NON-ROUTINE PROBLEMS IN MATHEMATICS (WITH SOLUTIONS) Editor Prof. V. K. Krishnan The Association of Mathematics Teachers of India B-19, Vijay Avenue, 85/37, V. R. Pillai Street, Triplicane, Chennai — 600 005. E.Mail : [email protected] Phone : (044) 844 1523.BETWEEN US Dear Reader, It is with a sense of pride and satisfaction that we place in your hands the edited version of the lessons dispatched as Correspondence Course in Non-Routine Problems Solving, The idea was conceived and executed by our senior member Prof M.S.Rangachary, formerly Director of Ramanujan Institute for Advanced Study in Mathematics, University of Madras and ably assisted by Prof. G. Rangan, Prof. ¥: Shankaram, Prof. VK. Krishnan, Dr. S. Muralidharan, Dr. Hemalatha Thyagarajan, Sri S.R. Santhanam and Ms. R. Vijayalakshmi. They were sent as 11 lessons to the registered candidates. Though their number was few, the quality of the work was appreciated by the recipients and those who saw the lessons. Itwas felt that the same might be brought out in the form of.a book for reference by the prospective Olympiad participants, IIT type entrance examination aspirants and the general public to spend time with Math- ‘ematical thinking as a hobby. ‘Prof. V. K. Krishnan from Trichur was kind enough to edit the les- sons to suit a textbook with improvements in the solutions of worked problems and in suggested solutions. I take this opportunity to express my grateful thanks to Prof: M. S. Rangachary, Prof. ¥.K. Krishnan and other learned teachers who made this publication possible. We are quite sure that the well-wishers of AMTI will ensure that this material reaches the needy and generate de- mand for further editions too. {tis my duty and responsibility to place on record our thanks to SriG. Narayanan of Ramanujan Institute for technical assistance and M/s Pagesmith Laser Typesetters, M.K, Graphics and Souri Printers for bring- ing out this book in time. With Kind regards/ best wishes, Yours sincerely, Ae 25.12.2001 (M. Mahadevan} A SecretaryLesson 2 Lesson -4 Lesson 5 Lesson 8 Lesson 9 Lesson 10 Lesson 11 CONTENTS Pigeon Hole Principle-I ..........+-++2-serseerereeeet 1 Tnequallities ..........0.cececeencesetareeeessecssoene 5 Divisibility of Numbers ..........--++++++ aseeieseess % Permutations and Combinations ..........--+-++++++ 32 Number Theory ........0605c0scccseeeeeeeeeeeeeeees 45 Concurrency and Collinearity ............ aavessoeaee 53 Mathematical Induction .............0eceeeeeseeeeie 73 Pigeon Hole Principle-I Quadrilaterals: Cyclic and Circumscribed ........... 87 Circles Related to Straight Lines Triangles ........ 105 Circles Connected with a Triangle ........s.sssc005 ur Solutions of Problems ..............--+0-:ssseeeeeeee 129LESSON 1 PIGEON HOLE PRINCIPLE-I Welcome to this course! Hope you will realize through this course the power of mathematical thinking. This programme will suceed if your confidence level goes up. In this lesson we start with a simple idea. Supposing you have 4 purses and 5 coins and you put the coins in these purses as you like. For instance, you can put all of them in one purse so that 3 purses are empty or you can put 2 in one, 1 in another, 2 in another and keep one empty. But some occurrence is certain to happen. Can you guess? At least one of the pu:ses contains more than one coin. Once you have been told this fact the ‘idea’ is silly and just an offshoot of commonsense (thinking!) You might have seen holes under the roof of towers in temples, churches, mosques etc. where pigeons live. If there are n holes and there are (n +1) pigeons and all of them get into the holes, then. At least in one hole there will have to be more than.one pigeon Bearing this in mind, the idea is called the PIGEON HOLE PRINCIPLE (PHP). See the power of this principle illustrated in the following examples. Example 1: Show that given 12 integers there-exists two of” them whase difference is divisible by 11. Solution: Possible remainders when any integer is divided by 11 are 0,1,2,--+ ,10.- Treat these remainders as ‘holes’ and the2 NON-ROUTINE PROBLEMS IN MATHEMATICS 12 given integers as pigeons. By PHP two of them should lie in the samie hole, i.e. should leave the same remainder when divided by 11. Hence their difference is divisible by 11. Example 2: There are balls of 4 colours in a basket. When picking up a random collection of balls, if at least 2 balls should have the same colour, what is the least number of balls in the collection? Solution: Treat the 4 colours as ‘holes’. The minimum number 5 of balls treated as ‘pigeons’ ensures that, at least 2 balls are of the same colour because of PHP. Thus the answer is 5. The following illustrates PHP being used along with some other idea. Example 3: There are 7 persons in a group. Show that some two of them have the same nuinber of acquaintances among them. Solution: Each of them can be acquainted with 0,1, 2,3,4;5,6 persons in the group. If there is one of the seven persons ac- quainted with the other 6, then there is no person acquainted with none of them, i.e., if we denote the corresponding classes of persons with an overhead bar, if 6 is not empty, then 0 is empty. In this case the seven persons are to be put in the classes T, 2, 3,4,3, 6 (6 classes) so that one of them should have more than one person by PHP. If 6 is empty,then the 7 persons are to be put in the classes 0,1, 2,3,4,5 so that in this case too there are two persons in some one of these classes. This proves the assertion. Example 4: Five points are marked at random in a square plate of length 2 units. Show that a pair of them are apart by not more than V2 units. Solution: By taking the midpoints of the edges and joPigeon HOLE PRINCIPLE-[ 3 the points on opposite edges we get a grid of 4 unit squares. By PHP two of the points lie in one of these grids and the maximum distance between these points is V2 units (the length of the diagonal of the unit square) Note: Here the pigeon holes are not pairwise disjoint. PROBLEM SHEET . The city of Madras has a population of 25 fakhs. If each citizen of Madras has an asset. of value not more than 2 lakhs, show that two of the citizens have assets of the same value, when corrected to the nearest integer. » Given 10 triangles show that two of them are either equi- lateral or isosceles but not equilateral or acute angled but not isosceles or obtuse angled but not isosceles or right angled but not isosceles. 3. A is a subset of the arithmetic progression 2,7, 12,--- ,152 having 16 elements. Show that there are two distinct, elements of A whose sum is 159. What can you say if A has 14 elements? = . Given three points in the interior of a right angled trian- gle, show that two of them are at a distance not greater than the maximum of the lengths of the sides containing the right angle. o . If a line is coloured with 11 colours show that there exist two points whose distance apart is an integer which have the same colour. = }. Show that in any set of ten distinct two digit numbers there exists two subsets which have the same sum.NON-ROUTINE PROBLEMS IN MATHEMATICS - What is the maximum number of squares in a 4 x 4 check- board which could be coloured red so that there is no red right angle in the board? . Show that, if in a class of 15 students the total of the marks in a subject is 600, then there is a group of 3 stu- dents the total of whose marks is at least 120. . Show that there exists a power of 3 which ends with the digits 001. If the. digits 1,2,3,4,5,6,7,8,9 are divided into three groups show that the product of the numbers in one of the groups exceeds 71.LESSON 2 INEQUALITIES § 1. We know thai given:any two distinct real numbers a,b, ither @ b. But to decide whether it is the former quality or the latter for any given pair of real numbers often mires use of certain facts in a clever way. Take for example 2,33. If 0 < m 0, 2" < 3”. But these inequalities as such do not help to compare 2° and 38. On the other hand, 3° = (2418 = Be3xP2¢3x2t+1< Prax P+ax 2= P4428 = 2x B42! = 2x 24 = 2. OF course 25 = 32 and 3% = 27 are easily computable and it is evident that 2° > 33. If, however, the indices m,n of 2”,3” are very s from which we lary is necessary to think of smaller indi could derive the inequality for larger indices. e.g. knowing that 2 > 33, we immediately deduce 25%12! = 2605 > 33121 — 3363, Also. 27 > 34 since 2? x 25 > 2 x 39 = 4x3 >3x 3% = 3 Hence, eg. 27 > 34. A simpler deduction is the foutine 9 = 3? > 23 = 8 = 37 > 29. Example 1. Which number is greater 31!? or 17!7? Solution. 31}? < (32)!? = (25)!2 = 260 < 268 = 24x17 = (2*)'7 < 167 < 17!7. So, 31! < 17'7, Example 2. Which is greater 7°? or 8°? Solution. Note that if a, b are positive and n a natural number (a+b)" = (a+b) x-!-x (a+b) = a"+na"~'b+ terms involving powers of a and / or b with natural numbers as coefficients. So (a +0)" > a" + na"~'d (equality occurring only for n = 1; this is an easy deduction from the binomial theorem and is some6 NON-ROUTINE PROBLEMS IN MATHEMATICS times called the Bernoulli inequality). Now 8° = (7 +1)% > 7) 491 x 7% = 7°9(7 +91) = 7%. 98 > 7.49 => 797, Hence 8°! > 7%. (Exercise: Use some other way to get this inequality). Example 3. Show that 49 < 2!33 + 3133 < 4108, Solution. Since 2.3!33 > 2!33 + 3!%3, to prove the inequality on the right. it suffices to show that 4198 > 2.3158, ie. 495 > 2. , 7, a Note that 4! > 3° so that fy > ($2) |= (1+28)” > 1+ 27x 33 > 2 (by Bernoulli’ the right is thus proved. We prove that 3'°3 > 49°-more than inequality). The inequality on s\ 19 95 the inequality on the left. Now 4° < 37. S0(#) = das <1 proving that 3'%3 > 4% and so the inequality on the left. Note: If0 <2 <1, then x? 1, then 2? > x. On the other hand, if 0 < z < y, then 2* < y® whatever be a>d. v2+Vv3 V24+ V3 i B22 (i Solution. Ss > Ga (since V3> V2 and V3<2)= we = v2, proving the inequality on the left. Now, Ari < 2¥3 = 9 (since VE re 2 (since ¥2 < V3 and V3 > 1). Observation. Consider the numbers ygpr jwgvt. Note that the pattern of the numbers in both the numerator and denom- inator of each fraction is the same, viz. Os flanked by 1. The only difference is that the denominator has one more 0 than Example 4. Show that V2-< <2. the numerator. If x is the numerator of any one of the two fractions, then the corresponding denominator is 102 — 9. De- noting the fraction by a, } = 12-2 = 10 - 2. Thus if z in- iB a = creases, 2 decreases and so ~$ increases. Hence } increases when z increases. Or, a decreases when = increases. In other words, A%; > yquour: The same argument applies generally asINEQUALITIES 7 indicated in Example 5. Show that k reros a aw (k+1) zeros (41) zeros Example 6. Show that (1.01)! > 1000. Solution. By Bernoulli's inequality (1 + .01)* > 1 +.08 and so (1.01)'00° > (1 + .08)'5 = ((1 + .08)5)8 > (1 + .4)?5 > (1 + .4)?4 = ((1.4)3)8 > (2.7)8 > 74 = 2401 > 1000. 12366665 _ 12366666 25633321 ~ 25633322" Solution. If the left hand number is £, then the right one is 241. From the difference Example 7. Show that if and only:if z > y, it follows that the right hand member is greater than the left hand member as desired. Example 8. Show that 1-}+4 Solution. }-}+}-f4--h 45 - eee i 4 1 1 1 (-3)+(-§) +--+) +5 1 é 1 6 +o 44 98x99” 100° Bod 07 0075 tote ne 20 ao > at % 1 $,1-(b-f4 poe as + 105) <2 Example 9. Which is greater 150° or 20000! x 19910028 NON-ROUTINE PROBLEMS IN MATHEMATICS Solution. 1503.< 150 x 150 x 150 = 30 x 30 x 30 x 125 = 27000 x 125 > 20000 x 100. Hence 150° = (150%)! > 20000! x 10019 Example 10. Find the largest of the numbers: 5100, G91 790, gas Solution. Write the powers of 5,7,8 as follows: Powerof 1 2 3 4 5 6 ul 5 5 25 125 625 3125 15625 78125 i 7 49 343 2401 16807 117649 823543 8 8 64 S12 4096 32768 262144 2097152 Thus 8¢ = 4096 > 3125 = 5° so that 88 > 8® = (81) > (55)? = 5! Next, take 2° x 39! = 6%" and 88 = 255, Canceling out the common factor, we need to compare 3°! and 264, To do this, note that 2° = 256 > 3° = 243 so that 264s 9182 — (28)19 > (35)!9 = 39 > 3°! Thus, 8® > 6%. It remains to compare 7” and 8°°, From the row relating to pow- ers of 8(= 23), 2"7 = 8 x 4 = 32768 x 4 = 131072 > 117649 = 79. Hence 88° = 2°59 = (2!7)!5 > (75)!5 = 7° proving that 8° > 7°. So, 88 is the largest of the four numbers. § 2. In the previous section we considered inequalities relat- ing to given-real numbers and observed that the inequalities can be obtained using some general inequalities like the Bernoulli inequality. We now consider general inequalities. Note that the variables which occur can take only real values not non- real complex values. Most of these general inequalities revolve around the fact that for any real number x, 2? > 0. Observe that this is not the case for 2” unless n is even. Also if we consider a rationalINEQUALITIES 9 number p/q, p € 2, the set of all integers, g € N, the set of all natural numbers, then 2?/* may not be defined for some as a real number. e.g. (—5/2)°/? is not defined. On the other hand, it is known that if n is any natural number and y is a non-negative real number, there exists a non-negative real number z such that 2” = y. 2 is deuoted in this case, by y'/* or 4g. The use of radical y/ and, in particular, 7" (for the square root), is‘conventionally limited to this non-negative real number z. Arithmetic Mean (A.M.) of a set of n positive real numbers, @1,42,-+- ,@p is defined to be ay +02 + n +an i.e. it is the average value of the n numbers. Geometric Mean (G.M.) of the same set of numbers is de- fined to be Yay X G2 X-** X Gn. If a,b > 0, a, then *$*,b are such that the difference between consecutive terms, viz. f+ — a, b- *4 are each 7 ile. the same. This is otherwise described by telling that the three terms form an arithmetic progression (A.P.). 4% is thus an AM. inserted between a and b to get an A.P. Again, a, Va x 6, ab an aby are each vi ile. the same a, Va-6, is said to form a geometric progression (G.P.) and Vab has been inserted (as a G.M.) between a and b to get a G.P. AM. - G.M. inequality: vas*>? 0) b are such that the quotient of consecutive terms, viz. Proof. (1) is equivalent to 2Vab < a + 6 (multiplying an10 NON-ROUTINE PROBLEMS IN MATHEMATICS inequality by a positive real number does not alter the inequal- ity), ie. 0 < a+b —2Vab (subtracting the same real number * from both sides of an inequality does not alter the inequality), ie. a+b—2Vab = (Va— Vb)? > 0 (Note that Vab, /a, Vb are uniquely defined.) which is true from our remark at the very beginning of § 2. Note. The proof shows that Vab = “#2 if and only if a = Vb or, equivalently. a = b (Recall the remark on the uniqueness of the nonnegative root.). To sum up Vab < “#4, ie. G.M. < A.M. and equality occurs if and only if both the quantities are equal. The G.M. - A.M. inequality is true in the general form Yar an < ate tan n equality occurring if and only if a = a) = Op, It is easy to deduce it for a’set of 4 positive numbers from the case of 2 numbers proved earlier. For, Yajazaza, = Yajaq- Yazaq Vara - Jase 1 3 (Vaia2 + Ya3a4) ; (Je +42) + Flas + 2) a) +42 +43 + a4 4 7 A IA Note. The above method of proof can be extended to.the case n= 2* for some k = 2,3,4,--+. To prove a +2 +3 Ya10203 < 3 (2)INEQUALITIES sae for a; >0, §=1,2,3, set a1 ==}, a2 = 23, a3= 23, 2; >0, i = 1,2,3. (2) becomes 3azars Sh +2} +23 or x} +2} + 23 - 32,2273 > 0. The I-h.s. of (3) can be rewritten as (x) + 22 + 23)(2? + 23 + 23 — 2109 — 2223 — 2321) 2 2 2 rh — + (x2 — 23)? + (23 - 21) (ert +m) (cele t | 3)? + (a3 — 21)"} 20 since the sum of the three squares is always nonnegative. Note also that equality can occur if and only if 2; = 72 = 73, or a, = a2 = a3. Note. The proof for the general case of n is by finite mathe- matical induction, a topic to be taken up in another lesson. Example 11. If a,b > 0 and a + 6 = 1, show that 1 1 8+ar [8+ gy 24. Solution. By A.M. — G.M. inequality 1 1 z 1 frat ys+p 28+ ay 8+ a 1 1 1 in af u+0(S+5) +an} 16 1 2y Ott ot ae 2V64 +16 x 4+ 16 2v144 = 2x12 24 Vv IV Vv12 NON-ROUTINE PROBLEMS IN MATHEMATICS since *$¢ = } implies } > Vab or Fez > 2. Example 12. For any positive a,b, show that 2 1)? (o+2) +(o+5) 28 a b Solution. By A.M. — G.M. inequality (3) otV ef) OY a) = bye 2) > 2) =8. 2(oos + $42) 2204 ) Vv Note. For x > 0, from A.M. — G.M. inequality it follows that rti>d2 Example 13. For any real x,y, 2, show that wey tz? > ay tyzt 20. Solution. By A.M. — G.M. inequality x? + y? > 2zy, y+2? > 2yz, z? +2? > 2y, irrespective of whether z,y,z are positive or negative. Adding these inequalities and canceling the factor 2 on both sides we get the result. Example 14. — If aj,a2,---,a, are nonnegative and 014) --+ Gy = 1 show that (1 +a1)(1 +@2)-+-(1 +n) > 2". Solution. By A.M. — G.M. inequality 1+ > i 7 7 2 Va, §= 1,2, Multiplying the inequalities (1+ a1)++-(L +n) > 2" Yayag= ay = 2",INEQUALITIES 13 Example 15. If ai,a2,a3 > 0, show that 1 1 1 9 a a2 a3 ~ a +a2+03° Solution. By A.M. — G.M. inequality a) ee ceca a a2 a3 Yera203 ; 1 3 By the same inequality T= 2 pate So, ale 9 + = +2 ay" a2 a3 > a +42 +43 Example 16. If a,b,c > 0, show that a6 c —+— 2 3/2. bret cra athe /2 Solution. sie RE da HE oe = HH so that, 1 1 L ins. =(orbro{ t+ teeth as, But 1 1 1 fea T T garish own ores —— + ——_ 4 — bret erat ate 2 Vote * cra tate Aatb+e) = (b+c)+(c+a) + (a+b) > 3Wb+etetatatd ‘80, IV Iv14 NON-ROUTINE PROBLEMS IN MATHEMATICS Note. A geometric interpretation of the A.M. - G.M. inequality. & Take segments of length a,b on a straight line putting them side by side along AB A € and BC. Take the midpoint D of AC. Erect a perpendicular to AC at B to cut the semicir- e cle on AC at BE. If E’ is the other point of intersection of ED produced with the complete circle, AC being a diameter perpendicular to the chord EE" bisects EE’. By the “secant theorem” AB - BC = a-b= E'B. EB = EB. Or, EB = Vab, ie., EB represents the G.M. of @ and 6 the A.M. being AD = radius of the circle. Since in any circle half of a chord is less than the radius (the diameter is the longest chord) Vab <= ‘#4. It is also clear that equality occurs if and only if B and D coincide, ie. a = b, §3. Another interesting inequality is Cauchy's inequality or Cauchy-Schwarz inequality as some people call it. If a, bzy are real, then lax + by| < Va? + 62/2? +92, |-| denotes the absolute value of the real number inside it, Proof. To prove (az + by)? < (a? + 6)(2? + y2) (4) a?z? + 2azby + by? < a2? + ay? + 62a? + by? 2azby < ay? + 622? ie. ay? — 2azby + 6%y? >0 ie. (ay — br)? >0 which is true.INEQUALITIES 15 ‘The same method of proof shows that if aj,a2,+-+ ,anj b1,b2,-++ ybn are all real, then layb + age +++ sAndal S at +43 +--+ ah y/OF + OF + ++ BR. Note. If aj,bj,i 1,2,-++ ,2 are all nonnegative. there is no need to put |-| on the Lhs. e@y) axtbY =O Note. The inequality (3) also has a simple geometric inter pretation. Let P(z,y) be a point in the plane referred to ~ a system of rectangular axes with origin O and let az + by = 0 be a given straight lax + by line through origin. Then is the perpendicular dis- d tance (PM) of (x,y) from the given straight line az + by = 0. Vr" + y? is the distance of x,y from origin. Cleatly, the former is less, than or equal to the latter whatever be a,b. The slope of the line is —¢ and of the line joining origin (0,0) to (2,y) is ¥. If these two lines are perpendicular and only then ¥(—§) = -1 or $ = 4 and in this case and only in this case does equality occur. viz. PO = PM i.e. equality occurs in (3) if and only if ¥. This is also clear from the proof given éarlier. Example 17. Find the minimum value of V2? +y? when 32 + dy = 15. Solution. [32 + 4y| < J+ PVFTR = sty by Cauchy's inequality. So, under the given condition v2? +? > 3 =3 and this value is reached (why?).16 NON-ROUTINE PROBLEMS IN MATHEMATICS Example 18. If a; > 0, i =1,2,-+* .7, show that ay tapt- tan < nl +a;). Solution. ‘Take-6; = 1, i = 1,2,-+-, and apply Cauchy's inequality. Example 19. Prove that if x; > 0, i= 1,2,---.n. 1 1 2 Ff Boy =))> 2. (a+ +2) (2 + +2) 2n Solution. Apply Cauchy’s inequality choosing a; = 1;, 6; = +. lsisn. Example 20. Prove that for any natural number n > 1 2 ’ (b+ qn tagty ttt) S(ht ett a)’. Solution. Choose a; = } and b; = gh, i= 1,2,--,n and apply Cauchy's inequality. § 4. Ifa,b are any real numbers and'|-| stands for absolute value, we know that Ja + | < Jal + fa}. (4) Recall that for any real number a, Ja| > 0 and |al = 0 if and only if a= 0. —a < |a| 2 for all z. Solution. If a = 2-5,b=3-x la +6] =|e-54+3—2]=2< fe—5] +|3—zI. Example 22. Show that x? — |r| +1 > 0 for all z. Solution. If |z| < 1, -|z]+1 = 1-2] > 0. Hence x? — |x] +1 > 0. If |x| > 1, 2? = |x|? > |x| so that x? - |x| > 0,18 NON-ROUTINE PROBLEMS IN MATHEMATICS Example 23. Show that. ({:r| + yl)? — 2lz + yl +22 0 for all ay. Solution. |z| + |y| > |z + yl. so that (lal + ly)? —2(le+ yl) +1 > (lel + lull? — 2lel + byl) +2 ((2| = 1)? + (Iyl = 1)? > 0. Note. a +] = |a| + |b] if and only'if ab > 0. For. in such a case ja +b)? = (a +0)? = a? + 2ab +0? = (lal + [bl)? = {o2] + 2jal [o| + [bl2. Since a? = Jaf’, 62 = [o/%, it follows that equality occurs if and only if 2la| |b] = 2ab or |ab| = ab or ab > 0. Example 24. For what values of x does fe — 10] + |8 — 2] = 22 Solution. 2 = |x — 10 +8 — | so that equality occurs when (2 - 10)(8 — x) > 0, ie, when both x — 10 and 8-2 >0or both x — 10 and 8 - 2 > 0. ic. either 8 > x > 10 (impossible) or 8 << 10. The values of x desired are those in the closed interval [8, 10]. § 5. We consider some other inequalities in this section. Example 25. Prove that, ifa >> 0, 1e=t? < a4? ~Vab (= AM. - GM) < 1(¢=9? zs Vab (= A.M. - G.M.) 1; . ath fa—VJo)? (a—t + Solution. *f4 — Vab = WE5VIP < 16-9 if we prove that (Va + Vb)? > 4b, ne. (\V/F+1) > 4, which is the case § 2 1. The first inequality is similarly proved.INEQUALITIES 19 Recall that if a1,@2,--- ,@n are nonnegative quantities in A.P. (see §2), then a +ag +--+ an Hence, the A.M. > G.M. inequality takes the form ata, 5 Yaran-@n <—>— J (8) if'a,,+++ jan are in A-P. Also, if d is the common difference, viz. d = a2 — ) = G3 — 2 = ++- = Og — OR-y = +8) = An — An-1, then ay = a) + (k — 1)d, 1 4)@n, k = 1,2,--- ,n. From these inequali- ties (a1a2-+-an)? = (a14n)(a2@n-1)--- (any) 2 (aan). Hence Vaan < V/aja2--- On (9) Combining (8) and (9) we have for an A.P. a1,a2,--- ,an of Positive terms: G.M. of first and last term = /ajGn < Yaja7- Gy = G < agen A.M. of the first and last term (10) Example 26. Prove that for any natural number n vas Vals 2h20 NON-ROUTINE PROBLEMS IN MATHEMATICS where n! (n factorial) = 1-2-3+-+n. Solution. Put a; = 1,a2 = 2,--- ,an =n in (10). For «« > 0. we have the identity l+ne 1+(n-1)t +(n—Dz 1+(n-2)e (a) lt+nr= for any natural number n. Note that, for = 0,1,2,---,2-1, z__lt(ntiz T+nz lt+nz * Hence, from (11), EGE DEY" ie. tna)! > (14 (n+ 10)" l+ur> l+nzr In particular putting z= x25, a >0 a \nl ae (+545) >(+8) and when a = 1 eee Inequalities proved using other techniques and relating to geometry will be taken up in other lessons. PROBLEM SHEET 1. Prove that <$-3-3-- Mc, 2. P. 1 1 sae 1 g rove that Th + Sig + + Fev > 3 3. If u is a natural number, show that 1 1 2(vn+1-1) <1+—24+—}4.. ( Isla aSon 10. INEQUALITIES 21 . If a,b,¢ > 0, then show that. ab + be + ca > aVbe + bea + eVab. . For any real numbers a,b,c shaw that al +b! +c! > abela+btc). If x > 0, show that. 3x3 — 6x? + 4 > 0. . Prove that for any real x x(a + 1)(2 +2)( +8) > -2. . If zy € [0,1]. show that z y . If z,y > 0, find the maximum of the minimum of the values of x, be y+i. Prove that, if a,b,c,d > 0, then Vat o)(b +d) > Vab + Ved.LESSON 3 DIVISIBILITY OF NUMBERS In this lesson we take up the idea of divisibility in the nat- ural number system. For example 6 = 3.2. 3 and 2 are called divisors or factors of 6. Similarly 2,4,8,16 and 1 are divisors of 16. In general if a = be where a,b,c are natural numbers then 6 and.c are called divisors or factors of the natural number a. If and m are two given natural numbers, the largest or the greatest natural number d among:the common divisors-of | and % is called the greatest common divisor of | and m written in short as g.c.d. Note that g.c.d.(I,m) is at the least 2. Also, that g.c.d.(a,0) = a by definition, though 0 is not a natural number. The smallest.natural number k such that both / and m are divisors of k is called the JeAst common multiple of | and ‘m, written in short as lcm. Similarly, we can talk of g.c.d and |.e.m of 3,4,5 etc., or more generally, a finite set of natural numbers. Ifa and b are two integers such that a—b is divisible by a natural number d or equivalently a and b leave the same remainder when divided by d, we write this as a = b (mod d). a = 0 (mod d) means that d is a divisor of a, which we also write as dla. Recall that if a natural numiber p is such that 1 and p are the only divisors of p, then p is called a prime number. If the “natural numbers a and.b have d as a g.c.d. of a and b, then we write this as (a,b) = @. If (a,b) = 1, then we see that a and b have no common divisors other‘than 1. ‘There are certain simple facts that one has to remember in testing the divisibility criteria for a number by a given number. It is obvious that.any even number is divisible by 2. Any suchDivisipiLity OF NUMBERS 23 number should end up with 0, 2, 4, 6, 8 in its unit's place, A number is divisible by 3 if and only if the sum of the digits of that number is divisible by 3. Similarly a number is divisible by 9 if and only if the sum of the digits of that number is divisible by 9 (Prove these two statements.). A number is divisible by 4 if and only if the two digit num- ber formed by the last two digits of the given number (i.e. digit in the 10th place and the digit in the unit place taken together) is divisible by 4. Similarly you can think of divisibility test for 5,6 etc. It is clear that if d divides a (notation: dja) and 6, it di- vides, a + 6. In our notation a = 0 (mod d), 6 = 0 (mod d) imply together a + 6 = 0 (mod d). More generally, a = ¢ (mod m), 6 = d (mod m) together imply a+ b= c+d (mod). eg. a = c (mod m) > a—c = 0 (mod m), b = d (mod m) = 6d =0 (mod m) so that a-c+b—d = 0 (mod m) by our earlier observation, i.e. (a +b) ~ (c + d) is divisi- ble by m. a +6 = (c+) (mod m). In the same way we have @=c (mod m), b =d (mod m) = ab = ed (mod m). In particular, a@=c (mod m) > a? = (mod m). More generally a = c (mod m) = a® = " (mod m), nEN. (Supply the details for justifying these assertions.) We now take up some problems involving these ideas. Example 1. All the 3-digit numbers from 130 to 164 are writ- ten consecutively to form the number N = 130131132--+ 164. Find the largest power of 3 that divides N. Solution. Let us first find the sum of the digits of N.24 NON-ROUTINE PROBLEMS IN MATHEMATICS loccurs (164—129)+4 = 39times; 2 occurs 4 Soccurs 10+4 4occurs 10+4 5 occurs 10+3 39 4times; theirsumis 8 14 times; their sum is 42 14 times; their sumis 56 13 times; their sum is 65 u 6 occurs 54+3 = 8times; theirsumis 48 7 occurs 3 = 3times; theirsumis 21 8 occurs 3 = 3times; theirsumis 24 9 occurs 3 = 3times; theirsumis 27 Total 330 Total sum of all digits in N is 330, which is divisible by 3 but not 9. Hence 3! is the highest power of 3 that divides N. Examples 2. Find the g.c.d. of 228 and 177. Solution. We now give a general process which is used to find gc.d. of two numbers using a fact known as Buclidean algo- rithm. ar 228 = LITT+51 in| 2-1-9 177 = 351424 si] -a—% Sl = 22443 “|! si-2—3 2 = 8340. 3] mu-8-o9 ‘The above process can be exhibited as Thus 3 is the g. of 228 and 177: For, 228 = 1.177451 1x (3 x 51 + 24) +51 = 1 x (3 x 514+ 24) 42x 2443 = 1x (3x 51424) +2x8x 343 =3x514+3x843%K8x343 =Divisipitity OF NUMBERS 25 3(51 +8 + 16 +1) =3 x 76. More generally we can find the g.c.d. of any two given numbers a and 6 using the above process. The fact which we used is the following algorithm: Euclidean Algorithm. If a is any integer greater than or equal to 0 and 6 is a’ natural number, then there exists a non- negative integer q such that a=bqtr where r is an integer satisfying the inequality 0 < r < 6. For, first observe that either a itself is a multiple of b or lies between two successive multiples of b. ie. bg < a < 6(q +1) for some integer q. If bg = a, then we can take r = 0 and write a = bg +r. If not, bg 4, >1z>-++ and there can be only a finite number of r’s between 0 and 6. When rn—1 = rndn41s then we see that (a,b) = g.ccd. of a and b= rp. For, it is easy to see from each of the above equatious that (a,8) = (byr1) = (risr2) = ++ = (ns 0) = Tn Check this argument with reference to Example 2. Exercise: Find the greatest common divisor of (i) 184, 7: (ii) 163, 24; (iii) 265, 53 using the above method. Example 3. Find one solution of 7z + lly = 13 in integers < and y and then the general solution of the equation. Solution. Note that (7,11) = 1. We have 7(-3) + (2) = 1 so that 7(13 x (-3)) + 13 x 2) = 13 (See Example 6 below). So z = —39 and-y = 26 is a particular solution. Let 2, y be any solution. Then a+ ly =13, 7(-39) +11(26) = 13 Hence T(x + 39) = 11(26 ~ y). Since 7 divides the left hand side, it must divide the right hand side. but 7 is relatively prime 11. So 7 divides 26-y. Thus +39 26-y 1 = an integer r. Consequently, z+39=1r, 26-y=7r, g=Ur-39, y=26-7r. On the other hand, if x = 11r-39, y = 26—7r, then the given equation is satisfied. Hence the general solution is given by z=~39+l1lr, y= 26-77,Divisieitity OF NUMBERS 27 where r = 0, £1, +2,- Example 4. Show that the equation 3z + Gy = 22 has no integer solutions x and y. Solution. First note that (3,6) = 3. If we can find integers p and q such that 3p + 6g = 22, then 3 must divide 22 which is impossible. Example 5. Find the least positive number that leaves the same remainder 3 when divided by 5, 7, 15, 21 and 33. Solution. The least positive number that leaves as remainder 0. when divided by 5, 7, 15, 21 and 33 is simply the Lc.m of these numbers which is 1155. So 1158 is the required number. Example 6. Given two natural numbers a and 6, let r be the least positive integer expressible in the form az+by=r with z and y integers. Then r is the g.c.d. of a and 6. Solution. For, it is easy to see that any common divisor of a and 6 is a divisor of r. So, it is enough to show that r is divisor of a and 6. Suppose that r does not divide a. Then, using Euclidean algorithm we can express a as a=qrtswithO 1, there exist PisP2s-** ;Pk which are primes and a1,02,--- ,a% € N, such that n= pt! - py? pyt where p1,p2,-- ,Pk are uniquely determined but for the order of their arrangement and _~—the-_correspond- ing a1,-++ a4 are uniquely determined. If m = pft.--pft, n = q-+-gft,(m,n) # 1 if and only if there exists a pair (A, LSE SKU SG SL p= Gy. Also (mn) = Perf where r; = py = qj for some pair (7, j') and j= max(ay, B;) and ane er Loam. (myn) = pips? --gfhglt of! where Bj = 0 if pi = gj for some pair (i, j) and a! = a, Bi = By if pi # gj for any pair (i,j) af = max(ai,f;) if pi = q for some j. Example 11, Find g.c.d. and l.c.m. of 228 and 177.30 NON-ROUTINE PROBLEMS IN MATHEMATICS Solution. 228 = 2?.3-19; 177 = 3-59. So, g.c.d. (228,177) = 3 Le.m. (228, 177) 3-19-59. Example 12. Find g.¢.d and L.c.m of the two numbers: 2.37.11? and 3°- 117-79, Solution. g.c.d of the two numbers is 3°- 112. L.c.m. of the two numbers is 2°. 37. 117. 75. Example 13. Find the least n such that n! is divisible by 990. Solution. 990 = 2-5-3?.11. n should be at least 11. In 11! we have the factors 2,5,3.and another 3 from the factor 6. Hence n= 11 is the least desired value of n. N o PROBLEM SHEET - Find the g.c.d. of 576 and 73 and find integers x and y such that 576x + 73y = (576,73) . Find the g.c.d. of the numbers 2! — 1, 9120 _ 1. . Show that for any natural number n, g.cd. (Qn +13,n+7) =1. - Show that for any natural number n the fraction }2+t cannot be reduced further. . Solve the following equations in integers x and y if they have solutions and then find the general solutions10. DivistBiLity OF NUMBERS 31 (i) 3a — 4y = 29 (ii) Liz + 12y = 58 (iii) 462 + 345y = 69 (iv) 462 + 345y = 92 (v) 46x + 345y = 24 (vi) 46x + 345y = 41 From the answers to (iii), (iv), (v) and (vi) above can you guess when an equation of the form az + by = c, where a,b and c are integers will have integer solutions for z and y. . Show that if the product of an odd number of odd integers is of the form 4n+1, then at least one is of the form 4n+1. . Solve the congruence equations: (i) 52 =3(mod 11) (ii) 7¢ = 4 (mod 3) (iii) 62 = 22 (mod 8) (v) Lz = 33 (mod 58) In each case find how many distinct solutions exist. . Solve the equation (i) 23 = 2% (mod 7) (ii) 8 = 2? (mod 7) (iii) 18 = 2? (mod 7) (iv) 10 = 2? (mod 7) (v) 19 = 2? (mod 7) (vi) 20 = 2? (mod 7) . Can a number whose digits which are one hundred zeros, one hundred 1's and one hundred 2’s be a perfect square? Justify your answer. Show that there exists n such that n+1,n+2,--» ,n+1999 are all composite.LESSON 4 PERMUTATIONS AND COMBINATIONS In this lesson we will discuss, how to solve non-routine prob- lems in permutations and.combinations. The following points may prove helpful in solving these problems. 1. Multiplication Principle of Counting: If the sets X1,X2,+++Xn contain a), a2,---a_ number of elements respec- tively, then n elements, one from each set can be selected in @1a2-+-@p ways. The reason is that if there are just two sets X;, Xo, for each choice of an element of X}, there are a2 choices of elements of X2 to go with this element. So, for the number a; of choices of elements of X1, the number of choices is a; x az. The same argument can be repeated for each choice of elements of X1,X2 when choosing element of X3 and so on. This principle of counting is clearly illustrated in the fol- lowing example. Example 1. In the following figure, there are 3 different routes from Madurai to Trichy, 4 routes form Trichy to Chennai and 2 routes from Chennai to Tirupati. Then in how many ways can one go from Madurai to Tirupati? Trichy Tirupati Madurai ChennaiPERMUTATIONS AND COMBINATIONS 33 Solution. Let the set A denote the set of routes from Madurai to Trichy, the set B those from Trichy to Chennai and the set C from Chennai to Tirupati. Now with each element of A we can associate any of the four elements of B. The number of elements in A is 3, each represeriting a route from Madurai to ‘Trichy and the number of elements in B is 4 each denoting a route from Trichy to Chennai. Also each association represents a route from Madurai to Chennai through Trichy. The number of such associations is clearly 3 x 4 = 12. Now let X be the set of 12 routes from Madurai to Chennai through Trichy. Arguing as before to form associations between elements of X and C, we can show that the number of routes from Madurai to Tirupati through Trichy and Chennai in that order is 12 x 2 = 24. Hence one can reach from Madurai to Tirupati in 24 ways. 2. Permutations and Combinations: Let there be three persons a,b.c to be arranged in a row. ‘There is op- tion therefore in the order in which they are placed. All possible ways are exhausted by the following arrangements. a, bc a,c, b baa ba, oa, b ¢ ba whose number is 6. The multiplication principle discussed in §1 can be used to conclude this fact as follows. The first place can be filled in 3 ways by putting one of a,b,c. After this the second place can be filled in.2 ways by putting one of a,b,c left out after filling the first place. After this there is only one left to be put in the third place. Thus the number of ways of34 NON-ROUTINE PROBLEMS IN MATHEMATICS arranging the three persons is 3 x 2x 1 = 6. Note. The persons are distinct. On the other hand if there is a red ball and two white balls and if they are to be put in three holes H;, H2, H3 in that order, there are only two ways of filling H (if the colour of the ball alone matters). Depending on whether we have chosen the red ball or the white ball, Hz can be filled in either only one way or two. In any case H3 can be filled after this in only one way. In other words, if R stands for red ball while W for a white ball the arrangements are Hy Hy Hs RWW w RW WWwWR In this case if the two white balls are put in two chosen holes the red ball has to find its place in the third one. Thus it is just a choice with no arrangement of two holes out of the three holes Hi, H2, Hs. With due regard to order of occurrence, the number of ways of arranging n things is known as the number of permutations of the n things. If we are not to arrange all the n things given but only r of them, r < n, the number of ways of arranging r things taking them out of n things is known as the number of permutations of r things out of 1n things and is denoted by nP,. We have seen above that 3Py = 3x 2x 1 = 3! (read as factorial 3; the old notation was | 3.) More generally, the argu- ment given earlier shows that ,Py =1x2x3x---xn= nl. The same argument helps us to get an expression for ,P, too. First, give names for the r places to be filled in the order Phy Payer Pre The first place can be filled in n ways taking out one of the n things, the second place in (n — 1) ways from the remain-PERMUTATIONS AND COMBINATIONS 35 ing things and so on. Finally, the rth place can be filled in n-(r—1) =n—r+1 ways. Thus the expression for nPr is given by nt Py moar n(n-1)---(n-r+1)= Note. After choosing for each place P,,--- ,P-, what matters is the arrangement in that order. Just the choig alone is not enough. Example 2. There are 5 chairs put in order Gi, C2, C3, Ca, Cs. In how many ways can 5 persons be seated in them. Solution. This is a straight forward question. The answer is Ps = 5! = 120. Example 3. Three gentlemen and two ladies are to be accom- modated one in each of 5 rooms in a row numbered 1 to 5 so that the ladies are put in rooms with even number. Find the number of such arrangements. Solution. If r1,r2,73,74,75 are the rooms, one lady has to be put in rp and another in r4. The number of these arrangements is 2P) = 2! = 2. The number of ways of arranging the three gentlemen in rooms r1,13,75 is sPs = 3! = 6. Thus the number of arrangements in the 5 rooms under the given condition is 2x6=12. Example 4. How many 3 digit numbers are there with all odd digits? Solution. The odd digits are 1,3,5,7,9, five im num- ber and the number of ways of arranging three of them is sPy= 3H =5x4x3=60. In the concept of permutations arrangement in an order was taken into account. If there are 5 chess players and we36 NON-ROUTINE PROBLEMS IN MATHEMATICS have to make games with a pair of them, the number of games possible does not depend on any ordering among them. It is only choice of a pair. We need not go in for any ordering (e.g. in the last example the order of the place values mattered) for this choice. Having chosen one of the players (possible in 5 ways) we can choose the other player in 4 ways. But we have allowed any one player to be chosen first as also second which does not matter for the game. So, of the number of ways, viz. 5 x 4 = 20 obtained earlier each choice of the pair is counted twice. Thus the actual choice of players for the game can be done in 20/2=10 ways. Supposing we have n things from which r things are to be chosen, without caring for any arrangement of the 7 things chosen, we should count all the arrangements af- ter the choice of r things among the r things as one choice. In other words, the number of choices alone of r things out of n things, called the combination of n things taken r at a time, denoted by ,C,("C; or (")) is given by ! n(n=1 Example-5. In how many ways can ‘two red ‘balls and two black balls be chosen from a bag containing 4 red balls and 3 black balls? Solution. The number of ways of choosing the ted balls is 4Cp and the number of ways of choosing the black balls is 3Cz. So, the number of ways of choosing the 4 balls is 4x3 3x2 4C2x 3C,=—~ x 7 = 18. Example 6. Ram has 7 friends. Each evening of 5 days in the week he invites 3 of his friends for dinner taking care that the same group of friends do not turn up once again. In how many ways can the groups be formed?PERMUTATIONS AND COMBINATIONS 37 Solution. If we want distinct groups we can fix up 2 and choose one from the remaining 5. But each group of 3 will be counted. So, the number of ways of forming the groups under the given condition is (7C2 x 5C;)/3 = 228 x § = 35. 3. Addition Principle of Counting: Some times the ways in which a selection or arrangement can be done may have to be divided into certain number of mutually exclusive and exhaus- tive cases - each case denoting a particular type of a condition or situation. In such a circumstance the sum of the number of ways belonging to each case will give the total number of ways in which the selection or arrangement can be done. This principle is illustrated through the following example. Example 7. There are five balls of different sizes and three distinct urns. Each urn can contain all five balls. In how many ways can the balls be placed in the urns so that no urn remains vacant? Case 1. Suppose that the balls are placed in the urns as fol- lows: 1 ball in the first urn, 1 ball in the second urn and 3 balls in the third urn. 1 ball in the first urn can be placed in sC; ways, 1 ball in the second urn can be placed in 5-10, = 40) ways and 3 balls in the third urn can be placed in s_2C3 = 3C3 ways, So, by multiplication principle the balls can be placed in sC1 x 4C1 X 3Cy ways. But the two urns (each containing 1 ball) can be selected from 3 urns in 3C> ways. So the number of ways of placing the balls in the urns in this case is sC, x 4C1 x 3C3 x 302 = 5x4x1x3= 60. Case 2. Suppose that 5 balls are placed in 3 distinct urns as follows: 1 ball in the first urn 2 balls in the second and 2 balls in the third urn. Arguing as in Case 1, we can show that the38 NON-ROUTINE PROBLEMS IN MATHEMATICS number of ways of placing the balls in the urns in this case is sCt x 4C2 x 202 X 30, = 5 X6 x1 x3 = 90. Now Case 1 and Case 2 are mutually exclusive and exhaus- tive cases. So the total number of ways of placing the 5 balls in three distinct urns so that no urn remains empty (by using addition principle of counting) = Number of ways in Case 1 + number of ways in Case 2 = 60 + 90 = 150. 3. Some times the construction of a function f(n) and a linear equation connecting f(n), f(n — 1) and f(n — 2) where n € N or NU {0} may prove useful as in the solution of the following example. Example 8. There are n letters and there are n envelopes with n distinct addresses to which the n distinct letters are to be sent. In how many ways can one place all the letters in all wrongly addressed envelopes. Solution. Let all the letters be denoted by ai,a2,++-aq and the envelopes with corresponding addresses by Ay, Ag, +++ , An. Let f(n) denote the required number of ways ie., the number of ways in which all the n letters can be placed in the wrong envelopes. Here n is a natural member. Now the letter a; can be wrongly placed in any one of the (n ~ 1) envelopes Az, Ag," , Ax. Suppose that a; is placed in the envelope Ay. Then two cases arise:- Case 1. Let the letter a,,k # 1, be placed in the envelope Ay which is a wrongly addressed envelope for the letter a,. In this case remove the letters ay, ay and the envelopes Ay, Ag the number of ways in which all the remaining (n — 2) letters are Placed in the (n — 2) envelopes all wrongly is equal to /(n — 2), ie., all the n letters are placed in the wrongly addressed en- velopes in f(n — 2) ways in this case,PERMUTATIONS AND COMBINATIONS 39 Case 2. Suppose that the letter a, is not placed in the envelope Ai. Then in this case all the (n — 1) letters 4, 03,-+ ,@q have to be placed wrongly in the (n—1) envelopes Aj, Aays+* )Ak-1) Ai, Asis >> sAn(2 Sk » Ak-1, Ai, Ak4i,**+ , An and we are to get the number of ways in which the (n — 1) letters should all be placed in the (n — 1) envelopes wrongly. By definition of f this number is {(n—1). So, if the letter a; is placed in a particular envelope Ax,k # 1, wrongly, then the total number of ways in which all the n letters are placed in the n envelopes wrongly, is, by Case 1 and Case 2, equal to f(n— 2) + f(n— 1). But the same argument can be repeated if the letter 2 is placed into any of the (n — 1) wrongly addressed envelopes. So the total number of ways in which all the n letters can be placed in the wrongly addressed envelopes is equal to (n — 1){f(n—1) + f(n—2)}. So we have the following linear relation connecting f(n), f(n — 1) and f(n—2): f(r) ~I{f(n-1) + f(r—2)} > f(n)-nf(n—1) = -{f(n-1)-(n-1f(n—-2)}. Similarly, f(n=1) —(n-1)f(n-2)=-{F(n - 2) -(n- 2) f(n - 3)} £(3) - 3f(2) = -{F(2) - f(1)}- But clearly we have f(2) = 1 and f(1) = 0. So (3) - 3f(2) -{1-0}=-1, £(4) — 4f(3) (-1)(-1) = (-1), u40 NON-ROUTINE PROBLEMS IN MATHEMATICS 7(5) -57(4) = —(-1)? = (-1)5, Hn) =nf(n-1) = (-1"? = (-1)" which implies that f(x) _ nf(m=1) _ (-1)" n! n! n! f(r) _ (m=) _ (-1" a @ei Q) Similarly, f(n=1) _ f(m=2) _ (-1)"*! (2) (=I)! (n-2)! f(n~2)_ f(n-3) _ (1)? (n=2)! ~ (3)! (2 &) £2) _ FQ) _ (-1)? Adding these n — 1 equations (from (1) to (n — 1), we have nee =I)" foyant {edad a}. Hence the total number of ways in which n distinct letters can be placed into n wrongly addressed envelopes in equal to ede (-1)" ann gtgct al }. Note. The solution can be simplified using the exclusion in- clusion principle. 4. Division of objects into groups. We illustrate this method by examples. tPERMUTATIONS AND COMBINATIONS 41 Example 9. A fruit-seller sells 4 types of mangoes; neelam, malgoa, sendur and daseri. In how many ways is it possible to purchase 7 mangoes? Solution. This problem is not a problem of permutating with repetitions, for the order of the mangoes in any purchase is immaterial. This problem is closer to the combination type; and the combination may involve repeated elements. For convenience we will represent each purchase by-means of zeros. First write the number of units to denote the number of neelam mangoes. Then put a zero to separate neelam from malgoa. Then put the number of units to indicate the number of malgoas purchased. Again put a zero to seperate malgova from sendurs, and then write the number of units to denote the number of sendurs purchased. Again put a zero to separate sendur from daseri and finally put the number of units to denote the number of daseris purchased. For example, if we purchase 2 neelams, 3 malgoas, 1 sendurs and 1 daseri, the purchase will be denoted by 1101110101. If there is no unit between two zeros, it means that the particular variety of mango has not been purchased. For example, 00111011011 means no neelam, 3 malgoas, 2 sendurs and 2 daseris. So, the number of distinct purchases is equal to the number of permutations of 7 units and 3 zeroes, taken all at a time. That is, the required number is #g1 = {29% = 120. Example 10. Find the number of ways in which (m+n +p) different things can be divided into three groups containing m,n and p things respectively. Solution. First select m things out of (m-+n-+p) things which can be done in min+pCm Ways, then select n things from the remaining (n + p) things in n4pCn ways and finally, p things42 NON-ROUTINE PROBLEMS IN MATHEMATICS and to be selected from p things in »C, ways. So the required number of ways is imentp)Om X ntpCn * pCp (m+n-+p)! i (n+p) elie (m+n+p)! (n+ ppl nip! minipl Caution. If m = n = p, then all the 3 groups have equal number of things say, m. So the required number of ways will be = 5. Applications to geometry. In combinatorial problems in- volving geometry the students should note that (1) Two points determine a straight line and two non-parallel straight lines de- termine a point called the point of intersection. (2) 3 points which do not lie on a line determine a triangle. (4) Four points in a plane, no three of which lie in a line determine a quadri- lateral. These facts facilitate in counting the required number of geometrical objects (points, line etc.) in specific problems as illustrated in the following example. Example 11. There are n points in a plane no three of which are in the same straight line with the exception of m(< n) points which are all in a straight line. Compute the number of (a) straight lines and (b) triangles which can be formed by joining the points. Solution. (a) As two points determine a straight line, the number of straight lines formed by ing n points is »Cp. But m of these n points lie on a straight line, so joining any two of these m points will determine the same line and no other line is formed. So to find the required number of lines we have to substract mC2— 1 from C2. That is, the required number of straight lines is »C2 = mC2 +1. (b) Arguing as in (a) we can conclude that the number ofPERMUTATIONS AND COMBINATIONS 43 triangles formed is »C3 ~ mC. Note that any 3 points lying in the same straight line do not produce any triangle. PROBLEM SHEET 1. A cricket team of 11 players is to be formed from 16 play- ers including 4 bowlers and 2 wicket keepers. In how many different ways can a team be formed so that the team consists of at least 3 bowlers and at least one wicket keeper. 2. Find the number of ways in which 12 students may be equally divided into 3 groups. 3) Find the number of ways in which 12 mangoes may be equally divided among 3 boys. 4. Mr. Mohan invites a party of 15 guests to a dinner and places 10 of them at a round table and the other 5 at another round table, Find the total number of ways in which he can arrange all the guests. 5. How many different packets of mangoes, eight mangocs in each packet, can be formed from four varieties of man- gocs? 6. A lift with seven people stops at ten floors. Froin zero to seven people go out at each floor. How many ways are there for the lift to empty. (The ways only differ. by the number of people leaving at each floor)? 7. Show that the number of different n-sided polygons formed by n straight lines in a plane, no three of which meet in a point, is }(n — 1). 8.-A parallelogram is cut by two sets of m lines parallel to its sides. Find the number of parallelograms then formed.44 NON-ROUTINE PROBLEMS IN MATHEMATICS 9. If n straight lines of indefinite length are drawn in a plané. no two of them being parallel, and no three meet- ing at a point, then show that the plane is divided into $(n? +n + 2) parts. 10. In how many ways can the candidates Ay, Az, A3++» , A1o be ranked so that A, is always above Aj? 11. In how many ways can an odd number of objects be se- lected from n objects?LESSON 5 NUMBER THEORY Divisibility Let a,b be integers and b # 0. We say that a is divisible by b if there is an integer c such that a = bc; in this case, we write bla. Clearly, 1[¢ for any a, ala for any a £ 0. Moreover, if cla and clb then c|(az + by) for all integers x,y. Primes An integer n > 1 is said to be prime if it has no divisor other than +1, tn. An integer n > 1 is said to be composite if it is not prime. Let n be an integer greater than 1. Whether n is prime or not it has divisors a with 1 < a < n; the least of such divisors must be prime. Consequently, every integer n > 1 is either a prime or is a product of primes. The fundamental theorem of arithmetic states that an expression of n as a product of primes is unique except for the order of the factors. Greatest common divisor Let a and 6 be integers of which at least one is nonzero. Consider the set of all positive integer n of the form n = ar-+by where z and y are integers. (Verify that such positive integers nexist). Let d be the least of such positive integers n. Then d is a positive integer such that (i) d= ax + by where x and y are integers (ji) dla,d|b, and (iii) cla, clb = cla46 NON-ROUTINE PROBLEMS IN MATHEMATICS The number d is called the greatest common divisor (gcd) of a and b, and is denoted by (a,6). If (a,b) = 1 then a and b are said to be relatively prime. Let a, be positive integers and m = oh then (i) alm, 6m, and (ii) a|n, dfn > min ‘The number m is called the least common multiple (Iem) of a and b, and is denoted by [a,b]. Congruence Let m be a positive integer. If a and 6 are integers such that a — 6 is divisible by m then we say that a is congruent to b modulo m, and write a = b(mod m). Theorem 1. Let a = b (mod m) and c = d(mod m). Then (i) az = be(mod m) for all integers 2 (ii) @+c=b-+d(mod m) (iii) ac = bd(mod m) (iv) a* = 8*(mod m) for all positive integers n () if f(z) is @ polynomial with integer coefficients then f(a) = f(b)(mod m). Theorem 2. (Cancellation Law). If ac = be(mod m) and c is relatively prime to rm then a = b(mod m). Theorem 3. (Fermat’s Theorem). If p is prime and a is relac tively prime to p then a?! = 1(mod Pp). Consequently, if p is prime then a? ger a. a(mod p) for any inte-Num ‘THEORY 47 Euler’s extension of Fermat's Theorem Euler's function (nr) is defined as follows: (1) a n> 1 then (1) is the umnber of positive integers less than ively prime tou. ‘Thus (2) = 1,4(3) = 2, and — 1 if p is prime. ‘The general formula for finding (72) Ifn > 1 then 4(n) = nfl,,,(1-1/p), whe is extended over all prime he product sors p of 1. Theorem 4. (Euler's Theorem). [fa is relatively prime to m then a") anc so Fermat’s theorem is a special case of Euler's (mod m). If m = p,uprime, then g(m) = p= 1 heorem. Theorem 5. (Wilson’s Theorem). If p is prime then (p— 1)! +1 =0 (mod p). Wilson’s theorem has a converse which states that if n is an integer greater than I and (n ~ 1)!+1 = 0(mod n) then n is prime. However, no such converse of Fermat's theorem is true. Infact, if a is relatively prime to 561 then a = 1(mod 561) while 561 is not prime [prove this}. Examples. Example 1. Find all integers n for which “£1384? also is an integer. Solution. Divide n? +13n +2 by 3n+5. We get the quotient n/3 + 34/9 and remainder ~152/9. So if 2 n+13n 429045 _ 3a +5 then " (3n + 34)(3n +5) — 152 (3n + 5)(3n + 34 — 9m). 9m(3n +5) = Y(n? + 13n + 2) 152 If m is an integer then 3n + 5 is a divisor of 152 = 8x 19. The48 Non-ROUTINE PROBLEMS IN MATHEMATICS divisors of 152 are £1, £2, +4, £8, £19, +38, +76 and +152, Heuce Bn +5 = £1, £2. £4. 48,419. £58. 476. £152 Bn + 34-9 = £152,476. 438.19. +8, £4. £2, 41 Considering the 1-47 signs 9n-29 = -151.-74. Ym = ~122.-45, Bd = 5. 18.40. 63, 103. 180. So the possible values af nave —5.2.7 and 20: these correspond to $n +5 = 2.8.38, and 152. Thus nm =—1.1.11.49. Next consider the (/~/) signs. We get Ym — 29 = 151, 74,94. 1. - LL, =H. 74. 151 These give the same possible values of mm. namely 20,7, 2and=5; -1,-4.-19 and -76. Thus Thus desired values of n are 1.11,49, -1 5.2.7 and 20. —8 and —27. The possible values of m are Example 2. Each of the numbers 1,22,--- 2, equals 1 of =1, and wy Tatgty +2: Fye5 A LaTEM HG +--+ Sy 1 LyX Tz + Ty T2Ty = 0. Prove that n is divisible by 4, Solution. Let yi = risitiy2ti43 for i = 1,2,.-+,n, where 2 and sings = x3. Then yy tyg+---+¥n = 0 and yi = £1 for each i. Suppose that y; = 1 for m; values of ¢ and yj = —1 for np values of i. Then ny + nz Ing) = Ty Tn42 Om ty2te ty = m1(1) +n2(-1) =n —n2.NUMBER THEORY 49 Hence n = n2,n = 2ng and ives Ym = (1)™(—1)" = (1), But Vive Yn = Bigg s+, = 1 This shows that ng is even; hence n = 2nq is divisible by 4. Example 3. Which of the numbers 101, 10101, 1010101, with alternating 0's and 1’s, beginning and ending with 1, can be primes? Solution. Let Nz = 101,.Nj = 10101, and in general, 100* — Ng = 1010---01 = 14 100+--- + 100°? Here N, involves k 1's. If k is odd then 10k 1) (10% +1 m= (S2) FE) This is composite since 10* + 1 is divisible by 10+1 if k is odd. If k is even and k = 2r then 100-1 _ 100" -1 9 ~=~=—C(Y 99 N= ) (100° +1) This is composite if r > 1. Thus only one number in the list, namely Np = 101, is prime; the remaining are composite. Example 4. Prove that 2222555 + 5555?22? is divisible by 7. Solution. Dividing 2222 and 5555 by 7 we get the remainders 3 and 4 respectively. So 2222 = 3(mod 7) and 5855 = 4(mod 7). Thus 22225885 35555 (mod 7) and 55557??? = 4?22(mod 7).50 NON-ROUTINE PROBLEMS IN MATHEMATICS Fermat's theorem statés that if a is not divisible by p then @-! = 1(mod p). Thus 36 = 1(mod 7) and 4° = 1(mod 7). Hence 3" = 1"(mod 7) and 4%" = 1"(mnod 7) for all positive integers n. Divide 5555 and 2222 by 6. we see that 5555 = 925 x6 +5, 2222=370x6+2 Hence 38555 925x635 = 35(mod 7), 47222 = 48706 4? = 42(mod 7). Since 3° = (mod 7) and 4? = 2(mod 7), we see that 35555 4 42222 = 5 4.2 = 0(mod 7) This proves the result. Example 5. Find the last two digits in the decimal represen- tation of 31999 + 21999 | Solution: We are required to find the least nonnegative integer r such that 31999 4, 91999 (mod 100). Use Euler's theorem which states that if a relatively prime to m then a®) = 1(mod m). Since 100 = 2. 5? we see that (100) = 100 (: ~ ) ( - 5) = 40. 1(mod 100) and so Thus 3° 32000 (mod 109). Since 3 x 67 = 1(mod 100) this gives 31999 = 31999(3 x 67)NuMBER THEORY 51 Since 2 is not relatively prime to 100 we proceed as follows: Since $(25) = 25(1 - 3 = 20 we have 29 = 1(mod 25), and so 22000 = 1! (mod 25) Since 8 x 22 = 1(inod 25) this gives 21997 = 91997(g x 22) = 22(mod 25). which implies that 21997 = 25m + 22, 21899 = 100m +88 Hence 88(mod 100). 31999 4 91999 67 + 88 = 55(mod 100) so the last two digits are 55. PROBLEM SHEET 1. Find all integers n (positive, negative or zero) for which n? +n+41 is a perfect square. 2. Find all integers n (positive, negative or zero) such that n? +73 is divisible by n + 73. 2 . Show that every positive integer n > 12 can be expressed as the sum of two composite numbers. = . Find all positive integers which can be expressed as the sum of three distinct composite numbers. on |. Is there a power of 2 such that it is possible to rearrange its digits giving another power of 2752 ° °° ° 10. ll. NON-ROUTINE PROBLEMS IN MATHEMATICS . Prove that each member of the set of n — 1 consecutive integers nl+2nl4+3,--,nltn is divisible by a prime which does not divide any other member of the set. . If p is a prime other 2 and 5, prove that p divides an infinite number of the numbers 1,11, 111, 1111,--- . Each of the numbers 1?, 2, 32, --- 1999? is replaced by the sum of the digits forming the number. In this operation one-digit numbers do not change whereas all other num- bers decrease. Then each of the 1999 numbers is again replaced by the sum of its digits, and the operation is per- formed repeatedly until we obtain a sequence of one-digit numbers. Find the sum of these 1999 one-digit numbers. ). Find the remainder obtained when the number 10! + 10007) 4... 4 y9(t0) is divided by 7. Prove that if a is relatively prime to 561 then a560 = 1 (mod 561). Find another composite number n with the property that if a is relatively prime to n then a"-! = 1 (mod n). (a) Show that 279.= 37(mod 83) (b) Find a positive integer n such that 7n25—10 is divisible by 83.LESSON 6 CONCURRENCY AND COLLINEARITY § 1. We are now entering into the fascinating branch of mathe- matics, namely, Geometry. It is like the branches of any science which were developed from time immemorial. Human mind pondered over two important mathematical concepts, namely the quantitative study of phenomena and the physical forms in nature. Geometry arose from the second idea. We hope that the reader is familiar with the basic concepts of geometry upto the school level. Most of the pure geometrical theorems in elementary geometry are concerned with: triangles and circles. Regarding triangles, we deal with in school courses special points like centroid, orthocentre, incentre, ex-centres. We also find that these special points are points of concurrence of some three lines through the vertices of the triangle. In this lesson, we learn two general theorems called Ceva’s theorem and Menelaus theorem. Menelaus of Alexandria, who lived around 100 BC wrote on geometry and trigonometry. His theorem was forgotten for a long period and was rediscovered by Giovanni Ceva, an Italian hydraulic engineer and mathe- matician. Ceva published both theorem in 1678. Definitions 1. A line segment joining a vertex of a triangle to.any point on the opposite side (the point may be on the extension of the opposite side also) is now called a Cevian. Altitudes, medians, angle bisectors are all Cevians.54 NON-ROUTINE PROBLEMS IN MATHEMATICS 2. Three points A, 8, C are said to be collinear if they lie on a straight line. 3. Three straight lines are said to be concurrent if all the three pass through a point. Theorem 1. /f A,B,C and'A’, B';C’ are points on two parallel lines such that $f, = Ber then AA’, BB',CC' are concurrent if they are not parallel. Note: Here the points are taken in the same order and in the same direction. Proof. Let AA’ and BB! meet at O when AA! and BB’ are not parallel. Join OC and let it cut A’B’ at Cy. By basic proportionality theorem we have ° BC _ AB BC BC AB” BC = BIC, = BIC’ => Gand C' coincide. ‘Thus CC" passes through O. Theorem 2. From a point O, OD,OE,OF are drawn perpendicular to the sides BC, CA, AB respectively angle ABC, then BD? — DC? + CE? — EA? + AF? — FB? = 9, of a tri- Proof. a BD OB?-oD? | pe 00? — op? Fie? > BD?-DC? = OB*-o¢?. Similarly, CB-EA = 008-04? Z AF? - FB? = OA?-0B?, : .CONCURRENCY AND COLLINEARITY 55 Adding we get the result. Exercise 1. In the above mentioned proof the point is taken inside the triangle. The reader can prove the theorem taking the point O outside the triangle. What will happen if the point lies on one side of the triangle? Theorem 3 (Converse of Theorem 2). If D, E, F be points on the sides BC,CA, AB of a triangle ABC such that BD? - DC? + CE? - BA? + AF? - FB? 0 then the perpendiculars at D,E,F to the respective sides are concurrent. Proof. Let the perpen- diculars at D,E to BC,CA respectively meet at O. Let OF! be the perpendicular from O to AB. Using Theo- rem 2 we have Fig. 3 BD? - DC? +CE*- EA? + AF" - F'B?=0. (1) But BD? - DC? + CE? - EA? + AF*— FB? =0. (2) > AF! - F'B? = AF? — FB? or (AF! + F'B)(AF'—F'B) = (AF +FB) (AF - FB) => AB(AF! - F'B) AB(AF - FB)56 NON-ROUTINE PROBLEMS IN MATHEMATICS > AF! - F'B = AF-FB (since AB ¥ 0) AF~AF! = FB-F'B or FF! = -FF! or 2FF! = 05> FF'= ie. FF coincide Exercise 2. Use Theorem 3 to prove that the perpendicular bisectors of the sides of a triangle are concurrent. The point of concurrence is called the circumcentre. ‘ Theorem 4 (Ceva’s Theorem). ABC is a triangle and AX, BY and CZ are three concurrent cevians. Then XC YA ZB-* Proof. Let the cevians meet a at P. Let [ABC] denote the area of the triangle ABC. 1 gBX-h 1 gXen [BPX] il [CPx] B where h is the length of the perpendicular from P to BC Fig. 4 [BPX] _ Bx }X0-h ~ [CPX] ~ XC Similarly, if A’ is the perpendicular from A to BC, then ABX-h! }XG-h ” (CAX]ConcuRRENCY AND COLLINEARITY 57 Thus #5 ae From (1) and (2) (2) _ [BAX] _ [BPX] @) ~ (GAX] ~ (CPX) From elementary algebra we have the result: ac ac a-c oaS ti ios’ =, = is te If 5 = jy then each of the ratios’, $ is equal to f— Bx BAX|-[BPX APB So, (3) = 8% = fee = HEA. Therefore, 3X _ [APB) = (AES. P « Similarly GY = (BPC) og AZ = (APC! spttinlying the YA [APB] 3 (BPC) equations we get 2%. GX. HH =1 Note: 1. If A,B,C are three points on a straight line in the order; we introduce a direction in the following manner: AB, BC, AC ate taken-to be positive and BA,CB,CA are taken to be negative. Thus AB + BC = AC and AB+ BC+CA=0. 2. In fog. 4 BX and XC are of same sign, CY and -YA are of same sign and AZ and ZB are of same sign. ‘Thus BX CY AZ Sa tz’ is positive. 3. In Ceva’s theorem, if P lies outside, as in the following di- agram, then, BX,XC are pos- itive CY positive, YA is nega- tive AZ is positive, ZB is nega- tive. Thus 3% is positive; $% is negative; 2% is negative. Hence BX CY AZ58 NON-ROUTINE PROBLEMS IN MATHEMATICS Theorem § (Converse of Ceva’s Theorem). If three ce- A vians AX,BY,CZ satisfy $2: - Fa 2% = +1 then they are con- current: YY Proof. Let BY and CZ ineet at P. Let AP meet BC at X'. Then by Ceva's theorem w® x c Fig. 6 (2) BX! _ BX |. 5 ‘We have XO Xe Adding 1 to both sides, BX! | _ BX wig **=xer BX'+X'C _BX+XC__ BC _ BC XC ~~ xe * xd = xe X'C=XC = X'C-XC=043Xxx!=0 Therefore, X,X! coincide. Thus the three cevians are con- current. Note. The converse of Ceva's theorem is more useful than the theorem in the sense that most of the elementary theorems regarding concurrency can be proved using the theorem, As an example you may remember the lengthy proof of the theorem: The medians of a triangle are concurrent,CONCURRENCY AND COLLINEARITY 59 If D,B,F are the mid points of BC, CA and AB of m a triangle ABC, then BD CE AF _ iz ¢ c Do BA Fp 7 lXixt=l. Thus the cevians AD, BE and GF are concurrent. B D ce Exercise 3. Prove the following Fig.7 . The altitudes of a triangle are concurrent. nw The internal bisectors of the angles of a triangle are con- current. 3. The internal bisector of an angle of a triangle and the other two external bisectors are concurrent. Example 1. M is an interior point of a triangle ABC. Bisectors of interior angles BMC,CMA,AMB_ intersect BC,CA, AB respectively at X,Y,Z. Prove that AX, BY,CZ are concurrent. If P is the point of concurrence and show that M is the cir- cumcentre of AABC. Solution. MX is the bisector of BMC. So, 24 = M8. In the same way $4 = MG, 3% = Mh s0 that60 NoN-ROUTINE PROBLEMS IN MATHEMATICS Hence, by the converse of Ceva’s theorem, AX, BY,CZ are concurrent. If P is the point of concurrence of AX, BY, CZ, let Fig. 9 area of triangle BPC = Ai, area of triangle CPA = Ag, area of triangle APB = A3. PA SAPB _ AAPC _ AAPB+AAPC PX QPBX ~~ APCX” ABPX + QPCX — ArtAs eee Ane Similarly PB_A3+A, PC A\+ Ay PyY~~ a, *™4 pz = A We know that for any two positive real numbers a,b. oS vas. Thus A, + Ag > 2VAi,Az, Ap + As > 2VApAG, and Ag + A, 2 2VA3A), or (Ai + A2)(Aa + A3)(43 +41) > 8VAARVARASVAGAL = 8A) A2A3. (Ai + A2)(A2 + As)(Aa + Ai) or Frreerecree Ay Away ieee 28.Conourrency AND COLLINEARITY 61 The equality holds when Ay = Ay = Ap. BX _ AABX _ APBX _ AABX -APBX XG ~ BACX ~ APCX ~ BACX-APOX = ed Ay > BX = XC. But BX is the bisector of BMC. MB = MC. Similarly MC = MA. Thus M is the circumcentre of AABC.. Since BX = XC, AX is the median. Similarly BY and CZ are also medians, Thus P is the centroid of the ‘triangle. Example 2. AD,BE,CF are three concurrent. lines drawn from the vertices of the triangle ABC to points D, B, F on the opposite sides. If AD is an altitude of the triangle A'BC, show that AD bisects the angle FDE. Solution. Through A * draw a line parallel to BC-DE meets this line at X and DF meets this line at Y, Consider the triangles AX E and EDC AXB = EDC, AEX = DEC and XAC = ECD. Thus the two triangles are similar. So, cE_po Ba" Ex (1) Similarly from the similar triangles AFY and BFD62 NoN-ROUTINE PROBLEMS IN MATHEMATICS AF _YA Oe cies (2) FB BD Sinee the lines AD. BE, CF are concurrent, by Ceva’s theorem BD CE AF _, ©) DC EA FB ' From equations (1), (2) and (3) we have AY = AX. Since AD is the altitude DAX = DAY = 90°; so DX = DY. ie. YDX is an isosceles triarigle. Hence DA bisects FDE. ° § 2. Menelaus Theorem. Before giving the statement of the theorem, the reader is advised to draw the following figures and come to a conclusion. (i) Draw a straight line cutting only one side of a triangle. (ii) Draw a straight line cutting only two side of a triangle, (iii) Draw a straight line cutting three sides of a triangle. ‘Theorem 6 (Menelaus Theorem). If a transversal cuts the sides .BC,CA, AB of.a triangle ABC at X,Y,Z respectively, then Proof. Let hi,hz,h3 be the lengths of Perpendiculass AP, BQ, CR respectively from A,B,C on the transversal, Triangles BQX and CRX are similar so that (1) Similarly YA"h (2)ConcurreNcY AND COLLINEARITY 63 AZ _ ly ZB i (3) A x¢ As per the directed lengths we have 4% ix negative and the other two ratios are positive XO YA ZB Exercise 4. Prove Menelaus theorem in the-case of a trausver sal cutting all the sides externally. Theorem 7 (Converse of Menelaus Theorem). /f X.Y,Z ure three. points on each of the sides BC,CA. AB of AABC or on their extensions such that $4. Gh . 3% = 1, then X.Y,Z are collinear, 7,a NON-ROUTINE PROBLEMS IN MATHEMATICS Proof. Let ux yine the proof af thin theorem for a transversal which euts all the three sides externally. X,¥,Z are points on the extensions of BC.CA and BA respectively such Uat (a) Produce ¥Z to mect BC extended at X’. As in the case of . we shall prove that X, X’ coincide. converse of Ceva’s theo We have Meuelans Theorem (2) From (1) aud (2) we have os xe Subtract 1 from both sides ie. ie. yo 7 X'C-XC=0, ie. XX" =0; or X,X" coincide. Note: The reader may have a doubt that whether X lies on the right or left of X’, Tn the proof given above X has beeat taken on the left of X‘, Exercise 5. (1) If X lies ou the right of X', prove the theorem. (2) Prove the theorem for the transversal cutting the sides in all possible anamer. Example 3. The tangents at the vertices of a triangle to its ci ircle meet the opposite sides in three collinear Points.Concurrency AND COLLINEARITY 65 Solution. The tan- gents at A,B,C meet the opposite sides re- spectively at D,E, F. DAB = ACB (using alternate segment. the- orem). Consider-trian- gles DCA and DAB. Dis common, DAB = ACD. Thus DBA = BAC and the triangles are similar. So, BD AD BD BD AD_ AB? @ a DG ~ AD GD~ AC? ~ BR Since D lies out- side BC BD_ é& A DEB k Similarly Sh=-S, HB = 8. ‘ Therefore BD. CE. AR 1, So® : © 7" D,E, F are Collinear. Fig Example 4. In a triangle ABC, AB = AC. A transversal intersects AB and AC internally at K and L respectively. It intersects BC produced at M. If KL = 2LM, find 42. Solution. The transversal BCM intersects the sides KA,AL,LK of the triangle KAL at B,C,M respectively. By -66 NON-ROUTINE PROBLEMS IN MATHEMATICS Menelaus Theorem, KB AC LM __ BA CL’MK~ But AB = AC and KM = 3LM. So, Example 5. ABC is a trian- A gle and D and E are interior points of the sides AB and BC > { respectively such that $2 = + and €§ = 3. If AB and CD intersect at F. find £5. Solution, Consider the trian- gle BCD and the transversal EFA, It cuts BC, CD, DB at E,F and A respectively. By 7 7 Menelaus theorem Fig. 15 Q Example 6. Three points X,Y,Z are taken on the sides BC,CA, AB respectively of a triangle ABC such that " AX, BY,CZ are concurrent. YZ meets BC in X'".ZX CA in Y',XY meets AB in Z', Prove that meets (i) X',¥!, 2! are collinear (ii) AX, BY',CZ! are concurrent (iii) AX’, BY, C2! are concurrent (iv) AX’, BY’, CZ are concurrent,CoNcURRENCY AND COLLINEARITY 67 Solution. The cevians AX, BY,CZ are concurrent. So BX Gy Az xe YA ZB~* ® The transversal X'YZ cuts the sides of the AABC. By Menelaus Theorem se Se eel. (2) Similarly consider the transversals XYZ' and XY'Z with re- spect to the triangle ABC. We get Re ya ppt (3) xe A ZB~"! “68 NON-ROUTINE PROBLEMS IN MATHEMATICS By multiplying the equations (2), (3), (4) we get BX' CY AZ BX CY AZ' CY' AZ _ X'C YA ZB XC YA 2B Y'A ZB Using (1) we get BX! oy" Az! XC YA 7B X',Y', Z/ are collinear by the converse of Menelaus Theorem. It is easy to prove using (1), (2), (3), (4) the other results also. §3. In 1746 M. Stewart stated a theorem regarding the met- ric relations among a cevian and the sides of a triangle. It is believed that this particular result was known to Archimedes about 300 BC. The first proof was given by R. Simson in 1751. Theorem 8 (Stewart's Theorem). Let AX be a cevien of length p, dividing BC into segments BX =m and XC =n, Then a(p? +mn) = b’m+ en, A Proof. Using cosine rule for triangles ABX and ‘ACX we a b have Fig. 16a cos@ = (a) cos(180°— 0) = (2) where AXB = @ (and AXC = 180 - 6). So,ConcurRENCY AND COLLINEARITY 69 n(p* + m? — ¢?) or = -pn-nm+en ie. Pmt+en = p'm+p'n+mn? + nm? p(m+n) +mn(m+n) or Pm+en = -(m+n)(p? +mn) or m+n = a(p?+mn). Example 7. To find the length of the medians of a triangle. Solution. Using Stewart's theorem we have a ‘ale Oe Pee 242% Boras = alates f i .. Pte a? © ie SEE = aves P+ a? ie, AD? = <= 2 46a DO 1. = 1 2 ua git he AD = 5 VOIP Ey Exercise 6. Apply Stewart's theorem to prove that the length of the bisector of the internal angle A of AABC is , iG] Note. Consider two triangles ABC and A'B'C' which are not congruent but the respective sides are parallel and one within® the other. Then AA’, BB',CC’ extended are concurrent. ‘The proof of this result resembles that of the converse of Ceva's or Menelaus Theorem. Assuming BB! and CC" meet at O, let OA to pass through O. Let it cut A’B’ at Ay (say). Prove that A, coincides with70 NoN-ROUTINE PROBLEMS IN MATHEMATICS A! (use properties of similar triangle). Here the triangles are called “homothetie triangles”. Now we shall see that most fascinating theorem of plane ge- ometry. This theorem A has lot of applications in a branch of mathe matics called projective geometry. This theo- ZA\ rem was first proved by aN the great mathemati- Zorn TeS 2 cian Pappus of Alexan- dria (about 300 BC) and is known as Pappus Fig. 18 % Theorem, Theorem 9 (Pappus Theorem). If A,C, E are three points on one straight line, B,D, F on another, and if the three lines AB,CD,EF meet respectively DE,FA and BC at L,M,N, then these three points L, M,N are collinear,ConcuRRENCY AND COLLINEARITY 7 It is quite interesting to draw (indeed very easy to draw only with a ruler) the situation and convince yourself. The proof is left as a good exercise for the reader to usc Menelaus Theorem for different sets of triangles and transver- sals. To facilitate the reader, the adjacent diagram is given. Use Menelaus Theorem for the triangle UVW, considering the transversals LDE, AMF, BCN, ACE, BDF. v PROBLEM SHEET . Ifa straight line cuts the sides AB, BC,CD and DA of a quadrilateral ABCD in P,Q, R and S respectively, show that op oe RD sa7t PB QC RD SA~ Prove that the necessary and sufficient condition that lines from the vertices A, B, C of AABC to points X,Y, Z on the opposite sides are concurrent is that sinBAX sinCBY sinACZ _ sinCAX sinABY sinBCZ — +1. . Prove that the external bisectors of the three angles of a scalene triangle meet their respective opposite sides at three collinear points. |. P is an interior point of a triangle ABC and AF, BP,CP. meet the opposite sides BC,CA,AB in D, E,F respec- tively. Show that AF AE _ AP FB EC” PD72 NoN-ROUTINE PROBLEMS IN MATHEMATICS . Three. circles (whose centres form the vertices of a tri- angle) touch two by two. Prove that the three common tangents at the points of contact are concurrent. - A circle touches the sides of a triangle. Prove that the lines joining the vertices to the points of contact with the opposite sides are concurrent. . If AD, BE,CF are eevians of AABC concurring at a point O, then prove that oD , OF , OF AD BE CF . (Ifa given straight line AB is divided internally at P and externally at Q in the same ratio, then AQ is said.to be divided harmonically at P and Q. P and Q are called the harmonic conjugates of A and B.) D,E,F are poirits on the sides BC,CA, AB of a trian- gle such that AD,BE,CF are concurrent. If EF cuts BC produced at D’, prove D and D’ are the harmonic conjugates of B and C.LESSON 7 MATHEMATICAL INDUCTION § 1. Almost. everyone has had fun arranging playing cards in a row (cards need to be folded so that they can stand vertically) and pushing the first card to start a wave of falling playing cards. The first card when pushed, falls and topples the second card, the second will topple the third and so on till all the cards are toppled. If we have a series of propositions ;. P2, P3.--* , numbered by positive integers, assume that we can prove: (B) the first proposition P; is true (S) the truth of P; implies the truth of the next proposition Pa. Then we would have established the truth of all P;’s. Just as in case of playing cards. The truth of P,, by (S), implies truth of Pp which in turn implies truth of P; and so on. This method of proving a series of propositions is called the Method of Mathematical Induction (MMI). (B) is called the base of induction and (S) is called the inductive step. Let us look at a simple example: Example 1. For all n > 1, show that 34" — 1 is divisible by 80. ° Solution. Here Py is the statement: 3*" — 1 is divisible by 80. We will use MMI. We need to establish: (B) Py is true. ’ (S) If Py is true sv is Pas. Proof of (RB). P; : 3'—1 is divisible by 80. 3*-1 81-1 = 80,7a NON-ROUTINE PROBLEMS IN MATHEMATICS clearly divisible by 80. Proof of (S). Assume P,, is true, That is 3! — 1 = 80k for some integer k. We need to prove 344!) — 1 is also a multiple of 80. pet pe gleet_ 2 1.3" -1 81(3"" — 1) +81-1 81 - (80k)-+ 80 = 80(81k + 1) W by truth of Py. Thus 34+") — 1 is also a multiple of 80. Note that we have used the truth of P, in proving P.41. That is, whenever Pp is true so is Py41. Hence by MMI. Ph is true for alln > 1. Let. us look at a visual problem:MATHEMATICAL INDUCTION 75 Example 2. Given a 2" x 2" checker board with any one square deleted, show that it is possible to cover this board with L-shaped pieces. For example, a 4 x 4 checker board with a corner square removed can be covered by L-shaped pieces as in Fig. 1. Solution. Check the base case: n = 1. We have a 2 x 2 board with one square removed. Clearly, the remaining piece is L-shaped. So, for n = 1, the proposition is true (see Fig. 2). Assume that P, is true. That is any 2" x 2" checker board with one of the squares removed can be covered with L-shaped pieces. We need to establish Py) : any 2"+! x 2"+! checker board with a square removed can be covered with L-shaped pieces. Figure 3 Figure 4 Fig. 3 shows 2°+! x 2"+! checker board with one square removed. To establish truth of P41, we need to use the truth of Py. Fig. 4 shows the 2"+! x 2"+! board subdivided into four 2" x 2" quadrants. Let us remove an L-shaped piece in the centre. Now we have four 2" x 2" boards each with one6 NoN-ROUTINE-PROBLEMS IN MATHEMATICS square removed. By P,, each of these can be covered by L- shaped pieces. We put back the centre L-shaped piece and we are done. Remark. Some students tend to argue with areas of boards. A 4x 4 square with one piec and each L-shaped pi removed has 15 small squares ¢ has 3 small squares. Since 3 divides 15, it should be possible to cover the board. This argument is not correct since a 6 x 1 board has 6 squares and cannot be covered by L-shaped pieces. . Example 3. Find the sum of the first 1 odd.uatural numbers. Solution. Here, we are asked to find the sum 14+345+---+(2n-1) in closed form. Let us first check a few cases., 143 4 14+34+5 9 14+34+5+7 = 16. fl It appears that 1+3+5+---+(2n—1), the su of n odd numbers equals n?, (This is our conjecture.) We know that it is true for n = 1. Assume it is true for n. 143454---+(2n-1) =n? For n + 1, consider 143 4 54-04 (Qn-1)4+(2n41) (L434---4(2n-1))+(2n41) n+ In+1=(n+1)%, 1 ‘Thus our conjecture is also true for (n+ 1)..By (MMI) we have established:MATHEMATICAL INDUCTION 7 14+345+-+++(2n-1) =n? for allneN. Remarks. In Example 3, we tried to conjecture the general form from a few particular cases. This procedure can be a very difficult one. Even great mathematicians have erred in this step. Fermat, a notable seventeenth century French math- ematician conjectured that 22" +1 is a prime for all n > 1. He was guided by 27 +1 = 5, 2% 41 = 17, 2” +1 = 257 and 2 41 = 65537 being all primes. But in eighteenth century, L. Euler, an academician of Petersburg discovered that 2% 41 = 4 294 967 297 = 641 x 6 700 417 is a composite number. If we evaluate 991n? + 1 for n = 1,2,3,---, we do not get a perfect square for even first 100 numbers. Can we conjecture that 991n? + 1 is not a perfect square for all n and try MMI? We would be wrong because, 991n? +1 can be a square but the least n for which it is a square is very large n= 12 055 735 790 331 359 447 442 538 767. Let us look at a geometrical example. Consider n points on acircle. How many regions will the circle be divided into if each pair of points are joined? 3 points 4 regions 2 points ‘2 regions: Figure 6 ‘Figure 5 4 points 8 regions 5 points 16 regions Figure 7 Figure 878 NON-ROUTINE PROBLEMS IN MATHEMATICS We expect to find 32 regions for 6 points. Try it out and see! § 2. In the above we have seen the simplest form of induction. A very beautiful “back and forth” induction is illustrated in Cauchy's proof of Arithmetic Mean - Geometric Mean inequal- ity: Example 4. Let a1,a2,:-» ,@n be nonnegative real numbers. Let. A See, G = (ayap-++an)/" be their arithmetic mean and geometric meani respectively. Then show that A > G with equality if and only if ¢. = a2 = ne Solution. Cauchy's proof: Let P, be the proposition that if 1,-++ ,@q are non-negative real numbers, not all equal, then Q, +--+ Gn > n(ay Gn). We establish P, for all n as follows: (i) Pp holds (ii) If Py holds for any n > 3 then Py also holds (iii) If Py holds for any n > 2 then Pym also holds. By proving the above, we show how they establish P, for all n. When P; holds; by repeated application of (iii), we see that all of Pay Pay Pas, Psay + hold. All we have to do is to fill in the interspaces between any two consecutive pairs of this sequence. This is achieved by using (ii). For example, from the truth of Ps, using (ii) we get the truth of Pes and hence that of Pé2, and so on upto the