فهرست نشریات علمی دسترسی آزاد ایران

‎We refine the classical Heinz and P\'olya inequalities by establishing sharper bounds and developing a unified interpolation framework that applies to real numbers, matrices, and strictly positive elements of both unital and non-unital C*-algebras. Motivated by numerical inaccuracies in earlier quadrature approximations, we introduce several new quadrature schemes—adaptive, higher-order, and hybrid Gaussian–adaptive—and demonstrate their effectiveness in estimating the integral representation of the logarithmic mean. These refinements yield improved inequalities lying strictly between the logarithmic mean and the Heinz–Heron interpolating mean $F_{1/3}$​. We further generalize these results to operator settings, including non-commutative cases, spectral perturbations, and multi-operator frameworks, and provide examples illustrating the sharpness of the new bounds. The methods developed here enhance both the theoretical understanding and numerical precision of inequalities central to operator theory and matrix analysis.