Where did the other 7 equations go?

Algebra Level 5

Let x 1 , x 2 , . . . , x 9 x_1,x_2,...,x_9 be non-negative reals such that x 1 + x 2 + + x 9 = 10 x_1+x_2+\cdots+x_9=10 and x 1 + 2 x 2 + + 9 x 9 = 18. x_1+2x_2+\cdots+9x_9=18.

Find the sum of the minimum and maximum values of ( 1 0 2 9 2 ) x 1 + ( 1 0 2 8 2 ) x 2 + + ( 1 0 2 1 2 ) x 9 . \big(10^2-9^2\big)x_1+\big(10^2-8^2\big)x_2+\cdots+\big(10^2-1^2\big)x_9.


Bonus: Generalize!


The answer is 596.00.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Apr 7, 2018

Consider linear functions f , g , F f,g,F on R 9 \mathbb{R}^9 defined by f ( x ) = j = 1 9 x j g ( x ) = j = 1 9 j x j F ( x ) = 100 f ( x ) j = 1 9 ( 10 j ) 2 x j f(\mathbf{x}) \; = \; \sum_{j=1}^9 x_j \hspace{1cm} g(\mathbf{x}) \; = \; \sum_{j=1}^9 jx_j \hspace{1cm} F(\mathbf{x}) \; = \; 100f(\mathbf{x}) - \sum_{j=1}^9 (10-j)^2x_j We are interested in finding the extreme values of F F over the domain P P , where P = { x R 9 : f ( x ) = 10 , g ( x ) = 18 , x 1 , x 2 , x 3 , . . . , x 9 0 } P \; = \; \big\{ \mathbf{x} \in \mathbb{R}^9 \; : \; f(\mathbf{x}) = 10\;,\;g(\mathbf{x}) \,= \, 18\;,\;x_1,x_2,x_3,...,x_9 \ge 0 \big\} Since F F is continuous and P P is compact, F F is bounded and achieves its bounds on P P . Let the maximum and minimum values of F F over P P be M M and m m respectively.

If x P \mathbf{x} \in P then x 2 + 2 x 3 + 3 x 4 + + 8 x 9 = 18 10 = 8 x_2 + 2x_3 + 3x_4 + \cdots + 8x_9 \,=\, 18-10=8 , and so 8 + x 1 = x 1 + x 2 + 2 x 3 + + 8 x 9 f ( x ) = 10 8 + 8 x 1 = 8 x 1 + x 2 + 2 x 3 + + 8 x 9 8 f ( x ) = 80 8+x_1 \; =\; x_1 + x_2 + 2x_3 + \cdots +8x_9 \; \ge \; f(\mathbf{x}) \; = \; 10 \hspace{1cm} 8 + 8x_1 \; = \; 8x_1 + x_2 + 2x_3 + \cdots + 8x_9 \; \le \; 8f(\mathbf{x}) \; = \; 80 Thus 2 x 1 9 2 \le x_1 \le 9 .

If x P \mathbf{x} \in P and there exist 2 u < v 9 2 \le u < v \le 9 such that x u , x v > 0 x_u,x_v > 0 , consider the vector y R 9 \mathbf{y} \in \mathbb{R}^9 defined by y j = { v u j = 1 1 v j = u u 1 j = v 0 o . w . y_j \; = \; \left\{ \begin{array}{lll} v-u & \hspace{1cm} & j = 1 \\ 1 - v & & j = u \\ u - 1 & & j = v \\ 0 & & \mathrm{o.w.} \end{array} \right. Then f ( y ) = g ( y ) = 0 f(\mathbf{y}) = g(\mathbf{y}) = 0 , while F ( y ) = ( v u ) ( u 1 ) ( v 1 ) < 0 F(\mathbf{y}) = -(v-u)(u-1)(v-1) < 0 . Since x 1 , x u , x v x_1,x_u,x_v are all positive, it follows that there exists δ > 0 \delta > 0 such that x + λ y \mathbf{x} + \lambda\mathbf{y} belongs to P P for λ < δ |\lambda| < \delta . Since F ( x + λ y ) = F ( x ) + λ F ( y ) F(\mathbf{x} + \lambda\mathbf{y}) \,=\, F(\mathbf{x}) + \lambda F(\mathbf{y}) , we deduce that F ( x + λ y ) < F ( x ) F(\mathbf{x} + \lambda\mathbf{y}) < F(\mathbf{x}) for all 0 < λ < δ 0 < \lambda < \delta , while F ( x + λ y ) > F ( x ) F(\mathbf{x} + \lambda\mathbf{y}) > F(\mathbf{x}) for all δ < λ < 0 -\delta < \lambda < 0 . Thus F ( x ) F(\mathbf{x}) can be neither m m nor M M .

It is not possible for any x P \mathbf{x} \in P to have x 1 x_1 as the only nonzero component. Thus, if x P \mathbf{x} \in P is such that F ( x ) F(\mathbf{x}) is either m m or M M , then there must exist some 2 u 9 2 \le u \le 9 such that x u > 0 x_u > 0 , but that x 1 x_1 and x u x_u are the only nonzero components of x \mathbf{x} . This means that x 1 = 10 8 u 1 x u = 8 u 1 x_1 \; = \; 10 - \tfrac{8}{u-1} \hspace{2cm} x_u \; = \; \tfrac{8}{u-1} and so F ( x ) = 100 × 10 { 9 2 ( 10 8 u 1 ) + ( 10 u ) 2 8 u 1 } = 342 8 u F(\mathbf{x}) \; = \; 100 \times 10 - \left\{9^2\left(10 - \tfrac{8}{u-1}\right) + (10-u)^2 \tfrac{8}{u-1}\right\} \; = \; 342 - 8u after some simplification. Thus m m must be the smallest of these values, and M M must be the largest. Hence m = 342 8 × 9 = 270 M = 342 8 × 2 = 326 m \; = \; 342 - 8\times9 \; = \; 270 \hspace{2cm} M \; = \; 342 - 8\times2 \; = \; 326 and so the sum of m m and M M is 270 + 326 = 596 270 + 326 = \boxed{596} .

newbie question - if i understood well u showed that the maximum and minimum are received when only 2 variables are non-zero, but isnt that case outside the domain P u defined for F(x)? i mean F(x) is undefined for vectors with only 2 variables who are non-zeros. wouldnt it make so that u found the supermum and infimum for F(P) but F(P) doesnt have min and max values?

Danny Ofek - 2 years, 11 months ago

Log in to reply

The domain includes points where some of the x i x_i are zero (that is why P P is closed and bounded, so compact, and hence why the continuous function F F achieves its minimum and maximum on P P . Since F F is a polynomial function of x 1 , x 2 , . . . , x 9 x_1,x_2,...,x_9 , there are no points where F F fails to be defined and continuous. I think you are reading non-negative (i.e. 0 \ge 0 ) as positive (i.e. > 0 > 0 ). Since non-negative includes zero, there is no problem here.

Mark Hennings - 2 years, 11 months ago

Log in to reply

oh shit, i should read the question next time i guess :) thanks for responding

Danny Ofek - 2 years, 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...