Abstract:The first paper explored two strength indicators (extreme point ratio (EPR) and centroid distance (CD)), both of which predict that subtours of the spanning tree in hypergraph polytope STHGP having large cardinality are significantly stronger than the corresponding subtours of small cardinality. In this second paper, we exploit this previously unknown property algorithmically within GeoSteiner, presenting strong computational evidence that the EPR and CD indicators are highly predictive of actual computational strength, at least for subtour inequalities of STHGP. The GeoSteiner package for geometric Steiner trees in the plane uses STHGP to solve the Full Steiner Tree (FST) concatenation problem. The separation algorithms in previous versions of GeoSteiner find violated subtours of only relatively small cardinality, and we examine the underlying algorithmic causes. We present improvements in GeoSteiner specifically designed to strengthen violated subtours by augmentation (instead of reduction), yielding violated subtours of large cardinality. Across all distance metrics and instance classes studied, the computational results are remarkable -- culminating with an optimal solution of a 1,000,000 terminal random Euclidean instance. The conclusion is that the EPR and CD strength indicators presented in the first paper have strong predictive power regarding actual computational strength (at least regarding STHGP subtours). The ability to accurately measure the strength of inequalities has numerous applications of great importance, both in theory and practice.

Automated Discovering And Proving For Geometric Inequalities

Automated Discovering and Proving for Geometric Inequalities.

A Complete Algorithm for Automated Discovering of a Class of Inequality-Type Theorems

The Iyengar Inequality, Revisited

Supporting Proving and Discovering Geometric Inequalities in GeoGebra by using Tarski

A program to create new geometry proof problems

Automated Generation of Geometry Proof Problems Based on Point Geometry Identity

Proving Computational Geometry Algorithms in TLA+2

Program Verification by Using DISCOVERER

A Complete Algorithm for Counting Real Solutions of Polynomial Systems of Equations and Inequalities

Proving Olympiad Algebraic Inequalities without Human Demonstrations

Automated Theorem Proving Practice with Null Geometric Algebra

Geometric Computing for Isogeometric Analysis

Proving Information Inequalities by Gaussian Elimination

Automated generation of geometric theorems from images of diagrams

Constraint problem solving based on geometric abstraction

Proving Information Inequalities and Identities with Symbolic Computation

Approximate Solutions for Abstract Inequality Systems

Quantitative Indicators for Strength of Inequalities with Respect to a Polyhedron, Part II: Applications and Computational Evidence

Equality Detection for Linear Arithmetic Constraints

A Matroid-Based Automatic Prover and Coq Proof Generator for Projective Incidence Geometry