Possibilities Determine the Combinatorial Structure of Probability Polytopes

Samson Abramsky,Rui Soares Barbosa,Kohei Kishida,Raymond Lal,Shane Mansfield

DOI: https://doi.org/10.1016/j.jmp.2016.03.006

2016-03-25

Abstract:We study the set of no-signalling empirical models on a measurement scenario, and show that the combinatorial structure of the no-signalling polytope is completely determined by the possibilistic information given by the support of the models. This is a special case of a general result which applies to all polytopes presented in a standard form, given by linear equations together with non-negativity constraints on the variables.

Quantum Physics,Logic in Computer Science,Probability

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to explore how the combinatorial structure of the polytope of probability models (especially the no - signaling polytope) widely studied in quantum information and fundamental research is completely determined by the support sets of these models (i.e., the information about which events may or may not occur). Specifically, the authors studied the set of no - signaling empirical models in a given measurement scenario and proved that the combinatorial structure of the no - signaling polytope is completely determined by the support sets of the models. This result is applicable not only to specific no - signaling polytopes but also to a broader class of polytopes as long as these polytopes can be defined by linear equations and non - negativity constraints.

### Core contributions of the paper

1. **Main theorem**:
- **Theorem 1.1**: For any given measurement scenario, a polytope \( N \) of no - signaling probability models is defined. Let \( F \) be the face lattice of this polytope, and \( S \) be the support sets of the models in \( N \), which are naturally ordered according to the context - inclusion relationship. Then there exists an order - isomorphism \( F\sim S \), that is, the combinatorial type of the polytope \( N \) is completely determined by its support sets.

2. **Corollaries**:
- All models in the relative interior of each face of the polytope \( N \) have the same support set.
- The vertices of \( N \) are precisely those probability models with the smallest support sets. Moreover, for each such smallest support set, there is only one probability model in the polytope \( N \).
- The vertices of \( N \) can be expressed as the disjoint union of local deterministic models and strongly contextual models.
- Extreme contextual probability models are completely determined by their support sets.

3. **General version of the theorem**:
- **Theorem 1.2**: For any polytope \( P \), if \( P = H_{\geq 0}\cap\text{Aff}(P) \), where \( H_{\geq 0} \) is the non - negative orthant and \(\text{Aff}(P)\) is the affine subspace generated by \( P \). Let \( F(P) \) be the face lattice of \( P \) and \( S(P) \) be the "support sets" of the points in \( P \), that is, 0/1 vectors, with positive components mapped to 1 and zero components remaining unchanged. Then there exists an order - isomorphism \( F(P)\sim S(P) \).

### Application to probability polytopes

- **Theorem 6.1**: For any measurement scenario \( \Sigma \), there exists an order - isomorphism between the face lattice of the no - signaling polytope \( N(\Sigma) \) and its support set lattice.
- **Corollary 6.2**:
- Two no - signaling empirical models have the same support set if and only if they are located in the relative interior of the same face.
- The vertices of the no - signaling polytope \( N(\Sigma) \) are precisely those no - signaling empirical models with the smallest support sets.
- A no - signaling empirical model \( e \) is a vertex if and only if it is the only model in \( N(\Sigma) \) with the support set \(\text{supp}(e) \).
- The empirical models in the relative interior of the no - signaling polytope \( N(\Sigma) \) are precisely those models whose support sets are the universal set.

### Conclusion

By proving that the combinatorial structure of the no - signaling polytope is completely determined by its support sets, this paper reveals the profound connection between probability models and their simplified possibility models. This result not only provides a new theoretical perspective but also provides a powerful tool for understanding and analyzing probability models in quantum information and fundamental research in practical applications.