{{Short description|Negative of a convex function}} {{Use American English|date = January 2019}}

In [[mathematics]], a '''concave function''' is one for which the function value at any convex combination of elements in the domain is greater than or equal to that convex combination of those domain elements. Equivalently, a concave function is any function for which the [[Hypograph (mathematics)|hypograph]] is convex. The class of concave functions is in a sense the opposite of the class of [[convex function]]s. A concave function is also [[synonym]]ously called '''concave downwards''', '''concave down''', '''convex upwards''', '''convex cap''', or '''upper convex'''.

==Definition== A real-valued [[function (mathematics)|function]] f on an [[interval (mathematics)|interval]] (or, more generally, a [[convex set]] in [[vector space]]) is said to be ''concave'' if, for any x and y in the interval and for any \alpha \in [0,1],{{cite book |last1=Lenhart |first1=S. |last2=Workman |first2=J. T. |title=Optimal Control Applied to Biological Models |publisher=Chapman & Hall/ CRC |series=Mathematical and Computational Biology Series |year=2007 |isbn=978-1-58488-640-2 }}

:f((1-\alpha )x+\alpha y)\geq (1-\alpha ) f(x)+\alpha f(y)

A function is called ''strictly concave'' if

:f((1-\alpha )x+\alpha y) > (1-\alpha ) f(x)+\alpha f(y)

for any \alpha \in (0,1) and x \neq y.

For a function f: \mathbb{R} \to \mathbb{R}, this second definition merely states that for every z strictly between x and y, the point (z, f(z)) on the graph of f is above the straight line joining the points (x, f(x)) and (y, f(y)).

[[Image:ConcaveDef.png|class=skin-invert-image]]

A function f is [[quasiconvex function|quasiconcave]] if the upper contour sets of the function S(a)={x: f(x)\geq a} are convex sets.{{Cite book|last=Varian, Hal R.|title=Microeconomic analysis|date=1992|publisher=Norton|isbn=0-393-95735-7|edition=3rd|location=New York|pages=489|oclc=24847759}}

==Properties== [[File:cubic_graph_special_points_repeated.svg|class=skin-invert-image|thumb|A cubic function is concave (left half) when its first derivative (red) is monotonically decreasing i.e. its second derivative (orange) is negative, and convex (right half) when its first derivative is monotonically increasing i.e. its second derivative is positive]] ===Functions of a single variable===

A [[differentiable function]] {{mvar|f}} is (strictly) concave on an [[interval (mathematics)|interval]] if and only if its [[derivative]] function {{mvar|f ′}} is (strictly) [[monotonically decreasing]] on that interval, that is, a concave function has a non-increasing (decreasing) [[slope]].{{Cite book| last=Rudin| first=Walter| title=Analysis| year=1976| pages= 101}}{{Cite journal |last1=Gradshteyn|first1=I. S.| last2=Ryzhik|first2=I. M.| last3=Hays|first3=D. F.| date=1976-07-01| title=Table of Integrals, Series, and Products| journal=Journal of Lubrication Technology| volume=98|issue=3|pages=479| doi=10.1115/1.3452897|issn=0022-2305 |doi-access=free}}

[[Point (geometry)|Points]] where concavity changes (between concave and [[convex function|convex]]) are [[inflection point]]s.{{Cite book|last=Hass, Joel | title=Thomas' calculus| others=Heil, Christopher, 1960-, Weir, Maurice D.,, Thomas, George B. Jr. (George Brinton), 1914-2006.|date=13 March 2017| isbn=978-0-13-443898-6| edition=Fourteenth| location=[United States]| pages=203| oclc=965446428}}

If {{mvar|f}} is twice-[[Differentiable function|differentiable]], then {{mvar|f}} is concave [[if and only if]] {{mvar|f ′′}} is [[non-positive]] (or, informally, if the "[[acceleration]]" is non-positive). If {{mvar|f ′′}} is [[negative numbers|negative]] then {{mvar|f}} is strictly concave, but the converse is not true, as shown by {{math|1=''f''(''x'') = −''x''4}}.

If {{mvar|f}} is concave and differentiable, then it is bounded above by its first-order [[Taylor approximation]]: f(y) \leq f(x) + f'(x)[y-x]

A [[Lebesgue measurable function]] on an interval {{math|'''C'''}} is concave [[if and only if]] it is midpoint concave, that is, for any {{mvar|x}} and {{mvar|y}} in {{math|'''C'''}} f\left( \frac{x+y}2 \right) \ge \frac{f(x) + f(y)}2

If a function {{mvar|f}} is concave, and {{math|''f''(0) ≥ 0}}, then {{mvar|f}} is [[subadditivity|subadditive]] on [0,\infty). Proof:

#* Since {{mvar|f}} is concave and {{math|1 ≥ t ≥ 0}}, letting {{math|1=''y'' = 0}} we have f(tx) = f(tx+(1-t)\cdot 0) \ge t f(x)+(1-t)f(0) \ge t f(x) . #* For a,b\in[0,\infty): f(a) + f(b) = f \left((a+b) \frac{a}{a+b} \right) + f \left((a+b) \frac{b}{a+b} \right) \ge \frac{a}{a+b} f(a+b) + \frac{b}{a+b} f(a+b) = f(a+b)

===Functions of ''n'' variables===

A function {{mvar|f}} is concave over a convex set [[if and only if]] the function {{mvar|−f}} is a [[convex function]] over the set.

The sum of two concave functions is itself concave and so is the [[pointwise minimum]] of two concave functions, i.e. the set of concave functions on a given domain form a [[semifield]].

Near a strict [[local maximum]] in the interior of the domain of a function, the function must be concave; as a partial converse, if the derivative of a strictly concave function is zero at some point, then that point is a local maximum.

Any [[local maximum]] of a concave function is also a [[global maximum]]. A ''strictly'' concave function will have at most one global maximum.

==Examples==

• The functions f(x)=-x^2 and g(x)=\sqrt{x} are concave on their domains, as their second derivatives f''(x) = -2 and g''(x) =-\frac{1}{4 x^{3/2}} are always negative.
• The [[logarithm]] function f(x) = \log{x} is concave on its domain (0,\infty), as its derivative \frac{1}{x} is a strictly decreasing function.
• Any [[affine function]] f(x)=ax+b is both concave and convex, but neither strictly-concave nor strictly-convex.
• The [[sine]] function is concave on the interval [0, \pi].
• The function f(B) = \log |B|, where |B| is the [[determinant]] of a [[nonnegative-definite matrix]] ''B'', is concave.{{cite journal |author-link=Thomas M. Cover |first1=Thomas M. |last1=Cover |first2=J. A. |last2=Thomas |s2cid=5491763 |title=Determinant inequalities via information theory| journal=[[SIAM Journal on Matrix Analysis and Applications]]| year=1988| volume=9|number=3| pages=384–392| doi=10.1137/0609033}}

==Applications==

• Rays bending in the [[computation of radiowave attenuation in the atmosphere]] involve concave functions.
• In [[expected utility]] theory for [[choice under uncertainty]], [[cardinal utility]] functions of [[risk aversion|risk averse]] decision makers are concave.
• In [[microeconomic theory]], [[production function]]s are usually assumed to be concave over some or all of their domains, resulting in [[diminishing returns]] to input factors.{{cite book |first1=Malcolm |last1=Pemberton |first2=Nicholas |last2=Rau |title=Mathematics for Economists: An Introductory Textbook |publisher=Oxford University Press |year=2015 |isbn=978-1-78499-148-7 |pages=363–364 |url=https://books.google.com/books?id=9j5_DQAAQBAJ&pg=PA363 }}
• In [[thermodynamics]] and [[information theory]], [[Entropy (information theory)|entropy]] is a concave function. In the case of thermodynamic entropy, without phase transition, entropy as a function of extensive variables is strictly concave. If the system can undergo phase transition, and if it is allowed to split into two subsystems of different phase ([[phase separation]], e.g. boiling), the entropy-maximal parameters of the subsystems will result in a combined entropy precisely on the straight line between the two phases. This means that the "effective entropy" of a system with phase transition is the [[convex envelope]] of entropy without phase separation; therefore, the entropy of a system including phase separation will be non-strictly concave.{{Cite book |last1=Callen |first1=Herbert B. |title=Thermodynamics and an introduction to thermostatistics |last2=Callen |first2=Herbert B. |date=1985 |publisher=Wiley |isbn=978-0-471-86256-7 |edition=2nd |location=New York |pages=203–206 |chapter=8.1: Intrinsic Stability of Thermodynamic Systems}}

==See also==

• [[Concave polygon]]
• [[Jensen's inequality]]
• [[Logarithmically concave function]]
• [[Quasiconcave function]]
• [[Concavification]]

==References== {{reflist}}

==Further References==

*{{cite book|last=Crouzeix|first=J.-P.|chapter=Quasi-concavity|title=The New Palgrave Dictionary of Economics|editor-first=Steven N.|editor-last=Durlauf|editor2-first=Lawrence E|editor2-last=Blume|publisher=Palgrave Macmillan|year=2008|edition= Second|pages=815–816|chapter-url=http://www.dictionaryofeconomics.com/article?id=pde2008_Q000008|doi=10.1057/9780230226203.1375|isbn=978-0-333-78676-5}} *{{cite book |title=Engineering Optimization: Theory and Practice|first=Singiresu S.|last=Rao| publisher=John Wiley and Sons|year=2009|isbn=978-0-470-18352-6|page=779}}

{{Calculus topics}} {{Convex analysis and variational analysis}} {{Authority control}}

[[Category:Convex analysis]] [[Category:Types of functions]]