Exploring properties and inequalities for geometrically arithmetically-Cr-convex functions with Cr-order relative entropy

Asfand Fahad,Yuanheng Wang,Zammad Ali,Riaz Hussain,Shigeru Furuichi
DOI: https://doi.org/10.1016/j.ins.2024.120219
IF: 8.1
2024-02-02
Information Sciences
Abstract:We introduce a new class of interval-valued functions which we call Geometrically Arithmetically Cr-convex functions (GA-Cr-convex functions) and investigate its properties. We explore necessary and sufficient conditions for an interval-valued function to be a GA-Cr-convex function through two distinct approaches. Firstly, we propose these conditions connecting GA-Cr-convex functions with two real-valued GA-convex functions, and secondly, we examine them with respect to a Cr-convex function. Furthermore, we employ our findings to establish generalizations of several renowned inequalities including Jensen's inequality, Jensen-Mercer inequality, and Hermite-Hadamard inequality for GA-Cr-convex functions. The derived inequalities are consistent with the corresponding inequalities for real-valued functions. Consequently, we re-capture several significant results from recent literature. We also define a strong version of the newly introduced class, investigate its properties and prove well-known inequalities for the strong version as well. Lastly, we demonstrate applications of the Cr-order inequalities in information science by defining Cr-relative entropy via log-sum inequality for interval-valued functions. For Cr-order relative entropy, we prove non–negativity, joint convexity and monotonicity to generalize these properties, previously known for Tsallis relative entropy. Consequently, we obtain a generalization of data (information) processing inequality to complete the impact of our work in three directions including generalization, consistency with literature, and applications in information theory.
computer science, information systems
What problem does this paper attempt to address?