Carleman's condition

In mathematics, particularly, in analysis, Carleman's condition gives a sufficient condition for the determinacy of the moment problem. That is, if a measure ${\displaystyle \mu }$ satisfies Carleman's condition, there is no other measure ${\displaystyle \nu }$ having the same moments as ${\displaystyle \mu .}$ The condition was discovered by Torsten Carleman in 1922.^[1]

For the Hamburger moment problem (the moment problem on the whole real line), the theorem states the following:

Let ${\displaystyle \mu }$ be a measure on ${\displaystyle \mathbb {R} }$ such that all the moments $${\displaystyle m_{n}=\int _{-\infty }^{+\infty }x^{n}\,d\mu (x)~,\quad n=0,1,2,\cdots }$$ are finite. If $${\displaystyle \sum _{n=1}^{\infty }m_{2n}^{-{\frac {1}{2n}}}=+\infty ,}$$ then the moment problem for ${\displaystyle (m_{n})}$ is determinate; that is, ${\displaystyle \mu }$ is the only measure on ${\displaystyle \mathbb {R} }$ with ${\displaystyle (m_{n})}$ as its sequence of moments.

For the Stieltjes moment problem, the sufficient condition for determinacy is $${\displaystyle \sum _{n=1}^{\infty }m_{n}^{-{\frac {1}{2n}}}=+\infty .}$$

In ^[2], Nasiraee et al. showed that, despite previous assumptions ^[3], when the integrand is an arbitrary function, Carleman's condition is not sufficient, as demonstrated by a counter-example. In fact, the example violates the bijection, i.e. determinacy, property in the probability sum theorem. When the integrand is an arbitrary function, they further establish a sufficient condition for the determinacy of the moment problem, referred to as the generalized Carleman's condition.

• Akhiezer, N. I. (1965). The Classical Moment Problem and Some Related Questions in Analysis. Oliver & Boyd.
• Chapter 3.3, Durrett, Richard. Probability: Theory and Examples. 5th ed. Cambridge Series in Statistical and Probabilistic Mathematics 49. Cambridge ; New York, NY: Cambridge University Press, 2019.