中山大学学报(自然科学版)
中山大學學報(自然科學版)
중산대학학보(자연과학판)
ACTA SCIENTIARUM NATURALIUM UNIVERSITATIS SUNYATSENI
2013年
6期
38-43
,共6页
刘越畅%林晓骏%汤庸
劉越暢%林曉駿%湯庸
류월창%림효준%탕용
人工智能%时态推理%析取时态问题%约束可满足问题%弱演化析取时态网络
人工智能%時態推理%析取時態問題%約束可滿足問題%弱縯化析取時態網絡
인공지능%시태추리%석취시태문제%약속가만족문제%약연화석취시태망락
artificial intelligence%temporal reasoning%disjunctive temporal problem (DTP)%constraint satisfaction problem (CSP)%weakly evolutional disjunctive temporal network (WEDTN)
时态推理是人工智能领域的重要研究方向。析取时态问题(disjunctive temporal problem,DTP)是一类表达力极强的定量时态模型,被广泛应用于智能规划和调度领域的时态建模,近年来吸引了研究人员的关注。提出DTP弱蕴含性和弱演化析取时态网络(weakly evolutional disjunctive temporal network,WEDTN)的定义,基于WEDTN设计和实现了可视化DTP求解器TRSE。系统演示发现基于WEDTN的可视化比常见的基于搜索树的可视化更能有利于人们直观理解DTP的求解算法过程、挖掘和发现问题的结构特征对求解过程的影响。
時態推理是人工智能領域的重要研究方嚮。析取時態問題(disjunctive temporal problem,DTP)是一類錶達力極彊的定量時態模型,被廣汎應用于智能規劃和調度領域的時態建模,近年來吸引瞭研究人員的關註。提齣DTP弱蘊含性和弱縯化析取時態網絡(weakly evolutional disjunctive temporal network,WEDTN)的定義,基于WEDTN設計和實現瞭可視化DTP求解器TRSE。繫統縯示髮現基于WEDTN的可視化比常見的基于搜索樹的可視化更能有利于人們直觀理解DTP的求解算法過程、挖掘和髮現問題的結構特徵對求解過程的影響。
시태추리시인공지능영역적중요연구방향。석취시태문제(disjunctive temporal problem,DTP)시일류표체력겁강적정량시태모형,피엄범응용우지능규화화조도영역적시태건모,근년래흡인료연구인원적관주。제출DTP약온함성화약연화석취시태망락(weakly evolutional disjunctive temporal network,WEDTN)적정의,기우WEDTN설계화실현료가시화DTP구해기TRSE。계통연시발현기우WEDTN적가시화비상견적기우수색수적가시화경능유리우인문직관리해DTP적구해산법과정、알굴화발현문제적결구특정대구해과정적영향。
Temporal reasoning is an important research direction of Artificial Intelligence (AI).Disjunc-tive temporal problem (DTP)is an expressive quantitative temporal model which is widely used in auto-mated planning and scheduling and is therefore being well studied.The definition of weak implication and weakly evolutional disjunctive temporal network (WEDTN)is proposed.Based on the WEDTN structure a visual DTP solver TRSE is designed.The demonstration reveals that the visualization based on WEDTN better facilitates the understanding and discovery of DTP solving algorithm and the impact of structural in-formation on that than usual search tree based visualization.