Lectures on the Thue Principle

Let $\{ {x_i } \}$ be any sequence approximating an algebraic number $\alpha $ of degree r, and let $x_{i + 1} = \varphi (x_i ,x_{i - 1} , \cdots ,x_{i - d + 1} )$, for some rational function $\varphi $ with integral coefficients. Let M denote the number of multiplications or divisions needed to compute $\varphi $ and let $\bar M$ denote the number of multiplications or divisions, except by constants, needed to compute $\varphi $. Define the multiplicative efficiency measure of $\{ {x_i } \}$ as $E = {{(\log _2 p)} / M}$ or as $\bar E = {{(\log _2 p)} / {\bar M}}$, where p is the order of convergence of $\{ {x_i } \}$. Kung [1] showed that $\bar E \leqq 1$ or equivalently, $\bar M \geqq \log _2 p$. In this paper we show that (i) $\bar M \geqq \log _2 [r(\lceil p \rceil - 1) + 1] - 1$; (ii) if $E = 1$ then $\alpha $ is a rational number; (iii) if $\bar E = 1$ then $\alpha $ is a rational or quadratic irrational number. This settles the question of when the multiplicative efficiency E or $\bar E$ achieves its optimal value of unity. Also, as a consequence of result (i), we show that the maximal efficiency $\bar E$ achievable by algebraic numbers of degree r drops at least as $O[(\log r)^{ - 1} ]$, provided that we only consider sequences $\{ {x_i } \}$ of bounded order of convergence.

In this paper, to identify irrational numbers while appropriately emphasizing basic functions. All rational numbers can be expressed algebraically, but not all irrational numbers. Rational, irrational, algebraic numbers and transcendental numbers are mostly concerned with irrational numbers using some elementary functions.

The set of real numbers and the set of complex numbers have the power of continuum. Among these numbers, those which are ``interesting'', which appear ``naturally'', which deserve our attention, form a countable set. Starting from this point of view we are interested in the periods as defined by M.~Kontsevich and D.~Zagier. We give the state of the art on the question of the arithmetic nature of these numbers: to decide whether a period is a rational number, an irrational algebraic number or else a transcendental number is the object of a few theorems and of many conjectures. We also consider the approximation of such numbers by rational or algebraic numbers.

where the Li(x) are homogeneous polynomials of degree 1 with coefficients which are algebraic numbers. One of the most interesting cases arises when the following two conditions are satisfied. (1) The Li(x) never vanish simultaneously at any of V. (2) Under the mapping a of V again becomes a of V. Generally speaking the only points which interest us are those which can be expressed with coordinates which are algebraic numbers. = (to, I, * * * n is such a point, the number field generated by the ratios ti/tj will be the smallest field in which is rational. The degree of this field over the rational numbers, we shall call the degree of rationality of P. Suppose now that conditions (1) and (2) are satisfied, then by iterating the mapping, a given of V will generate an infinite sequence = Po, P1, P2, *** of points on the variety. This sequence will consist either of distinct points or else there will be repetitions. In the latter case the sequence ultimately becomes periodic, and we shall describe this situation by saying P is an exceptional point. One of the results established can now be stated, namely, If we exclude the case in which the mapping is linear, then there will be at most a finite number of exceptional points with a given degree of rationality. We can illustrate this result by means of the well known Weierstrass function P(Z1 g2, g3), where g2 and g3 are algebraic numbers. A complex number 'a' is called a point if for some integer n, which is not zero, na is a period. The values of p(z) at the division points are called values, and it is easy to see that these are algebraic numbers. By comparing the elliptic function with the exponential function, we see that these division values are in some respects analogous to roots of unity, and it is therefore natural to inquire how they are distributed among the number fields. A special case of the general theorem stated above, shows that if we limit ourselves to number fields whose absolute degrees are bounded by some given integer, then amongst all these fields there will be only a finite number of division values. These facts are consequences of Theorem 2 which is an inequality proved under very general conditions. This inequality has another application (not discussed here), namely 167

In a recent paper [1] methods were introduced for investigating the accuracy with which certain algebraic numbers may be approximated by rational numbers. It is the main purpose of the present paper to deduce, using similar techniques, results concerning the accuracy with which the natural logarithms of certain rational numbers may be approximated by rational numbers, or, more generally, by algebraic numbers of bounded degree.

Contrary to popular misconception, the question in the title is far from simple. It involves sets of numbers on the first level, sets of sets of numbers on the second level, and so on, endlessly. The infinite hierarchy of the levels involved distinguishes the concept of "definable number" from such notions as "natural number", "rational number", "algebraic number", "computable number" etc.

We establish various new results on a problem proposed by Mahler [Some suggestions for further research. Bull. Aust. Math. Soc.29 (1984), 101–108] concerning rational approximation to fractal sets by rational numbers inside and outside the set in question. Some of them provide a natural continuation and improvement of recent results of Broderick, Fishman and Reich, and Fishman and Simmons. A key feature is that many of our new results apply to more general, multi-dimensional fractal sets and require only mild assumptions on the iterated function system. Moreover, we provide a non-trivial lower bound for the distance of a rational number $p/q$ outside the Cantor middle-third set $C$ to the set $C$, in terms of the denominator $q$. We further discuss patterns of rational numbers in fractal sets. We highlight two of them: firstly, an upper bound for the number of rational (algebraic) numbers in a fractal set up to a given height (and degree) for a wide class of fractal sets; and secondly, we find properties of the denominator structure of rational points in ‘missing-digit’ Cantor sets, generalizing claims of Nagy and Bloshchitsyn.

Let $r, \,m$ be positive integers. Let $x$ be a rational number with $0 \le x <1$. Consider $\Phi_s(x,z) =\displaystyle\sum_{k=0}^{\infty}\frac{z^{k+1}}{{(k+x+1)}^s}$ the $s$-th Lerch function with $s=1, 2, \cdots, r$. When $x=0$, this is a polylogarithmic function. Let $\alpha_1, \cdots, \alpha_m$ be pairwise distinct algebraic numbers of arbitrary degree over the rational number field, with $0<|\alpha_j|<1 \,\,\,(1\leq j \leq m)$. In this article, we show a criterion for the linear independence, over an algebraic number field containing $\mathbb{Q}(\alpha_1, \cdots, \alpha_m)$, of all the $rm+1$ numbers : $\Phi_1(x,\alpha_1)$, $\Phi_2(x,\alpha_1), $ $\cdots , \Phi_r(x,\alpha_1)$, $\Phi_1(x,\alpha_2)$, $\Phi_2(x,\alpha_2), $ $\cdots , \Phi_r(x,\alpha_2), \cdots, \cdots, \Phi_1(x,\alpha_m)$, $\Phi_2(x,\alpha_m)$, $\cdots , \Phi_r(x,\alpha_m)$ and $1$. This is the first result that gives a sufficient condition for the linear independence of values of the Lerch functions at several distinct algebraic points, not necessarily lying in the rational number field nor in quadratic imaginary fields. We give a complete proof with refinements and quantitative statements of the main theorem announced in [10], together with a proof in detail on the non-vanishing Wronskian of Hermite type.

Around the late 1970s, Rohrlich made a conjecture about multiplicative algebraic relations among the special values of the Γ- function. Later, Lang generalized the Rohrlich conjecture to polynomial algebraic relations among special values of the gamma function. In 2009, Gun et al. (J. Number Theory 129 (2009), no. 8, 1858–1873) formulated a variant of this conjecture of Rohrlich and a variant of the conjecture of Lang that deals with the linear independence of the values at non-integeral rational numbers of the logarithm of the gamma function over the field of rationals and algebraic numbers, respectively. In this direction, they proved a set of interesting results for the case of primes and their powers over the field of rationals. Further for the case of prime powers, they have extended their results assuming the Schanuel's conjecture. In this article, we improve their results without assuming Schanuel's conjecture. Further we provide counter examples to these variants of conjectures of Rohrlich and Lang for an infinite class of integers having at least two prime factors satisfying certain conditions.

We consider the sequences of fractional parts {ξαn}, n = 1, 2, 3,…, and of integer parts [ξαn], n = 1, 2, 3,…, where ξ is an arbitrary positive number and α > 1 is an algebraic number. We obtain an inequality for the difference between the largest and the smallest limit points of the first sequence. Such an inequality was earlier known for rational α only. It is also shown that for roots of some irreducible trinomials the sequence of integer parts contains infinitely many numbers divisible by either 2 or 3. This is proved, for instance, for [ ξ ( ( 13 − 1 ) / 2 ) n ] , n = 1, 2, 3,…. The fact that there are infinitely many composite numbers in the sequence of integer parts of powers was proved earlier for Pisot numbers, Salem numbers and the three rational numbers 3/2, 4/3, 5/4, but no such algebraic number having several conjugates outside the unit circle was known. 2000 Mathematics Subject Classification 11J71, 11R04, 11R06, 11A41.

We formalize algebraic numbers in Isabelle/HOL, based on existing libraries for matrices and Sturm’s theorem. Our development serves as a verified implementation for real and complex numbers, and it admits to compute roots and completely factor real and complex polynomials, provided that all coefficients are rational numbers. Moreover, we provide two implementations to display algebraic numbers, an injective and expensive one, and a faster but approximative version.

We derive an algorithm based on the ellipsoid method that solves linear programs whose coefficients are real algebraic numbers. By defining the encoding size of an algebraic number to be the bit size of the coefficients of its minimal polynomial, we prove the algorithm runs in time polynomial in the dimension of the problem, the encoding size of the input coefficients, and the degree of any algebraic extension which contains the input coefficients. This bound holds even if all input and arithmetic is performed symbolically, using rational numbers only.

We describe three ways to generalize Lenstra's algebraic integer recovery method. One direction adapts the algorithm so that rational numbers are automatically produced given only upper bounds on the sizes of the numerators and denominators. Another direction produces a variant which recovers algebraic numbers as elements of multiple generator algebraic number fields. The third direction explains how the method can work if a reducible minimal polynomial had been given for an algebraic generator. Any two or all three of the generalizations may be employed simultaneously.

We derive a bound on the computational complexity of linear programs whose coefficients are real algebraic numbers. Key to this result is a notion of problem size that is analogous in function to the binary size of a rational-number problem. We also view the coefficients of a linear program as members of a finite algebraic extension of the rational numbers. The degree of this extension is an upper bound on the degree of any algebraic number that can occur during the course of the algorithm, and in this sense can be viewed as a supplementary measure of problem dimension. Working under an arithmetic model of computation, and making use of a tool for obtaining upper and lower bounds on polynomial functions of algebraic numbers, we derive an algorithm based on the ellipsoid method that runs in time bounded by a polynomial in the dimension, degree, and size of the linear program. Similar results hold under a rational number model of computation, given a suitable binary encoding of the problem input.

Computational geometry algorithms branch on the signs of predicates. Prior predicate evaluation techniques are slow on degenerate (zero sign) predicates, especially on predicates on algebraic numbers. Degeneracy is common for predicates whose arguments have common antecedents. We present three randomized algorithms for degeneracy detection. The first algorithm uses modular arithmetic to detect degenerate predicates on rational numbers. The second algorithm uses quotient rings to reduce detecting degenerate predicates on algebraic numbers to multiple rational predicates, which can be evaluated deterministically or using the first algorithm. This algorithm is impractical because it is exponential in the number of algebraic numbers, yet it is still much faster than prior work that uses root separation bounds. The third algorithm uses a perturbation to eliminate degenerate algebraic predicates that are not identical to zero and a second perturbation to detect those that are. The first and third algorithms are incorporated into an exact geometric computation library. By sampling values generated by the algorithms, the library estimates the degeneracy detection failure probability over the lifetime of every calling program. We call this approach statistical degeneracy detection (SDD). Extensive testing shows that predicate evaluation is reliable and fast.