Swapnonil Banerjee
Function and Functional
We know a function is an input-output relationship. A
function takes a number as an input and spits out
another as an output. For example, for the input 𝑥 = 3,
the function 𝑓(𝑥) = 𝑥 2 assumes the value 𝑓(3) = 32 = 9.
A “functional” takes an entire function as the input and
spits out a single number as the output. For example, an
𝑡
integral, such as, ∫𝑡 𝑏 𝑥 2 (𝑡)𝑑𝑡 is a functional, since for
𝑎
the function 𝑥(𝑡) defined over an interval of 𝑡, the integral produces a single number.
𝑡
Symbolically, we can write a functional as 𝐹[𝑥(𝑡)]. For our problem, 𝐹[𝑥(𝑡)] = ∫𝑡 𝑏 𝑥 2 (𝑡)𝑑𝑡 .
𝑎
𝑡
Let’s evaluate 𝐹[𝑥(𝑡)] = ∫𝑡 𝑏 𝑥 2 (𝑡)𝑑𝑡 for 𝑥(𝑡) = 𝑡, and 𝑡𝑎 = 0, 𝑡𝑏 = 3.
𝑎
𝑡
𝐹[𝑥(𝑡)] = ∫𝑡 𝑏 𝑥 2 (𝑡)𝑑𝑡
𝑎
3
= ∫0 𝑡 2 𝑑𝑡 [Since 𝑥(𝑡) = 𝑡, 𝑥 2 (𝑡) = 𝑡 2 ]
𝑡=3
𝑡3
= 3| =9
𝑡=0
3 𝑑𝑥 2
Next let’s evaluate another functional 𝐹[𝑥(𝑡)] = ∫0 ( 𝑑𝑡 ) 𝑑𝑡 for the same function 𝑥(𝑡), i.e.,
𝑥(𝑡) = 𝑡.
𝑑𝑥 3 𝑑𝑥 2 3
Since, 𝑥(𝑡) = 𝑡, = 1. Hence, ∫0 ( 𝑑𝑡 ) 𝑑𝑡 = ∫0 1𝑑𝑡 = 3 − 0 = 3. So, 𝐹[𝑥(𝑡)] = 3 in this
𝑑𝑡
case.
The arc-length of an arbitrary curve is also a functional, since the arc length involves an integral
and depends on the curve (function) in question.
In physics, a functional called “action” plays an important role. The action (𝑆) is defined as
𝑡
𝑆 = ∫𝑡 𝑏(𝐾𝐸 − 𝑃𝐸)𝑑𝑡, where 𝐾𝐸 and 𝑃𝐸 are the kinetic energy and the potential energy of an
𝑎
object, and 𝑡 stands for time with 𝑡𝑎 and 𝑡𝑏 being the initial and final times respectively.
Swapnonil Banerjee
1
For example, an apple (of mass 𝑚), falling under gravity, will have a kinetic energy 2 𝑚𝑣 2 , where
𝑣 is the apple’s speed at any instant. Hence, if 𝑦(𝑡) denotes the apple’s distance from tree-top
𝑑𝑦 1 𝑑𝑦 2
at any instant, 𝑣 = and 𝐾𝐸 = 2 𝑚 ( 𝑑𝑡 ) . The apple’s gravitational potential energy is 𝑃𝐸 =
𝑑𝑡
−𝑚𝑔𝑦, where 𝑔 is the gravitational acceleration (we consider the downward direction to be
positive 𝑦 direction). Hence, the action for the falling apple under gravitation is
𝑡𝑏 𝑡𝑏
1 𝑑𝑦 2
𝑆 = ∫ (𝐾𝐸 − 𝑃𝐸)𝑑𝑡 = ∫ ( 𝑚 ( ) + 𝑚𝑔𝑦) 𝑑𝑡
𝑡𝑎 𝑡𝑎 2 𝑑𝑡
Next we consider a block attached to a spring that is stretched by an amount 𝑥. The kinetic energy
1 𝑑𝑥 2 1
of the block is 𝐾𝐸 = 2 𝑚 ( 𝑑𝑡 ) . The potential energy of spring-block system is 𝑃𝐸 = 2 𝑘𝑥 2 .
Hence, the action for the spring-block system
𝑡 𝑡 1 𝑑𝑥 2 1
𝑆 = ∫𝑡 𝑏(𝐾𝐸 − 𝑃𝐸)𝑑𝑡 = ∫𝑡 𝑏 (2 𝑚 ( 𝑑𝑡 ) − 2 𝑘𝑥 2 ) 𝑑𝑡
𝑎 𝑎
There is a deep connection between the action and Newton’s law of motion, which I will explain
in another post.