New results in the theory of normal numbers

Richard Crandall

A real number is called $b$-normal iff its base-$b$ digits are "truly random" in a certain technical sense. Until now, every explicit normal number has been of artificial construction; e.g., created by brute-force concatenation of digits. But the situation has changed; now there are normal numbers you can ``point to" in an algebraic sense. Whereas suspected normality of the celebrated constant $\log 2 = \sum_{n \in Z^{+}} 1/(n 2^n)$ remains unresolved to this day, the lecturer---and colleague D.~Bailey---discovered a restricted sum ($n$ running over an explicit subset of $Z^{+}$) that is provably $2$-normal. This work is interdisciplinary, touching upon random generators, number theory, chaotic maps, exponential sums, and discrepancy theory.