A characterization of generalized monotone normed cones (Q2644336): Difference between revisions

Let \(C\) be a cone, i.e. a triple \((C,+,\cdot)\) such that \((C, +)\) is a monoid and \(\cdot: \mathbb{R}^+\times C\to\mathbb{R}^+\) satisfies the axioms of a linear space when restricted to nonnegative scalars. A function \(p: C\to\mathbb{R}^+\) that is homogeneous for nonnegative scalars is called a quasi-norm on \(C\) if: (i) \(p(x+ y)\leq p(x)+ p(y)\) for all \(x,y\in C\); (ii) \(\max\{p(x), p(y)\}\neq 0\) for each pair \(x,y\in C\) such that \(x\neq 0\) and \(x+ y= 0\); (iii) \(p(x)\neq 0\) if \(x\neq 0\) and there is no \(y\in C\) such that \(x+ y= 0\). It is called monotone, if additionally \(p(x)\leq p(x+ y)\) for each pair \(x,y\in C\). Every quasi-norm \(p\) on \(C\) induces an extended quasi-pseudo-metric \(d_p\) on \(C\) which is fined by \(d_p(x, y)= \inf\{p(a)\mid y= x+ a\}\) if \(y\in x+ C\) and \(d_p(x, y)= +\infty\) if \(y\not\in x+ C\). The quasi-pseudo-metric \(d_p\) in turn induces a topology \(\tau_p\) on \(C\). Motivated by the observation that some topological structures that are used for the complexity analysis of programs and algorithms are induced from quasi-norms on cones which are monotone the authors characterize those quasi-normed cones \((C, p)\) for which there exists a monotone quasi-norm \(q\) on \(C\) such that \(\tau_p= \tau_q\). Additionally, they define and study the dual cone of a monotone quasi-normed cone and the monotone quotient of a general cone.