Part V
Important cases of polyhedra of generalized finite semimetrics
Chapter 12
Cones of partial semimetrics and weightable quasi-semimetrics
12.1Preliminaries
For given two partial semimetrics p1 and p2 on a set X their non-negative linear combination d = αp1 + βp2, α, β ≥ 0, is a partial semimetric on X. Here, as usual, for all x, y ∈ X it holds
Similarly, for given two weightable quasi-semimetrics q1 and q2 on a set X their non-negative linear combination q = αq1 + βq2, α, β ≥ 0, is a weightable quasi-semimetric on X.
Then we can speak about the cones of all partial semimetrics and all weightable quasi-semimetrics on n points, in fact, on the set Vn = {1, 2, ..., n}. We can consider already the similar cones, related to cuts semimetrics, and some corresponding polytopes.
In this chapter we consider, for small values of n, the cone of all partial semimetrics on Vn (as well as the cones of weak and strong partial semimetrics on Vn), the cone of all weightable quasi-semimetrics on Vn (including weak and strong weightable quasisemimetrics on Vn), the cone of all weighted semimetrics on Vn (together with down-weighted and strong-weighted semimetrics on Vn). For any cone C under consideration we construct its {0, 1}-C cone, generated by all extreme rays of C, containing a non-zero {0, 1}-valued point. In some cases we try to construct similar polytopes.
Partial semimetrics are generalization of semimetrics, having important applications in Computer Science (Domain Theory, Analysis of Data Flow Deadlock, Complexity Analysis of Programs, etc.). They are used for treatment of partially defined/computed objects in Semantics of Computation.
Partial semimetrics were introduced by Matthews in [Matt92] for treatment of partially defined objects in Computer Science. Weak partial semimetrics were introduced in [Heck99] as a generalization of partial semimetrics, introduced in [Matt92]. ...