Im having a bit of difficulty with the following question:
Given a DFA A = (E = {a,b,c} , Q ,q0 ,F , l) (where l is the transition function), build a new DFA B such that L(B) = L(A) - {a}.
Now I understand that Eb = Ea, but how can I define B transition function, or accepting states, without knowing the accepting states of A?
Thank you.
First, let us construct a DFA called C which accepts w. This DFA has |w| + 2 states: one initial state, one dead state, one accepting state.
Second, let us construct a DFA called D which accepts everything except w. Simply change all non-accepting states to accepting, and vice versa.
Third, let us construct B using the Cartesian Product Machine construction on input DFAs A and D with intersection as the operator. This DFA will hav |Q| x (|w| + 2) states in total.
The language of B is everything accepted by A and D simultaneously. D accepts anything that isn't w; so, B accepts anything in L(A) that isn't w, as required.
EDIT: Some more detail about what B ends up looking like.
Let the states of A be QA and the states of D be QD. Let the accepting states of A be FA and the accepting states of D be FD. Let dA be the transition function for A and dD be the transition function for D. From our construction above, we have:
As for what D looks like: