In mathematics, representation is a very general relationship that expresses similarities between objects. Roughly speaking, a collection Y of mathematical objects may be said to represent another collection X of objects, provided that the properties and relationships existing among the representing objects y_{i} conform in some consistent way to those existing among the corresponding represented objects x_{i}. Somewhat more formally, for a set Π of properties and relations, a Π-representation of some structure X is a structure Y that is the image of X under a homomorphism that preserves Π. The label representation is sometimes also applied to the homomorphism itself.
Perhaps the most well-developed example of this general notion is the subfield of abstract algebra called representation theory, which studies the representing of elements of algebraic structures by linear transformations of vector spaces.
Although the term representation theory is well established in the algebraic sense discussed above, there are many other uses of the term representation throughout mathematics.
An active area of graph theory is the exploration of isomorphisms between graphs and other structures. A key class of such problems stems from the fact that, like adjacency in undirected graphs, intersection of sets (or, more precisely, non-disjointness) is a symmetric relation. This gives rise to the study of intersection graphs for innumerable families of sets.^{[1]} One foundational result here, due to Paul Erdős and colleagues, is that every n-vertex graph may be represented in terms of intersection among subsets of a set of size no more than n^{2}/4.^{[2]}
Representing a graph by such algebraic structures as its adjacency matrix and Laplacian matrix gives rise to the field of spectral graph theory.^{[3]}
Dual to the observation above that every graph is an intersection graph is the fact that every partially ordered set is isomorphic to a collection of sets ordered by the containment (or inclusion) relation ⊆. Among the posets that arise as the containment orders for natural classes of objects are the Boolean lattices and the orders of dimension n.^{[4]}
Many partial orders arise from (and thus can be represented by) collections of geometric objects. Among them are the n-ball orders. The 1-ball orders are the interval-containment orders, and the 2-ball orders are the so-called circle orders, the posets representable in terms of containment among disks in the plane. A particularly nice result in this field is the characterization of the planar graphs as those graphs whose vertex-edge incidence relations are circle orders.^{[5]}
There are also geometric representations that are not based on containment. Indeed, one of the best studied classes among these are the interval orders,^{[6]} which represent the partial order in terms of what might be called disjoint precedence of intervals on the real line: each element x of the poset is represented by an interval [x_{1}, x_{2}] such that for any y and z in the poset, y is below z if and only if y_{2} < z_{1}.
Under certain circumstances, a single function f:X → Y is at once an isomorphism from several mathematical structures on X. Since each of those structures may be thought of, intuitively, as a meaning of the image Y—one of the things that Y is trying to tell us—this phenomenon is called polysemy, a term borrowed from linguistics. Examples include: