Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Introductory Worksheet 1. [Source: BMO1 β 2012] Isaac places some counters onto the squares of an 8 by 8 chessboard so that there is at most one counter in each of the 64 squares. Determine, with justification, the maximum number that he can place without having five or more counters in the same row, or in the same column, or on either of the two long diagonals. 2. [Source: BMO1] Determine the least natural number π for which the following result holds: No matter how the elements of the set {1, 2, β¦ , π} are coloured red or blue, there are integers π₯, π¦, π§, π€ in the set (not necessarily distinct) of the same colour such that π₯ + π¦ + π§ = π€. 3. [Source: Cayley] Four types of rectangular tile have sizes 300 mm × 300 mm, 300 mm × 600mm, 600 mm × 600 mm and 600 mm × 900 mm. Equal numbers of each type of tile are used, without overlaps, to make a square. What is the smallest square that can be made? Extra induction questions (if youβve seen the βEpilogueβ slides) 4. Prove that 11π β 6π is divisible by 5, βπ β β. 5. [Source: Frosty Special] Prove that 6π + 9π is divisible by 5 if and only if π is odd. Last updated: 12th September 2014 www.drfrostmaths.com/rzc Introductory Worksheet - ANSWERS Note about solutions in the worksheets: For problems that Iβve obtained from Olympiad or Maths Challenge papers, I tend to work out the question myself before I look at the model answers (although sometimes these werenβt available, so I can only hope that what Iβve provided is correct!). The solutions provided are a mix of my own and the βmodelβ solutions. Sometimes my own solutions uses a different approach from the one suggested β itβs indicative that thereβs often multiple approaches for many questions. 1. Isaac places some counters onto the squares of an 8 by 8 chessboard so that there is at most one counter in each of the 64 squares. Determine, with justification, the maximum number that he can place without having five or more counters in the same row, or in the same column, or on either of the two long diagonals. The maximum must be at most 32 counters, because we could place 4 counters in each row at most, and there are 8 columns. All that remains to see if itβs possible to have a valid arrangement of 32 counters. One possible such arrangement is shown below: It was slight luck that my βintelligent guessβ happened to be the right one. Notice that as the going across, the pattern repeats every 3 squares horizontally. The consequence is that because 3 and 8 share no common factors (weβll see soon that we call them βcoprimeβ), the column the two consecutive vertical counters will be in will be different each time (i.e. first the 4th column, then the 7th, then the 2nd and so on). This is good because we minimise column usage, and avoid getting to the 4-counter column limit too soon. 2. Determine the least natural number π for which the following result holds: No matter how the elements of the set {π, π, β¦ , π} are coloured red or blue, there are integers π, π, π, π in the set (not necessarily distinct) of the same colour such that π + π + π = π. The best way to get your head around this question is to consider simpler cases when π is small. Since also the question says βfor any colouringβ, we want to consider the βworst case scenarioβ, i.e. try to gradually colour in such a way that we avoid a colouring that where we have some numbers of the same colour satisfying π₯ + π¦ + π§ = π€. For example, suppose π = 4. Let the 1 be coloured red say. Weβd have to colour 3 blue, www.drfrostmaths.com/rzc because if it was red, then weβd have 1 + 1 + 1 = 3, so regardless of what the other numbers were coloured, weβd have π₯, π¦, π§, π€ available amongst the reds. We can easily break the condition by say colouring 1 red, 2 blue, 3 blue and 4 blue. This is because the condition π₯ + π¦ + π§ = π€ does not hold with the one number coloured red (1 + 1 + 1 = 3, but 3 is not a red number), and with the blue numbers, the minimum possible sum of three numbers is 2 + 2 + 2 = 6, so any sum of three blue numbers will give a number that is not a blue number. I tried some higher values of π next to see if I could spot a pattern. Suppose π = 8, and 1 was coloured red. 3 would have to be blue. Suppose 2 was also coloured red. Then we could combine these in sums to get 1 + 1 + 1 = 3, 1 + 1 + 2 = 4, up to 2 + 2 + 2 = 6. Thus we would have to colour 3, 4, 5 and 6 blue otherwise weβd have a π₯, π¦, π§, π€ available amongst the red numbers. But at this point, if we coloured 7 and 8 blue, then there canβt be an blue numbers that satisfy π₯ + π¦ + π§ = π€, because the minimum sum with blue is 3 + 3 + 3 = 9, and so we couldnβt have a blue number for π€. We will encounter a similar problem when π = 9 and π = 10. If 1 and 2 are coloured red, 38 coloured blue (colouring 3-8 blue avoids being able to have any satisfying π₯, π¦, π§, π€ amongst the blue, as the minimum sum of three blue numbers is 9, which is not blue) and 9 and 10 red (if there is a 10th), then we know thereβs no satisfying π₯, π¦, π§, π€ for the blue, and there canβt be any for the red either: if we sum both low and high red numbers, the minimum possible sum is 1 + 1 + 9 = 11, but 11 is not available since it exceeds π. Thus all that remains is to show that for π = 11 we can find π₯, π¦, π§, π€ for any possible colouring. As before, without loss of generality1 assume the first number, 1, is red. Now we have two cases to consider: a) The second number, 2, is the same colour, i.e. also red. Then as before, weβd have to colour 3, 4, 5, 6 blue to avoid there being a π₯, π¦, π§, π€ being available. If we coloured any of 9, 10 or 11 blue, weβd have a π₯, π¦, π§, π€ available because using the blue 3 to 6, we could get any sum between 9 and 11. So letβs colour them red. But now weβre guaranteed a π₯, π¦, π§, π€, because 1 + 1 + 9 = 11, where all numbers are red. So weβve managed to show that itβs impossible to avoid having a colouring where thereβs some π₯, π¦, π§, π€ available for one of the two colours. Symmetrically, we have the same logic when the first two numbers are blue instead of red. b) The second number, 2 is a different colour, i.e. blue. Letβs make 3 blue, because if it was red than we would have found a π₯, π¦, π§, π€. This means that if any of 6-9 are blue, weβd have a π₯, π¦, π§, π€ because these are all the sums we can get using 2 and 3. So consider them as red. But then weβre guaranteed a π₯, π¦, π§, π€ using red numbers, because 1 + 1 + 6 = 8 for example. So again, weβve shown that however we colour the numbers, we canβt avoid having a colouring where thereβs some π₯, π¦, π§, π€ available. 1 βWithout loss of generalityβ means weβre considering a specific case, but that doing so doesnβt make our proof any less general. If weβd instead let the 1 be blue, then we could have used the same logic thereafter but with the colours switched. www.drfrostmaths.com/rzc 3. Four types of rectangular tile have sizes 300 mm × 300 mm, 300 mm × 600mm, 600 mm × 600 mm and 600 mm × 900 mm. Equal numbers of each type of tile are used, without overlaps, to make a square. What is the smallest square that can be made? (Official solution) Let us say that one unit is 300 mm, so that the permitted tiles are 1 × 1, 1 × 2, 2 × 2 and 2 × 3. Since the sides of all the tiles have lengths that are a whole number of units, any square made out of them will have sides of length π that is a whole number of units. This square has area π 2 . Also, the total area of one tile of each type is 1 + 2 + 4 + 6 = 13, so is a multiple of 13. The smallest such π is 13 itself, so we ask ourselves if such a tiling is possible with 13 copies of each tile. It can indeed be done, as the example in the figure shows. This square, which we have proved to be the smallest possible, measures 13 units on each side, or 3.9 m × 3.9 m. [JAF note: Many Year 9 Tiffin students got 6 out of 10 for this question, because although they proved that the smallest square βmight beβ 13 of each tile, they didnβt prove that such a square was possible with this number of tiles)] 4. Prove that πππ β ππ is divisible by 5 βπ β β. Base Case: Show itβs true for π = 1: 11 β 6 = 5, so true. Inductive Step: Assume itβs true for π = π. Suppose now π = π + 1. 11π+1 β 6π+1 = 11 β 11π β 6 β 6π = 6(11π β 6π ) + 5 β 11π The first term in the sum is divisible by 5 as we assumed 11π β 6π was. And the second term is clearly divisible by 5. 5. Prove that ππ + ππ is divisible by 5 if and only if π is odd. This question is all about being rigorous with your proof by ensuring itβs βcompleteβ. Note that we can deal with the βif and only ifβ by showing that 6π + 9π is divisible by 5 if π is odd, and also (very importantly!) not if π is even. This is because βif is π is even, 6π + 9π is not divisible by 5β is equivalent to βif 6π + 9π is divisible by 5, π is oddβ. Weβll see more of negating/reversing conditional statements like this in the Logic module. Proof Part 1: If π is odd, ππ + ππ is divisible by 5. If π is odd, represent it as π = 2π + 1. Our problem becomes equivalent to showing 62π+1 + 92π+1 is divisible by 5 for all positive integers π. www.drfrostmaths.com/rzc Base Case: When π = 1, we have 216 + 729 = 945 which is divisible by 5. Inductive Step: Assume true for π = π. Now show true for π = π + 1. 62(π+1)+1 + 92(π+1)+1 = 62π+3 + 92π+3 = 36 β 62π+1 + 81 β 92π+1 = 36(62π+1 + 92π+1 ) + 45 β 92π+1 The first term is divisible by 5 because we assumed 62π+1 + 92π+1 was. And the second one is because 45 is. Proof Part 2: If π is even, ππ + ππ is not divisible by 5. Note that a counterexample is not sufficient: we need to show that for any even number, the statement does not hold. If π is even, then let π = 2π. Thus we need to prove that 62π + 92π = 36π + 81π is never divisible by 5 for any integer π. Base Case: When π = 1, 36 + 81 = 117 which is not divisible by 5. Inductive Step: Assume itβs not divisible for π, now show itβs still not divisible for π + 1. 36π+1 + 81π+1 = 36 β 36π + 81 β 81π = 36(36π + 81π ) + 45(81π ) Now, the first term is not divisible by 5 because neither 36 nor 36π + 81π is. The second term clearly is divisible by 5 because 45 is. And clearly a number not divisible by 5 plus a number divisible by 5 will give a number not divisible by 5. Youβll see this problem again in one of the Number Theory module worksheets. The complete proof can be reduced to just two lines using modular arithmetic! www.drfrostmaths.com/rzc