Two Undecidable Decision Problems on an Ordered Pair of Non-Negative Integers

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

For \(n \in {\mathbb{N}}\), let \(E_{n} = {\{{{1 = x_{k}},{{{x_{i} + x_{j}} = x_{k}},{{x_{i} \cdot x_{j}} = x_{k}}}}:{{i,j,k} \in {\{ 0,\ldots,n\}}}\}}\). For \(n \in {\mathbb{N}}\), \(f{(n)}\) denotes the smallest \(b \in {\mathbb{N}}\) such that if a system of equations \(\mathcal{S} \subseteq E_{n}\) has a solution in \({\mathbb{N}}^{n + 1}\), then \(\mathcal{S}\) has a solution in \({\{ 0,\ldots,b\}}^{n + 1}\). The author proved earlier that the function \(f:{{\mathbb{N}}\rightarrow{\mathbb{N}}}\) is computable in the limit and eventually dominates every computable function \(g:{{\mathbb{N}}\rightarrow{\mathbb{N}}}\). We present a short program in MuPAD which for \(n \in {\mathbb{N}}\) prints the sequence \({\{{f_{i}{(n)}}\}}_{i = 0}^{\infty}\) of non-negative integers converging to \(f{(n)}\). Since \(f\) is not computable, no algorithm takes as input non-negative integers \(n\) and \(m\) and decides whether or not \({\forall{(x_{0},\ldots,x_{n})}} \in\)\({{\mathbb{N}}^{n + 1}{\exists{(y_{0},\ldots,y_{n})}}} \in {{{\{ 0,\ldots,m\}}^{n + 1}{({{\forall k} \in {{\{ 0,\ldots,n\}}{({1 = x_{k}\Rightarrow 1 = y_{k}})}}})}}\ \land}\)\({(\forall i,j,k \in {\{ 0,\ldots,n\}}{(x_{i} + x_{j} = x_{k}\Rightarrow y_{i} + y_{j} = y_{k})})}\ \land\)  \({(\forall i,j,k \in {\{ 0,\ldots,n\}}{(x_{i} \cdot x_{j} = x_{k}\Rightarrow}} {y_{i} \cdot y_{j} = y_{k})})\). Similarly, no algorithm takes as input non-negative integers \(n\) and \(m\) and decides whether or not \({\forall{(x_{0},\ldots,x_{n})}} \in {{\mathbb{N}}^{n + 1}{\exists{(y_{0},\ldots,y_{n})}}} \in \\ {{{\{ 0,\ldots,m\}}^{n + 1}{({{{\forall j},k} \in {{\{ 0,\ldots,n\}}{({{x_{j} + 1} = x_{k}\Rightarrow{y_{j} + 1} = y_{k}})}}})}} \land \\ {({{{\forall i},j,k} \in {{\{ 0,\ldots,n\}}{({{x_{i} \cdot x_{j}} = x_{k}\Rightarrow{y_{i} \cdot y_{j}} = y_{k}})}}}).}}\) For \(n \in {\mathbb{N}}\), \(\beta{(n)}\) denotes the smallest \(b \in {\mathbb{N}}\) such that if a system of equations \(\mathcal{S} \subseteq E_{n}\) has a unique solution in \({\mathbb{N}}^{n + 1}\), then this solution belongs to \({\{ 0,\ldots,b\}}^{n + 1}\). The author proved earlier that the function \(\beta:{{\mathbb{N}}\rightarrow{\mathbb{N}}}\) is computable in the limit and eventually dominates every function \(\delta:{{\mathbb{N}}\rightarrow{\mathbb{N}}}\) with a single-fold Diophantine representation. The computability of \(\beta\) is unknown. We present a short program in MuPAD which for \(n \in {\mathbb{N}}\) prints the sequence \({\{{\beta_{i}{(n)}}\}}_{i = 0}^{\infty}\) of non-negative integers converging to \(\beta{(n)}\).

Article activity feed