MHB What is the minimum value of y in the Absolute Value Function?

anemone
Gold Member
MHB
POTW Director
Messages
3,851
Reaction score
115
Determine the minimum of $y$ where $y=|x-1|+|2x-1|+|3x-1|+\cdots+|nx-1|$.
 
Mathematics news on Phys.org
The minimum of $y$ is zero and is achieved when $n = 1$ and $x = 1$.

Seriously though, should we minimize $y$ over $x$ as a function of $n$? Also, are $x, y$ real? (I assume $n$ is a natural number).
 
Bacterius said:
The minimum of $y$ is zero and is achieved when $n = 1$ and $x = 1$.

Seriously though, should we minimize $y$ over $x$ as a function of $n$? Also, are $x, y$ real? (I assume $n$ is a natural number).

Ops...my problem isn't very clear, and yes, Bacterius, your instinct are all right, $n$ is a natural number and both $x$ and $y$ are real, and we should minimize $y$ over $x$ as a function of $n$ instead.:o
 
My solution:

Note that if we are to graph of $y=|x-1|+|2x-1|+|3x-1|+\cdots+|nx-1|$, we need to first determine the function for each piece from different interval, and each part in its interval is necessarily piecewise linear and hence the graph of $y$ will always looks like "elbow(\/)".

Therefore the minimum of $y=|x-1|+|2x-1|+|3x-1|+\cdots+|nx-1|$ occurs at $kx-1=0$ by the time we see the changes of the gradient of the linear function turns from negative to positive.

I will illustrate this with an example.

If we have $y=|x-1|+|2x-1|+|3x-1|+|4x-1|$ and we plot its graph, we see that we need to break the absolute function into 5 intervals:

[TABLE="class: grid, width: 820"]
[TR]
[TD]For $x>1$:

$y=x-1+2x-1+3x-1+4x-1=10x-4$[/TD]
[TD][/TD]
[/TR]
[TR]
[TD]For $\dfrac{1}{2}<x<1$:

$\begin{align*}y&=-(x-1)+2x-1+3x-1+4x-1\\&=-x+2x+3x+4x+1-1-1-1\\&=8x-2\end{align*}$[/TD]
[TD]For $\dfrac{1}{3}<x<\dfrac{1}{2}$:

$\begin{align*}y&=-(x-1)-(2x-1)+3x-1+4x-1\\&=-x-2x+3x+4x+1+1-1-1\\&=4x\end{align*}$[/TD]
[/TR]
[TR]
[TD]For $\dfrac{1}{4}<x<\dfrac{1}{3}$:

$\begin{align*}y&=-(x-1)-(2x-1)-(3x-1)+4x-1\\&=-x-2x-3x+4x+1+1+1-1\\&=-2x+2\end{align*}$[/TD]
[TD]For $x<\dfrac{1}{4}$:

$\begin{align*}y&=-(x-1)-(2x-1)-(3x-1)-(4x-1)\\&=-x-2x-3x-4x+1+1+1+1\\&=-10x+4\end{align*}$[/TD]
[/TR]
[/TABLE]
View attachment 4309From the graph, we can see that minimum of $y$ occurs at $3x-1=0$, i.e. $x=\dfrac{1}{3}$. This happens when the gradient of the piecewise linear function changes from $4$ ($y=4x$) to $-2$ ($y=-2x+2$). We can evaluate the minimum value by evaluating the function $-x-2x-3x+4x+1+1+1-1=-2x+2$ at $x=\dfrac{1}{3}$.

Now, back to the original problem to find the minimum of $y=|x-1|+|2x-1|+|3x-1|+\cdots+|nx-1|$. I know I need to focus on finding which $k$ value (from $kx-1$) does the gradient of the piecewise function changes from negative to positive. Since the only thing that matters is the gradient of the linear function, I will consider only all the coefficients of $x$ terms and discard the constants in finding for $k$. Here, $k$ is natural number.

$(1,\,2,\,3,\,\cdots,\,k-1),\,\,(k,\,k+1,\,\cdots,\,n)$

I know I need the sum from $(1,\,2,\,3,\,\cdots,\,k-1)$ is less than the sum from $(k,\,k+1,\,\cdots,\,n)$.

Therefore I get

$k\left(\dfrac{k-1}{2}\right)<(n+k)\left(\dfrac{n+1-k}{2}\right)$

I then solve the inequality above for natural number $k$, and obtain

$2k^2-2k-(n^2+n)<0$

$k=\left\lfloor\dfrac{1}{2}+\sqrt{\dfrac{1}{4}+\dfrac{n^2+n}{2}}\right\rfloor$

Now, to find for its minimum value, we subtract the sum of $kx-1+(k+1)x-1+\cdots+nx-1$ from the sum of $x-1+2x-1+\cdots+(k-1)x-1$ and remember to replace $x=\dfrac{1}{k}$ and get:

$\begin{align*}y_{\text{min}}&=(kx-1+nx-1)\left(\dfrac{n+1-k}{2}\right)-(x-1+(k-1)x-1)\left(\dfrac{k-1}{2}\right)\\&=(kx+nx-2)\left(\dfrac{n+1-k}{2}\right)-(kx-2)\left(\dfrac{k-1}{2}\right)\\&=\left(\dfrac{n}{k}-1\right)\left(\dfrac{n+1-k}{2}\right)+\left(\dfrac{k-1}{2}\right)\end{align*}$

where $k=\left\lfloor\dfrac{1}{2}+\sqrt{\dfrac{1}{4}+\dfrac{n^2+n}{2}}\right\rfloor$.
 

Attachments

  • Optimization Challenge.PNG
    Optimization Challenge.PNG
    3.8 KB · Views: 78
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. In Dirac’s Principles of Quantum Mechanics published in 1930 he introduced a “convenient notation” he referred to as a “delta function” which he treated as a continuum analog to the discrete Kronecker delta. The Kronecker delta is simply the indexed components of the identity operator in matrix algebra Source: https://www.physicsforums.com/insights/what-exactly-is-diracs-delta-function/ by...
Fermat's Last Theorem has long been one of the most famous mathematical problems, and is now one of the most famous theorems. It simply states that the equation $$ a^n+b^n=c^n $$ has no solutions with positive integers if ##n>2.## It was named after Pierre de Fermat (1607-1665). The problem itself stems from the book Arithmetica by Diophantus of Alexandria. It gained popularity because Fermat noted in his copy "Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos, et...
I'm interested to know whether the equation $$1 = 2 - \frac{1}{2 - \frac{1}{2 - \cdots}}$$ is true or not. It can be shown easily that if the continued fraction converges, it cannot converge to anything else than 1. It seems that if the continued fraction converges, the convergence is very slow. The apparent slowness of the convergence makes it difficult to estimate the presence of true convergence numerically. At the moment I don't know whether this converges or not.

Similar threads

Back
Top