USAMTS Problem 1/3/13

 

We will say that an rearrangement of the letters of a word has no fixed letters if, when the rearrangement is placed directly below the word, no column has the same letter repeated.  For instance, the blocks of letters below show that ESARET is a rearrangement with no fixed letters of TERESA, but REASTE is not.

T

E

R

E

S

A

 

T

E

R

E

S

A

E

S

A

R

E

T

 

R

E

A

S

T

E

How many distinguishable rearrangements with no fixed letters does TERESA have?  (The two E’s are considered identical.)

 

We can replace TERESA with 112345 because there are still six digits, with two being identical.

 

For a number that has no fixed digits with 112345, neither the first nor the second digit can be 1.

 

So, the first two digits can be any two of {2, 3, 4, 5}.  There are 4P2 = 4  3 = 12 such permutations.

 

We will see how many numbers with no fixed digits with 112345 there are that begin with “23”

 

 

1

1

2

3

4

5

 

2

3

1

1

4

5

 

2

3

1

1

5

4

 

2

3

1

4

1

5

 

2

3

1

4

5

1

 

2

3

1

5

1

4

 

2

3

1

5

4

1

 

2

3

4

1

1

5

 

2

3

4

1

5

1

 

2

3

4

5

1

1

 

2

3

5

1

1

4

 

2

3

5

1

4

1

 

2

3

5

4

1

1

 

There are seven numbers that begin with “23” that have no fixed digits with 112345.

 

So the total number of required permutations is 12  7 = 84.


USAMTS Problem 2/3/13

 

Without computer assistance, find five different sets of three positive integers {k, m, n} such that  k < m < n  and .

 

Let us first consider all possible sets of two integers {a, b}, such that   

 

So,

, or , or

 

 

Now, if a and b are integers, then  must also be integers. 

Thus,  are integer factors of 7056.

So, the possible pairs  are:

{2, 3258}

{21, 336}

{3, 2352}

{24, 294}

{4, 1764}

{28, 252}

{6, 1176}

{36, 196}

{7, 1008}

{42, 168}

{8, 882}

{48, 147}

{9, 784}

{49, 144}

{12, 588}

{56, 126}

{14, 504}

{63, 112}

{16, 441}

{72, 98}

{18, 392}

{84, 84}

 

By substituting these values in for , we find that the only pair that produces integer values for a and b is {49, 144}. 

{49, 144} gives us a = 7 and b = 12.

 

So, .

 

Now, we can find all integers {c, d} such that , and all integers {e, f} such that .

This will give us numbers in the form  and , which satisfy our given conditions. 

 

So,

, or , or

 

Possible values for (c  7) and (d  7) are {1, 49} and {7, 7}

 

If we don’t want c to equal d, then the only solution is  c = 8 and d = 56.

 

So, , or {8, 12, 56} is one value for {k, m, n}

 

Now we will find all integers {e, f} such that  

, or , or

 

Possible values for (e  12) and (f  12) are:

{1, 144}

{2, 72}

{3, 48}

{4, 36}

{6, 24}

{8, 18}

{9, 16}

{12, 12}

If we don’t want e to equal f, the solutions for e and f are:

e = 13

f = 156

e = 14

f = 84

e = 15

f = 60

e = 16

f = 48

e = 18

f = 36

e = 20

f = 30

e = 21

f = 28

Thus, {7, 13, 156}, {7, 14, 84}, {7, 15, 60}, {7, 16, 48}, {7, 18, 36}, {7, 20, 30},

and {7, 21, 28} are all valid values for {k, m, n}.

 

(In fact, there are many other solutions that can be found by splitting  differently.)

 


USAMTS Problem 3/3/13

 

Suppose p(x) = xn + an-1xn+1 + … + a1x + a0 is a monic polynomial with integer coefficients.  [Here monic polynomial just means the coefficient of xn is one.]  If (p(x))2 is a polynomial all of whose coefficients are non-negative, is it necessarily true that all the coefficients of p(x) must be non-negative?  Justify your answer.

 

If we take

p(x) = x5 + 100x4 + 100x3  x2 + 100x + 100

we get

 

Here, p(x) is a monic polynomial with the coefficient of x2 being negative.  All the coefficients of (p(x))2 are non-negative.  Thus, it is not necessary for all the coefficients of p(x) to be non-negative for all the coefficients of (p(x))2 to be non-negative.

 


USAMTS Problem 4/3/13

 

As shown in the figure on the right, in ΔACF, B is the midpoint of, D and E divide side  into three equal parts, while G, H, and I divide side  into four equal parts.  Seventeen segments are drawn to connect these six points to one another and to the opposite vertices of the triangle.  Determine the points interior to ΔACF at which three or more of these line segments intersect one another.  To make grading easier, we have embedded the triangle into the first quadrant with point F at the origin, point C at

 (30s, 30t), and point A at (60r, 0), where r, s, and t are arbitrary positive real numbers.  Please use this notation in your solutions.

 

.

Using one specific triangle (see Fig. 1), I drew, experimentally, the seventeen segments.  I found that there were only two likely cases of more than two lines intersecting at a single point.  These are the intersection of,  and , and the intersection of , , and .  We will  prove that these triplets of lines do indeed intersect at single points. (See proofs 1a and 1b) .

 

We will also prove that:

 

 


Using the given coordinate system, the coordinates of points A  F are:

A(60r, 0), B(15s + 30r, 15t), C(30s, 30t), D(20s, 20t), E(10s, 10t), F(0, 0), G(15r, 0), H(30r, 0), and I(45r, 0)

 

Proof 1a:

We will show that , , and  meet at one point. 

 

The equation of  is

 = 20tx   1200rt   …. (Eq 1)

 

The equation of  is

 = 15tx  450rt    ……………(Eq 2)

 

The equation of  is

 = 30tx  1350rt  ……  (Eq 3)

 

By multiplying Eq 1 by 3 and subtracting  Eq 2 multiplied by 4 we get for the point of intersection for lines  and :

 

y(60s  180r  60s) = -1800rt

or y = 10t

 

Similarly, by subtracting Eq 3 from 2 times Eq 2, we get for the point of intersection of lines  and  :

y(45r) = 450rt

or y = 10t

 

Since both points have same y value, we need not check the x value because none of the above lines are horizontal.

 

Thus the three lines intersect at one point

 

Substituting the value in Eq (1) we get x = 30r+10s and so the common point is,

(30r+10s, 10t)

 


Proof 1b:

We will show that, , , and  meet at one point. 

 

The equation of  is

, or

  (s + 2r)y = tx    ………… (Eq 4)

 

The equation of  is

y(30s  15r) = 30t(x  15r) = 30tx  450tr, or

  (2s  r)y =  2tx - 30tr  ………(Eq 5)

 

The equation of  is

y(10s  30r) = 10t(x  30r) = 10tx  300tr, or

  (s  3r)y = tx - 30tr    ………(Eq 6)

 

By subtracting Eq 4 from Eq 6, we get, for the point of intersection of lines  and :

-y(s  3r) + y(s + 2r) = 30tr, or

y = 6t

 

By subtracting two times Eq 4 from Eq 5, we get, for the point of intersection of lines  and :

-y(2s  r) + y(2s + 4r) = 30tr, or

y = 6t

 

Again, since both points have same y value, we need not check the x value because none of the above lines are horizontal.

 

Thus the three lines intersect at one point

 

Substituting the value of y in Eq (4) we get x = 12r + 6s,

So the common point is at 

(12r+6s, 6t)

 

 

 

 

 

 


Proof 2:

 

Let  and  be two intersecting lines in  ΔACF.  Let L and N be the points on  where these lines meet at the base. (We will later discuss the case  where the end points are not on the line AF)   Let the point of intersection be P

 

Let us take the case where both K and M are on the line CF.  (See Fig 2).

 

According to the problem, the values of  and  are  given constants  (they do not depend on r, s, or t)

 

Now, let us draw a line parallel to  through C, and a line parallel to  through C, meeting  at points L' and N', respectively.

 

So,

ΔFKL ~ ΔFCL'

 

So, the values of  and  are constants and hence  is also a constant.  

 

 

Now

ΔPNL ~ ΔCN'L'

 

Thus, the height of  ΔPNL is proportional to the height of  ΔCN'L' .  In other words, the height (or y-coordinate) of the point P depends only on the height of the original triangle, and not the shape.

 

Using similar logic, it is easy to see that this will hold true even if the point K is on  the line , as long as the ratio of AK to AC is a given constant.  Similarly, if  the two end points are on lines  AC and CF, one simply extends the line so that it meets the bases .  (In the given problem there are no horizontal lines, so the line will meet AF)  . In all  the cases , the horizontal segments of the small and large triangles would be proportional to the base of the triangle and the height would be proportional to  the height of the original triangle.) See following figures

 

 

 

 

 

 

 

 

 

 

                                                    F

 

 

Here AL’ is AL *(AC/KL) hence it is proportional to  AF.

CS= (CF/RC)*CK ; SA= CA-CS ;  and LA = (KA/SA)*FA and thus  LA/FA is a constant (depends only on the given ratios)

 


This method also suggest an easy way to find the heights of the points of intersection, and thus one can calculate if more than two lines meet at a point.  Each line now can be represented by two parameters. First lets us take, t=1, r=1, and s=0. The two parameters are:  x- intercept of the line (We will call it p of the line) and the x-intercept of the line parallel to the line and passing through C,  (We will call it q). Then the height of the point of intersection is :

             (difference in p values)/(different in q values)  * (Height of the triangle).

 

Part 3 -

Let us write for all these lines:

 

Line

p value

q value

p/15

q/15

AD

60

90

4

6

AE

90

180

6

12

BD

120

180

8

12

BE

-60

-180

-4

-12

BF

0

-60

0

-4

BG

15

-30

1

-2

BH

30

0

2

0

BI

45

30

3

2

CG

15

15

1

1

CH

30

30

2

2

CI

45

45

3

3

DG

15

45/2

1

3/2

DH

30

45

2

3

DI

45

135/2

3

9/2

EG

15

45

1

3

EH

30

90

2

6

EI

45

135

3

9

 

 

Note that we have added additional columns  as (p/15 and q/15) ,as , we are only concerned with the ratio, and these values will make it easier to calculate:

 

Using the above values, we find that   the “y” values of the points of intersection ( and multiplying the ratio by 30 to get the answer in the same form as before) we get:


The following table shows the y coordinates of all points of intersection:

AD

BE

40/3

AE

CH

6

BF

CG

6

BG

DI

120/13

CH

DI

12

AD

BF

12

AE

CI

10/3

BF

CH

10

BG

EH

15/4

CH

EI

30/7

AD

BG

45/4

AE

DG

60/7

BF

CI

90/7

BG

EI

60/11

 

 

 

AD

BH

10

AE

DH

20/3

BF

DG

60/11

 

 

 

DG

EH

20/3

AD

CG

18

AE

DI

4

BF

DH

60/7

BH

CI

10

DG

EI

8

AD

CH

15

 

 

 

BF

DI

180/17

BH

DI

20/3

 

 

 

AD

CI

10

BE

CG

150/13

BF

EH

6

BH

EI

10/3

DH

EI

5

 

 

 

BE

CH

90/7

BF

EI

90/13

 

 

 

 

AE

BF

15/2

BE

CI

14

 

 

 

CG

DH

15

AE

BG

45/7

BE

DG

100/9

BG

CH

15/2

CG

DI

120/7

AE

BH

5

BE

DH

12

BG

CI

12

CG

EH

6

AE

CG

90/11

BE

DI

140/11

BG

DH

6

CG

EI

15/2

 (Lines that obviously do not intersect inside the triangle have been ignored.)

 

 

It is easy to see that the only triplets of intersecting lines that have the same y coordinate are ((AD, BH), (AD, CI), (BH, CI); y = 10), and ((BF, CG), (BF, EH), (CG, EH); y = 6)

 

The only points where three or more lines intersect in a way described in the problem are:

 

(30r + 10s, 10t) and (12r + 6s, 6t)

 

 

 

 

 

 

 

 

 

 


USAMTS Problem 5/3/13

Two perpendicular planes intersect a sphere in tow circles.  These circles intersect in two points, spaced 14 units apart, measured along the straight line connecting them.  If the radii of the circles are 18 and 25 units, what is the radius of the sphere?

 

 

 

 

Let O be the center of the sphere, and P and Q be the centers of the circles with radii 18 and 25, respectively.  Let the circles meet at points A and B, and let C be the midpoint of .

 

So,

 

.

 

So,

 

Also, OQC = QCP = CPO = PCA = 90

 

So,

 

 

Now, from right ΔOQA, we get

 

 

Thus, the length of radius  is 30.