Lecture13 - Notes on Complexity Theory Last updated October 2011 Lecture 13 Jonathan Katz 1 Randomized Time Complexity 1.1 How Large is BPP We know

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Notes on Complexity Theory Last updated: October, 2011 Lecture 13 Jonathan Katz 1 Randomized Time Complexity 1.1 How Large is BPP ? We know that P ⊆ ZPP = RP ∩ co RP ⊆ RP ∪ co RP ⊆ BPP ⊆ PSPACE . We currently do not have a very good unconditional bound on the power of BPP — in particular, it could be that BPP = NEXP . Perhaps surprisingly, especially in light of the many randomized algorithms known, the current conjecture is that BPP is not more powerful than P . We will return to this point later in the semester when we talk about derandomization. What (unconditional) upper bounds can we place on BPP ? Interestingly, we know that it is not more powerful than polynomial-size circuits; actually, the following theorem is also a good illustration of the power of non-uniformity. Theorem 1 BPP ⊂ P / poly . Proof Let L ∈ BPP . Using amplification, we know that there exists a polynomial-time Turing machine M such that Pr[ M ( x ) 6 = χ L ( x )] < 2-| x | 2 . Say M uses (at most) p ( | x | ) random coins for some polynomial p . (Note that p is upper-bounded by the running time of M .) An equivalent way of stating this is that for each n , and each x ∈ { , 1 } n , the set of “bad” coins for x (i.e., coins for which M ( x ) outputs the wrong answer) has size at most 2 p ( n ) · 2- n 2 . Taking the union of these “bad” sets over all x ∈ { , 1 } n , we find that the total number of random coins which are “bad” for some x is at most 2 p ( n ) · 2- n < 2 p ( n ) . In particular, there exists at least one set of random coins r * n ∈ { , 1 } p ( n ) that is “good” for every x ∈ { , 1 } n (in fact, there are many such random coins). If we let the sequence of “advice strings” be exactly { r * n } (using the alternate definition of P / poly ), we obtain the result of the theorem. We can also place BPP in the polynomial hierarchy: Theorem 2 BPP ⊆ Σ 2 ∩ Π 2 . Proof We show that BPP ⊆ Σ 2 ; since BPP is closed under complement, this proves the theorem. We begin by proving some probabilistic lemmas. Say S ⊆ { , 1 } m is large if | S | ≥ (1- 1 m )2 m , and is small if | S | < 2 m m . For a string z ∈ { , 1 } m define S ⊕ z def = { s ⊕ z | s ∈ S } ....
View Full Document

This note was uploaded on 01/13/2012 for the course CMSC 652 taught by Professor Staff during the Fall '08 term at Maryland.

Page1 / 4

Lecture13 - Notes on Complexity Theory Last updated October 2011 Lecture 13 Jonathan Katz 1 Randomized Time Complexity 1.1 How Large is BPP We know

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online