i'm thinking about problems in P complexity, and i have a question:
If a problem "A" is a problem in P, his complementary have to be a problem in P? Why?
Because every algorithm to decide A
can be trivially turned into one deciding not(A)
simply by negating the answer. That doesn't change its complexity.
As you can see from this argument, that's not a special property of the class P. The same argument applies to all deterministic complexity classes, hence they are all closed under complement: DTIME(f(n)) = co-DTIME(f(n))
for all f
.