compiler-principle03_2.ppt

算法：NFA M=(S,Σ,δ,S0, F)→DFA M” = (S’,Σ,δ’,I0, F’) ①取I0=ε-closure(S0), 将I0加入到S’中，设I0为未标记 下图中： I0=ε-closure({X})={X,5,1} ②若状态集S’中存在有未标记的状态Ii=[S1,S2,…,Sj]，则标记之，且求It=ε-closure(Move([Ii, a))，令δ’(Ii, a)=[r1,r2,…,rt] =It ，将It加入S’中，设It为未标记 下图中，标记I0： Ia=ε-closure(Move(I0,a))=ε-closure({5,3})={5,3,1}=I1 Ib=ε-closure(Move(I0,b))=ε-closure({5,4})={5,4,1}=I2 例： Y a X a a ε 5 ε b 1 3 4 a b b 2 ε 6 ε b 算法：NFA M=(S,Σ,δ,S0, F)→DFA M” = (S’,Σ,δ’,I0, F’) ③ 重复②，直至S’中没有未标记状态为止。 下图确定化： ε-closure(Move(I1,a))=ε-closure({5,2,3})={5,2,3,1,6,Y}=I3 ε-closure(Move(I1,b))=ε-closure({5,4})={5,4,1}=I2 ε-closure(Move(I2,a))=ε-closure({5,3})={5,3,1}=I1 ε-closure(Move(I2,b))=ε-closure({5,2,4})={5,2,4,1,6,Y}=I4 ε-closure(Move(I3,a))=ε-closure({5,2,3,6})={5,2,3,1,6,Y}=I3 ε-closure(Move(I3,b))=ε-closure({5,4,6})={5,4,6,1,Y}=I5 例： Y a X a a ε 5 ε b 1 3 4 a b b 2 ε 6 ε b ε-closure(Move(I4,a))=ε-closure({5,3,6})={5,3,1,6,Y}=I6 ε-closure(Move(I4,b))=ε-closure({5,2,4,6})={5,2,4,6,1,Y}=I4 ε-closure(Move(I5,a))={5,3,1,6,Y}=I6 ε-closure(Move(I5,b))={5,2,4,6,1,Y}=I4 ε-closure(Move(I6,a))={5,3,1,2,6,Y}=I3 ε-closure(Move(I6,b))={5,4,6,1,Y}=I5 S’={I0, I1, I2, I3, I4, I5, I6} 初态为I0, 终态为含有原终态的状态集。 F’={I∈S’ | I∩F≠?}={I3, I4, I5, I6} 例： Y a X a a ε 5 ε b 1 3 4 a b b 2 ε 6 ε b I0 I4 a I3 a a I6 a a b b I5 b 确定化后的DFA图： I Ia Ib I0 I1 I2 I1 I3 I2 I2 I1 I4 I3 I3 I5 I4 I6 I4 I5 I6 I4 I6 I3 I5 例： Y a X a a ε 5 ε b 1 3 4 a b b 2 ε 6 ε b a b I1 I2 b b b b b a 例：将下面NFA确定化。 10 a 0 a ε 7 ε b 1 3 5 b 6 ε 8 ε 2 4 ε ε b 9 ε ε 取I0=ε-closure({S0})= ε-closure({0})={0,1,2,4,7} 标记I0:ε-closure(Move(T0,a))=ε-closure({3,8})={1,2,3,4,6,7,8}=I1 ε-closure(Move(T0,b))=ε-closure({5})={1,2,4,5,6,7}=I2 标记T1:ε-closure(Move(I1,a))={1,2,3,4,6,7,8}=I1 ε-closure(Move(I1,b))={1,2,4,5,6,7,9}=I3 标记T2:ε-closure(Move(I2,a))={1,2,3,4,6,7,8}=I1 ε-closure(Move(I2,b))={1,2,4,5,6,7}=I2 标记T3:ε-closure(Move(I3,a))={1,2,3,4,6,7,8}=I1 ε-closure(Move(I3,b))={1,2,4,5,6,7,10}=I4 标记T4:ε-closure(Move(I4,a