COMPSCI330 Design and Analysis of Algorithms

Midterm Exam

Guidelines
• Describing Algorithms If you are asked to provide an algorithm, you should clearly define
each step of the procedure, and then analyze its overall running time. There is no need to
write pseudo-code; an unambiguous description of your algorithm in plain text will suffice.
If the running time of your algorithm is worse than the suggested running time, you might
receive partial credits.

• Timing Exam starts at 3:05 pm and ends at 4:20 pm.

Name:

Duke ID:

1
Problem 1 (Graph Algorithms). (25 points)
(a) Run Dijkstra’s algorithm on the following directed graph, starting from vertex 1. The
numbers in the circles are the indices for nodes, and the numbers near the arrows are the weights
of the edges.
9
6 5

14 2 1 10
9 6
1 7 4

7 10 11 2

2 3
15
(a.1) (5 points) Give the ordering in which the vertices are visited in Dijkstras algorithm. (When
there are multiple options, the algorithm chooses the vertex with smaller index.) Also list the final
shortest path distance for all vertices.
Ordering 1 2 3 4 5 6 7
Vertex 1
Distance 0

(a.2) (5 points) After vertex 6 is visited by the algorithm, what are the distance values for all
vertices (if a vertex cannot yet be reached, the distance is +∞?)
Vertex 1 2 3 4 5 6 7
Distance 0

2
 (b) Run DFS on the following graph, starting from vertex 0. When you have the option of
choosing multiple vertices, always choose the vertex with smaller index first.
0 2

1 5 3

6 4
(b.1)(5 points) Draw the DFS tree. (Feel free to use the vertices we provided below, if you need
to change your solution, please point clearly where your solution is.)

0 2

1 5 3

6 4

(b.2)(5 points) List the pre-order and post-order.

(b.3) (5 points) For edges not on the DFS tree, classify them as forward edge, backward edge
or crossing edge.

3
Problem 2 (Graph Examples). (25 points)
(a) (10 points) Construct an undirected graph with 7 vertices and a starting vertex 1, such that
one of its BFS tree looks like the figure below, and the graph has as many edges as possible. You
should draw the graph with all the edges, label the vertices as in the given figure, and list the BFS
order of the graph (you need to list the specific BFS order that would lead to the given BFS tree).
(You do not need to prove that the graph you draw has the maximum number of edges possible.)
2 3 4

5 6 7

Feel free to use these vertices. If you prefer to draw the entire graph, please clearly indicate
where it is.

2 3 4

5 6 7

4
(b) (15 points) Construct a weighted directed graph with 5 vertices (with no negative cycles), such
that when running Bellman-Ford algorithm starting from vertex 1, the distance to vertex 5 changes
4 times. Draw the graph, and list the values of d[5, i] for i = 0, 1, 2, 3, 4, 5.

5
Problem 3 (Electricity Supply). (25 points) There are k communities that have n houses in total.
In this problem, our goal is to connect all of these houses to an power plant. We are given a
graph (see figure below) with n + 1 vertices representing the power plant (vertex 0) and the houses
(remaining n vertices). There is an undirected edge (u, v) with weight w(u, v) > 0 if we can build
a connection between vertex i and vertex j. A house has electricity as long as it is connected with
the power plant (by a path). We also know which houses belong to the same community.

5
h2 h5

5 16 3 5

h1 h3 h4 h6
7
4
4 7

plant

11 3
2
h7 h8
1
6 4

h9 h10
13

To prevent loss of power, each community also has its own generator (located in one of the
houses, which house it is does not matter for this problem). Therefore, when we are building the
connections, we want to make sure that for every community, even if all the connections with outside
(either with the power plant or with houses in another community) fail, all the houses within the
community are still connected. The thicker lines in the figure above illustrate the optimal solution
satisfying this requirement.
Design an algorithm that finds the minimum cost way of connecting all the houses to the power
plant, while satisfying the constraint in the above paragraph. You can assume there is always a
way to connect the houses while satisfying the constraint.

6
(a) (10 points) Describe your algorithm (it should not be slower than O(m log n), where m is the
number of edges in the graph).

7
(b) (5 points) Analyze the running time of your algorithm.

(c) (10 points) Prove the correctness of your algorithm.

8
(This page is left blank intentionally.)

9
Problem 4 (Shipping Products). (25 points) Company X is trying to figure out how to ship their
product from factories to shops. There are n1 factories and n2 shops. Factory Fi (i = 1, 2, ..., n1 )
can produce at most ui products at the cost of ci per unit. Shop Sj (j = 1, 2, ..., n2 ) can sell at
most vj products at the price of pj per unit. However, transportation is only available for a limited
pairs of factories and shops - you are given a list of possible transportation routes E = {(i1 , j1 ),
(i2 , j2 ), ..., (im , jm )}. Each route (i, j) ∈ E has a cost of wi,j per unit, where transporting each
product from factory Fi to shop Sj costs wi,j . See Figure below for an example. In this example,
the optimal solution would produce 3 products at each of F1 and F2 , and transport them to S1 and
S2 correspondingly. The total profit is 24.
Note: All of ui , ci , vj , pj , wi,j are constants given to you as input, these are not the variables.
You need to define your own variables. Your LP needs to work for all instances of the problem, not
just the example given.
w1,1 = 2
u1 = 5, c1 = 3 F1 S1 v1 = 3, p1 = 8
w1
,2 =
8
w2,2 = 1
u2 = 3, c2 = 4 F2 S2 v2 = 6, p2 = 10

(a) (10 points) Write a linear program to compute the maximum profit that the company can make.
Profit is equal to the total selling price, minus the total production cost and total transportation
cost.

10
(b) (10 points) Suppose you can overload the factories so that factory Fi can produce at most u0i
extra products, at the additional cost of c0i > ci (that is, the first ui products still cost ci each, but
the next u0i products will cost c0i each). Modify your LP so that it computes the maximum profit
in this setting. (For the example before, you can think about u01 = 1, c01 = 4, and u02 = 4, c02 = 7.
In this case the optimal solution would produce 3 products at F1 and 6 products at F2 . The total
profit for this solution is 30.)

11
(c) (5 points) If in the setting of part (b), we did not have the assumption that c0i > ci , does your
LP in part (b) still work? If not, give a feasible solution to your LP that does not correspond to
an actual feasible scenario in the problem. (You will only get points for this part if your solution
to (b) is at least close to correct.)

12