|  Types and Realizations of Posets | 
In General > s.a. Hasse Diagram.
  * Well founded: A poset P
    is well founded if no infinite decreasing sequences occur in P.
  * Well partially ordered: A well
    founded poset containing no infinite antichains.
  * Locally finite: A poset such that
    every interval in it is finite.
  * Prime poset: One such that all
    its autonomous subsets are trivial.
  @ General references: Bosi et al Ord(01) [interval orders, numerical representation];
    Hubička & Nešetřil EJC(05) [universal partial order].
  @ Types: Ng Ord(05) [linear discrepancy n−2];
    Malicki & Rutkowski Ord(05) [well partially ordered];
    Pouzet & Zaguia Ord(09),
    Boudabbous et al Ord(10) [prime];
    Dzhafarov Ord(11) [infinite saturated orders];
    Bonanzinga & Matveev Ord(11) [thin posets];
    Pouzet et al EJC(14)
      [posets embeddable in a product of finitely many scattered chains];
    > s.a. Relations [k-homogeneous].
Sphere Order > s.a. 2D geometries [Lorentzian].
  * Idea: A poset realization
    defined by sphere containment in Euclidean space.
  * Result: Every circle order
    has a normal representation, ∩i
    int(Pi) ≠ 0.
  * Result: For all n
    > 2, ∃ P with dim(P) = n which is not a
    circle order;  Example: [2] × [3] × \(\mathbb N\) is not,
    but perhaps all 3D finite posets are; Some infinite posets are not sphere orders in any dimension!
    [@ Fon-Der-Flaass Ord(93)].
  @ References: Fishburn Ord(88);
    Scheinerman & Wierman Ord(88);
    Hurlbert Ord(88);
    Sidney et al Ord(88);
    Brightwell & Winkler Ord(89);
    Lin Ord(91),
    Scheinerman Ord(92) [circle orders];
    Meyer Ord(93);
    Scheinerman & Tanenbaum Ord(97);
    Fishburn & Trotter Ord(98) [survey];
    Vatandoost & Bahrampour JMP(11) [and spacetime].
Other Realizations
  * Path order: A representation
    by finite oriented paths ordered by the existence of homomorphism between them;
    Any countable (finite or infinite) partially ordered set can be realized this way.
  * Function order: A proper set of
    functions fi: [0,1] →
    \(\mathbb R\) (two functions intersect, and cross, only a finite no of times,
    but not at 0 or 1), with pi
    < pj iff
    fi(x)
    < fj(x) for all
    x ∈ [0,1]; Every poset can be represented as a function order.
  @ References:
    Sydney, Sydney & Urrutia Ord(88) [function orders];
    Hubička & Nešetřil Ord(04),
    Ord(05) [path orders];
    Laison Ord(04)
      [(n, i, f)-tube orders];
    Berman & Blok Ord(06) [as algebras].
Special Posets
  > s.a. set theory [directed set]; lattices.
  * Discrete poset:  One in which no two elements are comparable.
  * Binomial poset: It is
    locally finite; All maximal chains between two points have the same length;
    Any two n-intervals contain the same number of maximal chains.
  * Linear order:
    Ln = The totally
    ordered n-element poset; They are all 1-dimensional.
  * Standard n-dimensional
    poset: Hn = (u1,
    ..., un, v1,
    ..., vn);
    ui < vj
    for i ≠ j; χ(Hn)
    = 2 for all n ≥ 3.
  * Unimodular poset: A ranked poset
    is unimodular if, for some j, r0
    ≤ r1 ≤ ...
    ≤ rj
    ≥ rj+1 ≥ ...
    ≥ rn,
    with rk
    the number of elements of rank k.
  * Sperner poset: A ranked poset
    which has a maximum antichain all of whose elements have the same rank.
  * Spectral poset: One which is
    order isomorphic to (Spec(R), ⊆) where Spec(R) is the
    set of prime ideals of some commutative ring with identity R.
  * Multipartite poset: A poset
    P = (X, ≤) is m-partite if X has a
    partition X = X1 ∩
    ... ∩ Xm such that
    (1) each Xi forms an
    antichain in P, and (2) x ≤ y implies x
    ∈ Xi and y
    ∈ Xj, where i
    < j.
  * Result: (Schnyder's Theorem)
    The incidence poset of a graph G has dimension at most three if and only if G is planar;
    > s.a. Wikipedia page. 
  @ Completely regular:
    Künzi Ord(90).
  @ Boolean, Distributive, Modular:
    Chajda & Rachunek Ord(89);
    Niederle Ord(95).
  @ Planar: Scheinerman pr(90);
    Hashemi et al Ord(97);
    Barrera-Cruz & Haxell Ord(11) [Schnyder's theorem].
  @ Other: Behrendt AC(90) [homogeneous];
    Hanna & McMaster Ord(00) [splittability];
    Bezrukov et al JCTA(04) [Macaulay posets];
    Belaid & Echi T&A(04) [spectral];
    Waphare & Joshi Ord(05) [uniquely complemented];
    Agnarsson DM(08) [multipartite posets, dimension];
    Kwaśniewski a0901,
    a0903 [graded posets, zeta matrix];
    Conflitti DM(09) [fences and crowns, Whitney numbers of order ideals];
    Kukieła Ord(09) [reversible and bijectively related];
    Streib & Trotter EJC(14) [posets with planar cover graphs];
    Trotter & Wang Ord(14) [incidence posets of graphs, cover graphs of posets].
Random Posets > s.a. Relation.
  * (i) Uniform random orders:
    As n → ∞, most n-element orders have 3
    levels; This is used for poset enumeration; 2004, We do not have a good way
    to  generate those random individual posets; 2015, Algorithm for generating uniform random orders.
  * (ii) Random k-dimensional orders:
    [Winkler Ord(85)]
    Pk(n)
    is the intersection of k linear orders obtained from
    random permutations of the  n labels, and the result
    is a k-dimensional order; The height is ~ C
    n1/k and the width
    ~ C' n1−1/k;
    A variation consists in sprinkling points in Minkowski space, although there
    are finite posets which cannot be embedded in flat spacetime in any dimension.
  * (iii) Random graph orders:
    (a.k.a. transitive percolation) Given any p ∈ [0,1],
    for every pair 0 < i < j ≤ n, draw
    a graph edge between ai and
    aj with probability p
    and interpret it as ai
    < aj, and take the
    transitive closure of the resulting relation; On average the poset will look
    like a series of "blobs" connected by  posts which are
    related to every other element (the fraction of posts is
    ~ exp{−π2/3p} as n → ∞)
    [@ Alon et al AAP(94)]; Can be generalized to classical
    growth models for causal sets, the only generalization that satisfies "general
    covariance" and "Bell causality" [@ Rideout & Sorkin].
  * Results: For n → ∞,
    every first-order property holds either in nearly all unlabelled 2D orders, or hardly any.
  @ Uniform random orders: Compton I&C(88) [0-1 laws for first-order logic]; Henson et al a1504 [algorithm].
  @ Random k-dimensional orders:
    Winkler Ord(85),
    Ord(85),
    Ord(89) [general];
    Winkler Ord(90) [2D];
    Spencer Ord(90) [2D, first-order properties];
    Bollobás & Brightwell AAP(92) [height];
    Brightwell Ord(92) [width and number of linear extensions].
  @ Random graph orders: Albert & Frieze Ord(89) [height, width, dimension, etc];
    Bollobás & Brightwell TAMS(91)*;
    Luczak Ord(91) [first-order properties];
    Bollobás & Brightwell MathSci(95) [width];
    Kim & Pittel JCTB(00) [interpost distance distribution];
    Bombelli et al a0809 [finite, expected number of posts].
Examples and Applications
  > s.a. Clustering; connection;
  Filters; logic; topological
  spaces; Well Ordering.
  * On the set of partitions of an
    integer m: A partition P1
    precedes P2 if the elements of
    P2 can be decomposed to give
    those of P1; For  m = 5,
    (2,2,1) < (3,2).
  * In differential geometry:
    Can be used to sample Lorentzian manifolds and define distances between them.
  * In graph theory: The
    set of acyclic orientations of a connected graph with a given sink has
    a natural poset structure.
  * In quantum gravity:
    The poset could be either spacetime, with causality as a partial ordering,
    or space; > see causal sets.
  @ Examples: Ehrenborg & Slone Ord(09) [acyclic orientations of a graph];
    Endo et al T&A(10) [knots and links].
  @ In physics:
    Borchers & Sen CMP(90) [for physicists],
    CMP(99);
    Bombelli JMP(00)gq [on Lorentzian manifolds];
    Baumgartner JPA(02)mp/01 [subsystems];
    Coecke qp/02  [probability];
    Sen 10 [causality, measurement and discrete vs differentiable structures].
 main page
  – abbreviations
  – journals – comments
  – other sites – acknowledgements
  send feedback and suggestions to bombelli at olemiss.edu – modified 12 jun 2020