. One of the unique feature of this hierarchy solver is that there is a clear division between the flat solver and the hierarchy solver. Thanks to this feature, the hierarchy solver is general and we can adapt it to different domain/constraint speci{uFB01}c flat solvers to produce quickly different hierarchy solvers. The research is also devoted to the design and implementation of an HCLP system over real intervals with locally-predicate-better (lbp) comparator in order to enhance the expressiveness and usefulness of CLP(Intervals). Our prototype HCLP(RI,1pb) interpreter is somewhat close to a full-implementation approach and employs a simpli{uFB01}ed version of the above mentioned hierarchy solver as the constraint solver. Its computational model which is based on a simple operational semantics brings out to a simple and ef{uFB01}cient implementation. Due to the " glass box " approach adopted in the implementation, the interpreter is also extendible.
650	1	0	_aConstraints (Artificial intelligence)
700	0		_aKanchana Kanchanasut, _eChairperson
700	1		_aHuynh, Ngoc Phien, _eExamination Committee
700	1		_aTabucanon, Mario T., _eExamination Committee
700	1		_aMaher, Michael J., _eExamination Committee
810	2		_aAsian Institute of Technology. _tDissertation ; _vno. CS-98-1
856			_3Full-Text _uhttp://203.159.5.9/ait-thesis/detail.php?q=B00314
907			_a.b11733718 _bmnait _cu
902			_a240407
998			_b0 _c981202 _dm _ea _fu _g0
945			_lmnait
945			_lmnait
945			_lmnarc
942			_c20
942			_c40
909			_aBarcode : 30050120562532 _bCREATED : 1998-02-12 _cRECORD # : i12110784 _dLPATRON : 1014142 _eLCHKIN : 2002-03-03 _f# RENEWALS : 0 _g# OVERDUE : 0 _hIUSE3 : 0 _iTOT CHKOUT : 4 _jTOT RENEW : 0
909			_aBarcode : 30050120562524 _bCREATED : 1998-02-12 _cRECORD # : i12110796 _dLPATRON : 1015151 _eLCHKIN : 2019-02-04 _f# RENEWALS : 0 _g# OVERDUE : 0 _hIUSE3 : 0 _iTOT CHKOUT : 4 _jTOT RENEW : 3
909			_aBarcode : 30050160037494 _bCREATED : 2016-01-25 _cRECORD # : i12873214 _dLPATRON : 0 _eLCHKIN : - _f# RENEWALS : 0 _g# OVERDUE : 0 _hIUSE3 : 0 _iTOT CHKOUT : 0 _jTOT RENEW : 0
999			_c109430 _d109430