A convolution type inequality for fuzzy integrals (Q2467404): Difference between revisions

Given a nonnegative Lebesgue measurable function \(f:[0,1]\to [0,\infty)\), the Sugeno (or fuzzy) integral of \(f\) is defined by \[ -\mkern-18mu\int_0^1f(t)dt :=\sup_{\alpha\geq0}\min\big(\alpha,\mu\{x\in[0,1]:f(x)\geq\alpha\}\big), \] where \(\mu\) stands for the Lebesgue measure. The main results of the paper state that, for \(s\geq2\), Bushell--Okrasinski type inequalities \[ s -\mkern-18mu\int_0^1(1-t)^{s-1}g(t)^sdt\geq \left(-\mkern-18mu\int_0^1 g(t)dt\right)^s \] and \[ s -\mkern-18mu\int_0^1t^{s-1}g(t)^sdt\geq \left(-\mkern-18mu\int_0^1 g(t)dt\right)^s \] hold if \(g:[0,1]\to[0,\infty)\) is a continuous strictly decreasing or increasing function, respectively.