Answers are grounded in this paper and cite their sources.
We study the harmonic Jensen functional of harmonically convex functions. We give lower and upper bounds obtained from the support inequality and an identity for the difference between the functional and its lower bound. This identity characterizes the equality cases. In particular, equality holds for monotone functions. Its integral form gives a refinement of the harmonic Hermite–Hadamard inequality. Further, we obtain sharp two-sided bounds under the assumption that tf(t)−m2t2 is convex for a real parameter m. For m<0, this assumption covers some functions which are not harmonically convex. For twice differentiable harmonically convex functions, the upper bound improves the bound of Dragomir. The results are applied to Csiszár f-divergences, where the harmonic mean of the ratios is equal to (1+χ2(p,q))−1. The bounds are illustrated for the Kullback–Leibler, squared Hellinger and Jeffreys divergences and compared with the Lah–Ribarič bound. An application to differential entropy is also given.
No takes yet. Share an insight, caveat, or question.
Barić et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: