{
  "id": "supremum",
  "title": "Supremum",
  "lang": "en",
  "source_lang": "en",
  "summary": "The supremum of a set is its least upper bound, whether or not that bound belongs to the set.",
  "category": "mathematics",
  "subcategory": "analysis",
  "type": "concept",
  "aliases": [
    "Least upper bound",
    "Sup"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/supremum/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/supremum/",
      "json": "https://aiwiki.page/zh/supremum.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The supremum of a set is the least element that is greater than or equal to every element of that set, relative to a specified ordering. Written \\(\\sup A\\), it gives the sharpest upper bound on \\(A\\), but need not belong to \\(A\\). The concept is fundamental in mathematical analysis, where it expresses the completeness of the real numbers, and in the study of partially ordered sets, where it generalizes the idea of taking a maximum."
    },
    {
      "heading": "Definition",
      "text": "Let \\(A\\) be a subset of a partially ordered set \\((P,\\leq)\\). An element \\(u\\in P\\) is an upper bound of \\(A\\) if\n\n\\[a\\leq u\\qquad\\text{for every }a\\in A.\\]\n\nAn element \\(s\\in P\\) is the supremum of \\(A\\) if:\n\n- \\(s\\) is an upper bound of \\(A\\);\n- \\(s\\leq u\\) for every upper bound \\(u\\in P\\) of \\(A\\).\n\nThus, a supremum must both bound the set from above and be no larger than any other upper bound. If it exists, it is unique: if \\(s\\) and \\(t\\) both satisfy the definition, then \\(s\\leq t\\) and \\(t\\leq s\\), so antisymmetry gives \\(s=t\\).\n\nThe order-dual concept is the infimum, or greatest lower bound, denoted \\(\\inf A\\). It is obtained by reversing all inequalities in the definition.\n\nFor an indexed collection, common notation includes\n\n\\[\\sup_{i\\in I}a_i \\quad\\text{and}\\quad \\sup\\{a_i:i\\in I\\}.\\]\n\nFor a function \\(f:D\\to\\mathbb R\\), the expression\n\n\\[\\sup_{x\\in D}f(x)\\]\n\nmeans the supremum of its set of values \\(\\{f(x):x\\in D\\}\\)."
    },
    {
      "heading": "Supremum versus maximum",
      "text": "A maximum of \\(A\\) is an element of \\(A\\) that is greater than or equal to every other element of \\(A\\). A supremum is not required to belong to \\(A\\).\n\nConsequently:\n\n- If \\(A\\) has a maximum, then \\(\\sup A=\\max A\\).\n- If \\(\\sup A\\) exists and belongs to \\(A\\), then it is the maximum.\n- A set can have a supremum without having a maximum.\n\nFor example,\n\n\\[\\sup(0,1)=1,\\]\n\nbut the open interval \\((0,1)\\) has no maximum. Every number in it is smaller than another number in the same interval. By contrast,\n\n\\[\\sup[0,1]=\\max[0,1]=1.\\]\n\nEvery nonempty finite subset of a totally ordered set has a maximum and therefore a supremum. This does not extend to arbitrary partial orders: two incomparable elements need not have either a maximum or a supremum."
    },
    {
      "heading": "Supremum in the real numbers",
      "text": "The real numbers have the least-upper-bound property: every nonempty subset of \\(\\mathbb R\\) that is bounded above has a supremum in \\(\\mathbb R\\). This is a form of order completeness and can be used as an axiom characterizing the real number system among ordered fields.\n\nThe assumptions are essential. A real set that is unbounded above has no real supremum. The empty set also has no supremum in \\(\\mathbb R\\), since every real number is an upper bound of it and there is no least real number.\n\nApproximation characterization\n\nFor a nonempty set \\(A\\subseteq\\mathbb R\\), a real number \\(s\\) equals \\(\\sup A\\) if and only if:\n\n- \\(a\\leq s\\) for every \\(a\\in A\\);\n- for every \\(\\varepsilon>0\\), there exists \\(a\\in A\\) such that\n\n\\[s-\\varepsilon<a\\leq s.\\]\n\nThe second condition says that elements of \\(A\\) can be found arbitrarily close to \\(s\\) from below, possibly equal to \\(s\\). If it failed for some \\(\\varepsilon>0\\), then \\(s-\\varepsilon\\) would itself be an upper bound, contradicting leastness.\n\nFor example, let\n\n\\[A=\\left\\{1-\\frac1n:n=1,2,3,\\ldots\\right\\}.\\]\n\nEvery element is less than \\(1\\), and the elements approach \\(1\\), so \\(\\sup A=1\\), although \\(1\\notin A\\).\n\nDependence on the ambient ordered set\n\nA supremum is always relative to an ambient ordered set. Consider\n\n\\[A=\\{q\\in\\mathbb Q:q\\geq0,\\ q^2<2\\},\\]\n\nwhere \\(\\mathbb Q\\) denotes the rational numbers.\n\nViewed as a subset of \\(\\mathbb R\\), this set has supremum \\(\\sqrt2\\). Viewed as a subset of \\(\\mathbb Q\\), it is bounded above but has no supremum, because its least real upper bound is not rational. This illustrates the failure of order completeness in \\(\\mathbb Q\\).\n\nMore generally, changing the ambient ordered set can change the supremum or whether one exists."
    },
    {
      "heading": "Extended real numbers and the empty set",
      "text": "In the extended real line\n\n\\[\\overline{\\mathbb R}=\\mathbb R\\cup\\{-\\infty,+\\infty\\},\\]\n\nevery subset has a supremum. An unbounded-above subset of \\(\\mathbb R\\) has supremum \\(+\\infty\\), and\n\n\\[\\sup\\varnothing=-\\infty, \\qquad \\inf\\varnothing=+\\infty.\\]\n\nThese conventions follow directly from the order definition. Every element of the ambient ordered set is an upper bound of the empty set, so its supremum, when one exists, is the least element of that ambient set.\n\nAccordingly, \\(\\sup\\varnothing\\) depends on the setting: in the ordered interval \\([0,1]\\), it is \\(0\\), whereas in \\(\\mathbb R\\) it does not exist."
    },
    {
      "heading": "Basic properties",
      "text": "For nonempty, bounded-above real sets \\(A\\) and \\(B\\), the following identities and inequalities hold.\n\nMonotonicity. If \\(A\\subseteq B\\), then\n\n\\[\\sup A\\leq\\sup B.\\]\n\nUnion.\n\n\\[\\sup(A\\cup B)=\\max\\{\\sup A,\\sup B\\}.\\]\n\nTranslation. For \\(c\\in\\mathbb R\\),\n\n\\[\\sup\\{a+c:a\\in A\\}=\\sup A+c.\\]\n\nPositive scaling. For \\(\\lambda>0\\),\n\n\\[\\sup\\{\\lambda a:a\\in A\\}=\\lambda\\sup A.\\]\n\nOrder reversal. If \\(A\\) is also bounded below, then for \\(\\lambda<0\\),\n\n\\[\\sup\\{\\lambda a:a\\in A\\}=\\lambda\\inf A.\\]\n\nIn particular,\n\n\\[\\inf A=-\\sup\\{-a:a\\in A\\}.\\]\n\nIndependent sums.\n\n\\[\\sup\\{a+b:a\\in A,\\ b\\in B\\}=\\sup A+\\sup B.\\]\n\nHere the two variables vary independently. For functions on the same nonempty domain, only an inequality holds in general:\n\n\\[\\sup_x(f(x)+g(x)) \\leq \\sup_x f(x)+\\sup_x g(x),\\]\n\nassuming the two suprema on the right are finite. Equality can fail because the functions may be large at different points."
    },
    {
      "heading": "Sequences, limits, and attainment",
      "text": "If \\(A\\subseteq\\mathbb R\\) is nonempty and has finite supremum \\(s\\), then for every positive integer \\(n\\), an element \\(a_n\\in A\\) can be chosen so that\n\n\\[s-\\frac1n<a_n\\leq s.\\]\n\nThe resulting sequence converges to \\(s\\). Thus every finite real supremum can be approached by a sequence of elements of the set.\n\nFor an increasing real sequence \\((x_n)\\) bounded above,\n\n\\[\\lim_{n\\to\\infty}x_n=\\sup\\{x_n:n\\geq1\\}.\\]\n\nThis is the monotone convergence theorem for sequences, a basic consequence of real completeness. The limit need not equal any individual term.\n\nWhether a function attains its supremum requires additional conditions. The extreme value theorem states that a continuous real-valued function on a nonempty compact space attains both its supremum and its infimum. Without suitable hypotheses, the supremum may be finite but unattained, as with \\(f(x)=x\\) on \\((0,1)\\).\n\nSuprema also define the limit superior of a sequence:\n\n\\[\\limsup_{n\\to\\infty}x_n = \\inf_{n\\geq1}\\sup_{k\\geq n}x_k,\\]\n\ninterpreted in the extended real line. This construction records the limiting behavior of the upper bounds of successive tails."
    },
    {
      "heading": "Suprema in partial orders",
      "text": "In a partial order, the existence of upper bounds does not guarantee a supremum. For example, suppose \\(P=\\{a,b,u,v\\}\\), with \\(a\\) and \\(b\\) incomparable, both below \\(u\\) and \\(v\\), and \\(u\\) and \\(v\\) incomparable. Then \\(\\{a,b\\}\\) has two upper bounds but no least upper bound.\n\nFor a pair of elements, the supremum is often called their join, written\n\n\\[a\\vee b=\\sup\\{a,b\\}.\\]\n\nA lattice is a partially ordered set in which every pair has both a supremum and an infimum. A complete lattice has suprema and infima for all subsets, including the empty set.\n\nExamples show that a supremum need not resemble a numerical maximum:\n\n- In the power set of a set \\(X\\), ordered by inclusion, the supremum of a family of subsets is their union.\n- Among linear subspaces of a vector space, ordered by inclusion, the supremum of a family is the linear span of their union, not generally the union itself.\n- Among positive integers ordered by divisibility, the supremum of two integers is their least common multiple.\n\nThese examples also show why a supremum can exist even when the elements being compared are incomparable."
    },
    {
      "heading": "Applications in analysis and optimization",
      "text": "Optimization\n\nIn mathematical optimization, the value\n\n\\[v=\\sup_{x\\in F}f(x)\\]\n\nis the best upper value of an objective function over a feasible set \\(F\\). This value is distinct from an optimizer: an optimizer is a feasible point \\(x^\\ast\\) satisfying \\(f(x^\\ast)=v\\).\n\nFor example,\n\n\\[\\sup_{x>0}\\frac{x}{1+x}=1,\\]\n\nbut no finite positive \\(x\\) attains that value. Writing “supremum” rather than “maximum” avoids assuming that an optimizer exists.\n\nFunction norms and uniform convergence\n\nFor a bounded real- or complex-valued function on a nonempty set \\(D\\), the supremum norm is\n\n\\[\\|f\\|_\\infty=\\sup_{x\\in D}|f(x)|.\\]\n\nIt measures the greatest magnitude of the function, allowing for the possibility that this magnitude is approached but never attained.\n\nThe corresponding distance\n\n\\[\\|f-g\\|_\\infty=\\sup_{x\\in D}|f(x)-g(x)|\\]\n\nis central to uniform convergence: a sequence \\(f_n\\) converges uniformly to \\(f\\) precisely when these suprema tend to zero.\n\nEssential supremum\n\nIn measure theory, the essential supremum disregards exceptional sets of measure zero. For a measurable real-valued function,\n\n\\[\\operatorname{ess\\,sup}f = \\inf\\{M\\in\\mathbb R:f\\leq M\\text{ almost everywhere}\\},\\]\n\nwith extended-real conventions when needed.\n\nFor example, on \\([0,1]\\) with Lebesgue measure, a function equal to \\(1\\) at one point and \\(0\\) elsewhere has ordinary supremum \\(1\\) but essential supremum \\(0\\). The distinction is fundamental in L^p spaces, particularly \\(L^\\infty\\), where the norm is the essential supremum of the absolute value rather than the pointwise supremum."
    }
  ],
  "links": [
    "mathematical-analysis",
    "partial-order",
    "subset",
    "mathematical-function",
    "real-number",
    "rational-number",
    "empty-set",
    "limit",
    "extreme-value-theorem",
    "continuous-function",
    "compact-space",
    "power-set",
    "linear-subspace",
    "linear-span",
    "mathematical-optimization",
    "feasible-set",
    "uniform-convergence",
    "measure-theory",
    "lp-space"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}