As another illustration of the Hopcroft-like DFA minimization algorithm, here we apply it to this DFA:
δ | Symbol | Transition Graph | |||
---|---|---|---|---|---|
Initial? | Final? | State | a | b | |
✓ | q0 | q2 | q4 | ||
q1 | q3 | q2 | |||
q2 | q1 | q4 | |||
✓ | q3 | q1 | q1 | ||
q4 | q5 | q0 | |||
q5 | q1 | q0 |
I0 has two equivalence classes, one containing the non-final states and the other containing the final states. For the given DFA, the latter contains only state q3.
[0] [3] ------------------- ----- 0 1 2 4 5 3 a [0] [3] [0] [0] [0] - b [0] [0] [0] [0] [0] - |
As a result of Round 1, we find that state q1 is 1-distinguishable from the other non-final states, all of which are related in I1.
[0] [1] [3] ---------------- ----- ----- 0 2 4 5 1 3 a [0] [1] [0] [1] - - b [0] [0] [0] [0] - - |
Round 2 determines that states q0 and q4 are 2-indistinguishable from each other, as are q2 and q5, but q0 and q2 are not.
[0] [2] [1] [3] ------- ------- ----- ----- 0 4 2 5 1 3 a [2] [2] [1] [1] - - b [0] [0] [0] [0] - - |
Round 3 determines that there is no need for a Round 4, as, in each equivalence class, all states have the same signature.
The conclusion is that, to minimize the given DFA, we need to merge states q0 and q4 into a single state, and similarly for the pair of states q2 and q5.
Original DFA | Minimized DFA |
---|---|