Problems:

1. Subproblem definition

• Describe the meaning of subproblem in words, in terms of parameters x

∈X

• Often subsets of input: prefixes, suffixes, contiguous substrings of a

sequence

• Often multiply possible subsets across multiple inputs

• Often “remember” information by maintaining some auxiliary state

(expansion)

• Often expand based on state of integers in problem space

(pseudopolynomial?)

2. Relate recursively

• Relate subproblem solutions recursively x(i) = f(x(j), . . .) for one or more j

<

• Identify question about subproblem solution whose answer would let

you reduce to smaller subproblem(s), then locally brute-force all possible
answers to the question

3. Topological order

• Argue relation is acyclic (defines “smaller” subproblems), subproblems

form a DAG
4. Base cases

© 2022 All Rights Reserved . Programming Exam Help

• State solutions for all (reachable) independent subproblems where
relation breaks down 5. Original problem

• Show how to compute solution to original problem from solutions to

subproblem(s)

• Possibly use parent pointers to recover actual solution, not just objective
function

6. Time analysis P

• work(x), or if work(x) = O(W) for all x ∈ X, then |X| · O(W) x∈X

• work(x) measures nonrecursive work in relation; treat recursions as taking

O(1) time

Problem 1. Future Investing (adapted from S18 PS9)

Tiffany Bannen stumbles upon a lottery chart dropped by a time traveler

from the future, which lists winning lottery numbers and positive integer
cash payouts for the next n days. Tiffany wants to use this information to
make money, but is worried if she plays winning numbers every day, lottery
organizers will get suspicious. As such, she decides to play the lottery
infrequently: at most twice in any seven day period. Describe a O(n)-time
algorithm to determine the maximum amount of lottery winnings Tiff can
win in the next n days by playing the lottery infrequently.

© 2022 All Rights Reserved . Programming Exam Help

Problem 2. Oh Charlie, Where Art Thou? (adapted from S18 PS9)

A wealthy family, Alice, Bob, and their young son Charlie are sailing around
the world when they encounter a massive storm. Charlie is thrown
overboard, presumed drowned. Twenty years later, a man comes to Alice
and Bob claiming to be Charlie. Alice and Bob are excited, but skeptical.
Alice and Bob order a DNA matching test from the genetic testing company
46AndThee. Given three length-n DNA sequences from Alice, Bob, and
Charlie, the testing center will determine ancestry as follows: if Charlie’s
DNA can be partitioned into two (not necessarily contiguous)
subsequences of equal length, where one is a subsequence of Alice’s DNA,
and the other is a subsequence of Bob’s DNA, the Charlie is their son. For
example, suppose Alice’s DNA is AATT and Bob’s DNA is CCGG. If Charlie’s
DNA were CATG, he would be matched as their son, since CATG can be
partitioned into disjoint subsequences CG and AT which are subsequences
of Alice and Bob’s DNA respectively. However, Charlie would be found to be
an imposter if his DNA were AGTC. Describe an O(n4)-time algorithm to
determine whether Charlie is a fraud.

Problem 3. Sweet Tapas (adapted from S18 Final)

Obert Ratkins is having dinner at an upscale tapas bar, where he will order
many small plates. There are n plates of food on the menu, where
information for plate i is given by a triple of nonnegative integers (vi, ci, si):
the plate’s volume vi, calories ci, and sweetness si ∈ {0, 1} (the plate is
sweet if si = 1 and not sweet if si = 0). Obert is on a diet: he wants to eat no
more than k calories during his meal, but wants to fill his stomach as much
as possible. He also wants to order exactly s < n sweet plates, without
purchasing the same dish twice. Describe an O(nks)-time algorithm to find
the maximum volume of food Obert can eat given his diet.

© 2022 All Rights Reserved . Programming Exam Help

Problem 4. Gokemon Po (adapted from S18 Final) Kash Etchum wants to
play Gokemon Po, a new augmented reality game. The goal is to catch a
set of n monsters who reside at specific locations in her town. Monsters
must be obtained in a specific order: before Kash can obtain monster mi,
she must have already obtained all monsters mj for j < i. To obtain
monster mi, Kash may either purchase the monster in-game for positive
integer ci dollars, or she may catch mi for free from that monster’s
location. If Kash is not at the monster’s location, she will have to pay a
ride share service to drive her there. The minimum possible cost to
transport from the location of monster mi to the location of monster mj
via ride sharing is denoted by the positive integer s(i, j). Given the price
lists of in-game purchases and ride sharing trips between all pairs of
monsters, describe an O(n2)-time algorithm to determine the minimum
amount of money Kash must spend in order to catch all n monsters,
assuming that she starts at the location of monster m1.

© 2022 All Rights Reserved . Programming Exam Help

 Solutions:
Solution: 1.

Subproblems
• Let L(i) be the cash payout of playing the lottery on day i ∈ {1, . . . , n}

• Need to keep track of most recent two plays (or equivalently, restrictions
on future plays)

• x(i, j): maximum lottery winnings playing on suffix of days from i to n,

assuming play on day i and next allowable play is on day i + j

• for i ∈ {1, . . . , n} and j ∈ {1, . . . , 6} 2. Relate

• Tiffany will play again on some day in future. Guess!

• It is never optimal to go 11 days without playing the lottery, as playing on

the 6th day would be valid and strictly increase winnings

• The next play can be on day i + k for k ∈ {j, . . . , 11}

• If next play on i + k for k ∈ {1, . . . , 6}, next allowable play is on day i + 7

• If next play on i + k for k ∈ {7, . . . , 11}, next allowable play is on day i + k

+1

• x(i, j) = L(i) + max{x(i + k, max{1, 7 k}) | k ∈ {i, . . . , 11} and i + k ≤ n} 3.

© 2022 All Rights Reserved . Programming Exam Help

3. Topo

• Subproblems x(i, j) only depend on strictly larger i, so acyclic 4.

Base

• x(n, j) = L(i) for all j ∈ {1, . . . , 6} 5. Original

• Solve subproblems via recursive top down or iterative bottom up

• Guess first play (within first seven days)

• Solution to original problem is max{x(i, 1) | i ∈ {1, . . . , 7}}

• (Can store parent pointers to reconstruct days played) 6. Time

• # subproblems: 6n

• work per subproblem: O(1)

• work for original: O(1)

• O(n) running time

© 2022 All Rights Reserved . Programming Exam Help

Solution 2:

1. Subproblems

• Let A, B, and C be the relevant length-n DNA sequences from Alice, Bob,
and Charlie.

• Want to match some characters of A and B to all characters of C

• x(i, j, ki, kj ): true if can match a length-ki subsequence of suffix A[i :] and
a length-kj characters from prefix B[j :] to all characters in suffix C[(n ki
kj ) :] (the suffix containing the last ki + kj characters), and false otherwise.

• for i, j ∈ {0, . . . , n} and ki, kj ∈ {0, . . . , n/2} (assume n is even) 2. Relate

• Must match character C[i]; if A[i] = C[i] or B[i] = C[i] recurse on remainder

• Alternatively, do not use either A[i] or B[i]

© 2022 All Rights Reserved . Programming Exam Help

3. Topo

• Subproblem x(i, j, ki, kj ) only depends on strictly smaller i + j, so acyclic

4. Base

• x(n, n, 0, 0) is true (all matched!)

• x(n, j, ki, kj ) false if ki > 0 (no more characters in A)

• x(i, n, ki, kj ) false if kj > 0 (no more characters in B)

5. Original

• Solve subproblems via recursive top down or iterative bottom up

• Solution to original problem is x(n, n, n/2, n/2) 4

6. Time

• # subproblems: O(n4)

• work per subproblem: O(1)

• O(n4) running time

© 2022 All Rights Reserved . Programming Exam Help

Solution: 3. Subproblems

• x(i, j, s0 ): maximum volume of food possible purchasing a suffix of

plates pi to pn1, using at most j calories, ordering exactly s0 sweet plates

• for i ∈ {0, . . . , n}, j ∈ {0, . . . , k}, s0 ∈ {0, . . . , s} 2. Relate

• Either order plate pi or not. Guess!

• If order pi, get vi in volume but use ci calories

• If pi is sweet, need to order one fewer sweet plate

3. Topo

• Subproblems x(i, j, s0 ) only depend on strictly larger i, so acyclic

4. Base

• x(n, j, 0) = 0 and any j (no more plates to eat)

© 2022 All Rights Reserved . Programming Exam Help

• x(n, j, s0 ) = ∞ for s0 > 0 and any j (no more plates, but still need to eat
sweet)

5. Original

• Solution given by x(0, k, s) 6. Time

• # subproblems: O(nks)

• work per subproblem: O(1)

• O(nks) running time

Solution 4:

1. Subproblems

• x(i, j): min cost of catching monsters mi to mn starting at location mj for

j ≤ i 2. Relate

• If at location of monster, catch it for free!

• Otherwise, either acquire monster mi by purchasing or ride-sharing to

location. Guess!

• If purchase spend ci dollars, else need to ride share to mi from mj

© 2022 All Rights Reserved . Programming Exam Help

3. Topo

• Subproblems x(i, j) only depend on strictly larger i + j so acyclic

4. Base

• x(n + 1, j) = 0 for any j (no cost to collect no monsters)

5. Original

• Solution is given by x(1, 1)

6. Time

• # subproblems: O(n2)

• work per subproblem: O(1)

• O(n2) running time

© 2022 All Rights Reserved . Programming Exam Help