<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Complexity theory on Thomas van Maaren</title><link>https://tvmaaren.nl/mathematics/complexity/</link><description>Recent content in Complexity theory on Thomas van Maaren</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Mon, 26 Jan 2026 12:57:01 +0100</lastBuildDate><atom:link href="https://tvmaaren.nl/mathematics/complexity/index.xml" rel="self" type="application/rss+xml"/><item><title>Circuits</title><link>https://tvmaaren.nl/mathematics/complexity/circuits/</link><pubDate>Mon, 26 Jan 2026 12:57:01 +0100</pubDate><guid>https://tvmaaren.nl/mathematics/complexity/circuits/</guid><description>&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; An \(n\)-ary boolean function is a function \(f:\{0,1\}^n\to \{0,1\}\).&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Proposition&lt;/strong&gt; For every boolean function \(f\) there is a propositional formula \(\varphi\) such that the function defined by \(\varphi\) is equal to \(f\).&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; The &lt;em&gt;size&lt;/em&gt; of a circuit is defined as the amount of and-, not-, and or-gates.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; Suppose that \(\{C_n\}_{n\in \mathbb{N}}\) is a family of circuits such that for every \(n\), \(C_n\) computes a boolean function of \(n\) inputs. A family computes a language \(L\) if for all \(x\in \{0,1\}^n\) we have&lt;/p&gt;</description></item><item><title>Randomized computation</title><link>https://tvmaaren.nl/mathematics/complexity/randomized/</link><pubDate>Sun, 25 Jan 2026 17:14:16 +0100</pubDate><guid>https://tvmaaren.nl/mathematics/complexity/randomized/</guid><description>&lt;h1 id="pp"&gt;PP&lt;/h1&gt;
&lt;p&gt;We are going to look at a variant of NP where the acceptance criterion is probabilistic over all paths. Given a nondeterministic Turing machine the &lt;em&gt;error probability&lt;/em&gt; is the ratio of the parths giving the wrong answer. These are &lt;em&gt;probabilistic&lt;/em&gt; Turing machines.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; PP is the class of &lt;em&gt;probabilistic polynomial time&lt;/em&gt; sets. Meaning a set \(L\) that is computed in polynomial time by probabilistic machine such that&lt;/p&gt;
\[x\in L \iff \text{the fraction of accepting paths is }&gt;1/2\]&lt;p&gt;Which implies that &lt;/p&gt;</description></item><item><title>Diagonalization</title><link>https://tvmaaren.nl/mathematics/complexity/diagonalization/</link><pubDate>Fri, 23 Jan 2026 16:14:41 +0100</pubDate><guid>https://tvmaaren.nl/mathematics/complexity/diagonalization/</guid><description>&lt;p&gt;Diagonalization is a proof technique first used by Georg Cantor to show that some infinities are greater than others. In proof theory it is used a lot. Most famously for proving the halting problem.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Theorem&lt;/strong&gt;(Halting problem) Not every problem is computable.&lt;/p&gt;
&lt;p&gt;Assume to the contrary that any problem is computable. We consider \(\{(M,x) \mid M(x)\text{ halts }\}\). Then this must also be a computable, hence we can define a machine \(N\) such that \(N(M)\) does not halt if and only if \(M(M)\) halts. Now let us think if \(N(N)\) halts. We see that \(N(N)\) does not halt if and only if \(N(N)\) halts which is a contradiction.&lt;/p&gt;</description></item><item><title>Chapter 4 Relativized computation and the polynomial hierarchy</title><link>https://tvmaaren.nl/mathematics/complexity/relativized/</link><pubDate>Thu, 22 Jan 2026 15:24:10 +0100</pubDate><guid>https://tvmaaren.nl/mathematics/complexity/relativized/</guid><description>&lt;h1 id="relativized-computation"&gt;Relativized computation&lt;/h1&gt;
&lt;p&gt;An oracle Turing machine is a machine which has a lookup table (oracle) which has constant lookup time. This lookup table can be read from and written to.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; Given oracle turing machine \(M\), we let \(M^B\) be the function computed using oracle \(B\).&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Definition&lt;/strong&gt; A set \(A\) &lt;em&gt;Turing reduces&lt;/em&gt; to \(B\) denoted \(A\leq_T^p B\), if \(A=M^B\) for some machine \(M\) working in polynomial time.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Proposition&lt;/strong&gt; \(A\leq_m^p B \implies A\leq_T^p B\)&lt;/p&gt;</description></item><item><title>Chapter 11 Proof Complexity</title><link>https://tvmaaren.nl/mathematics/complexity/proof-complexity/</link><pubDate>Wed, 21 Jan 2026 13:36:43 +0100</pubDate><guid>https://tvmaaren.nl/mathematics/complexity/proof-complexity/</guid><description>&lt;p&gt;We know that any propositional formula has a proof. Now we are interested if every propositional formula has a &amp;ldquo;short&amp;rdquo; proof. Where &amp;ldquo;short&amp;rdquo; means polynomial given the size of the propositional formula.&lt;/p&gt;</description></item></channel></rss>