OCR MEI D2 (Decision Mathematics 2) 2011 June

1

1. Heard in Parliament: "Will the minister not now discontinue her proposal to ban the protest?"
   The minister replied "Yes I will."
   To what had the minister committed herself logically, and why might that not have been her intention?
2. In a cricket tournament an umpire might be required to decide whether or not a batsman is out 'lbw', ie 'leg before wicket'. The lbw law for the tournament refers to parts of the cricket pitch as shown in the diagram (assuming a right-handed batsman):
   \includegraphics[max width=\textwidth, alt={}, center]{52b8153f-e655-4852-a0f8-6f1c1e9c9170-2_254_1045_717_507} The umpire has to make a number of judgements:
   A Would the ball have hit the wicket?
   B Did the ball hit the batsman, or part of his equipment other than the bat, without hitting the bat?
   C Did the ball hit the batsman, or part of his equipment other than the bat, before hitting the bat?
   D Was the part of the batsman or his equipment which was hit by the ball, between the wickets when it was hit? E Was the part of the batsman or his equipment which was hit by the ball, outside of the wicket on the off side when it was hit? F Was the batsman attempting to play a stroke?
   The law can be interpreted as saying that the batsman is out lbw if \([ ( \mathrm { A } \wedge \mathrm { B } ) \vee ( \mathrm { A } \wedge \mathrm { C } ) ] \wedge [ \mathrm { D } \vee ( \mathrm { E } \wedge \sim \mathrm { F } ) ]\).
   The tournament's umpiring manual, in attempting to simplify the law, states that the batsman is out lbw if \(\mathrm { A } \wedge ( \mathrm { B } \vee \mathrm { C } ) \wedge ( \mathrm { D } \vee \mathrm { E } ) \wedge ( \mathrm { D } \vee \sim \mathrm { F } )\). For an lbw decision this requires 4 conditions each to be true.
   1. Use the rules of Boolean algebra to show that the manual's rule is logically equivalent to the law as stated above, naming the rules used at each step. A trainee umpire, using the manual, considers each condition in turn and judges that the following are true: A; B; E; D.
   2. What is her decision and why?
   3. What is odd about her judgement, and does this make the logic invalid?

2 A government has just created a new ministry, the Ministry of Administrative Affairs. The ministry is to have four departments:
the Administration
the Bureaucracy
the Certification Service
the Duplication Section.
Each of these departments is to be established in a separate office on one of four existing sites. The diagram shows the direct journey times in minutes between these four sites.
\includegraphics[max width=\textwidth, alt={}, center]{52b8153f-e655-4852-a0f8-6f1c1e9c9170-3_342_403_721_829}

1. Use Floyd's algorithm to find the shortest journey times between the four office sites.
2. Draw a network showing your shortest times.
3. Use appropriate algorithms to find upper and lower bounds for the optimum solution to the Travelling Salesperson Problem in the original network, briefly explaining the steps taken.
4. A van is to be organised to deliver bundles of paperwork between the departments. Why might the optimum solution to the TSP be useful in planning this, and why might it not be?
5. Journeys to locations 2 and 3, from locations 1 and 4, are subject to a congestion charge which is equivalent in costing terms to 15 minutes of journey time. What sort of network would be needed to model this?

3 Magnus has been researching career possibilities. He has just completed his GCSEs, and could leave school and get a good job. He estimates, discounted at today's values and given a 49 year working life, that there is a \(50 \%\) chance of such a job giving him lifetime earnings of \(\pounds 1.5 \mathrm {~m}\), a \(30 \%\) chance of \(\pounds 1.75 \mathrm {~m}\), and a \(20 \%\) chance of \(\pounds 2 \mathrm {~m}\). Alternatively Magnus can stay on at school and take A levels. He estimates that, if he does so, there is a 75\% chance that he will achieve good results. If he does not achieve good results then he will still be able to take the same job as earlier, but he will have lost two years of his lifetime earnings. This will give a \(50 \%\) chance of lifetime earnings of \(\pounds 1.42 \mathrm {~m}\), a \(30 \%\) chance of \(\pounds 1.67 \mathrm {~m}\) and a \(20 \%\) chance of \(\pounds 1.92 \mathrm {~m}\). If Magnus achieves good A level results then he could take a better job, which should give him discounted lifetime earnings of \(\pounds 1.6 \mathrm {~m}\) with \(50 \%\) probability or \(\pounds 2 \mathrm {~m}\) with \(50 \%\) probability. Alternatively he could go to university. This would cost Magnus another 3 years of lifetime earnings and would not guarantee him a well-paid career, since graduates sometimes choose to follow less well-paid vocations. His research shows him that graduates can expect discounted lifetime earnings of \(\pounds 1 \mathrm {~m}\) with \(20 \%\) probability, \(\pounds 1.5 \mathrm {~m}\) with \(30 \%\) probability, \(\pounds 2 \mathrm {~m}\) with \(30 \%\) probability, and \(\pounds 3 \mathrm {~m}\) with \(20 \%\) probability.

1. Draw up a decision tree showing Magnus's options.
2. Using the EMV criterion, find Magnus's best course of action, and give its value. Magnus has read that money isn't everything, and that one way to reflect this is to use a utility function and then compare expected utilities. He decides to investigate the outcome of using a function in which utility is defined to be the square root of value.
3. Using the expected utility criterion, find Magnus's best course of action, and give its utility.
4. The possibility of high earnings ( \(\pounds 3 \mathrm {~m}\) ) swings Magnus's decision towards a university education. Find what value instead of \(\pounds 3 \mathrm {~m}\) would make him indifferent to choosing a university education under the EMV criterion. (Do not change the probabilities.)

4 A small alpine hotel is planned. Permission has been obtained for no more than 60 beds, and these can be accommodated in rooms containing one, two or four beds. The total floor areas needed are \(15 \mathrm {~m} ^ { 2 }\) for a one-bed room, \(25 \mathrm {~m} ^ { 2 }\) for a two-bed room and \(40 \mathrm {~m} ^ { 2 }\) for a four-bed room. The total floor area of the bedrooms must not exceed \(700 \mathrm {~m} ^ { 2 }\). Marginal profit contributions per annum, in thousands of euros, are estimated to be 5 for a one-bed room, 9 for a two-bed room and 15 for a four-bed room.

1. Formulate a linear programming problem to find the mix of rooms which will maximise the profit contribution within the two constraints.
2. Use the simplex algorithm to solve the problem, and interpret your solution. It is decided that, for marketing reasons, at least 5 one-bed rooms must be provided.
3. Solve this modified problem using either the two-stage simplex method or the big-M method. You may wish to adapt your final tableau from part (ii) to produce an initial tableau, but you are not required to do so.
4. The simplex solution to the revised problem is to provide 5 one-bed rooms, 15 two-bed rooms and 6.25 four-bed rooms, giving a profit contribution of \(€ 253750\). Interpret this solution in terms of the real world problem.
5. Compare the following solution to your answer to part (iv): 8 one-bed rooms, 12 two-bed rooms and 7 four-bed rooms. Explain your findings.