scientific article; zbMATH DE number 18644

From MaRDI portal
Publication:3976049





[For the entire collection see Zbl 0731.00026.] The paper presents some nice ideas of how to parallelize known algorithms to solve the word problem for some important classes of groups. To be more precise, the parallel complexity of some group constructions --- such as free and direct products, special \(HNN\)-extensions and wreath products --- is studied. So the question is as follows: Suppose the finitely presented group \(G\) is built from \(H\) and \(K\) by one of the above mentioned constructions and suppose the parallel complexities of \(H\) and \(K\) are known. What can be said about the parallel complexity of \(G\)? Here is one typical result: Let \(G\) be the wreath product of \(H\) and \(K\). Then the word problem of \(G\) is \(NC^ 1\)-equivalent to the maximal complexity of \(H\) and \(K\). Here \(NC^ 1\) is the set of functions \(f: \{0,1\}^*\to\{0,1\}^*\) computable by a uniform Boolean circuit family \(<\alpha_ n>_{n\in N}\) in which \(\alpha_ n\) has size polynomial in \(n\) and depth \(O(\hbox{log} n)\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3976049)