Scribd
Upload
129 views

COMP3121-Assignment 4 Solutions

The document provides solutions to 4 problems assigned as part of COMP3121/9101 Term 2, 2021 Assignment 4. The solutions involve using maximum flow algorithms to minimize the cost of disconnecting computers in a network, minimize the time to deliver goods from warehouses to shops, maximize the number of children that can complete a modified hopscotch game, and maximize the number of black rooks that can be placed on a chess board amidst white bishops.

Uploaded by

Ray Chen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
129 views

COMP3121-Assignment 4 Solutions

The document provides solutions to 4 problems assigned as part of COMP3121/9101 Term 2, 2021 Assignment 4. The solutions involve using maximum flow algorithms to minimize the cost of disconnecting computers in a network, minimize the time to deliver goods from warehouses to shops, maximize the number of children that can complete a modified hopscotch game, and maximize the number of black rooks that can be placed on a chess board amidst white bishops.

Uploaded by

Ray Chen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

COMP3121/9101 Term 2, 2021

Assignment 4 Solutions

Due date Friday, August 6, 2021 at 8 PM sharp.

You have four problems, marked out of a total of 100 marks.

1. There are N computers in a network, labelled {1, 2, 3, . . . , N }. There are M


one-directional links which connect pairs of computers. Computer 1 is trying
to send a virus to computer N . This can happen as long as there is a path
of links from computer 1 to computer N . To prevent this, you’ve decided to
remove some of the links from the network so that the two computers are
no longer connected. For each link, you’ve calculated the cost of removing
it. What is the minimum total cost to disconnect the computers as required,
and which edges should be removed to achieve this minimum cost? (25 pts)

Solution: Construct a flow network with computers as vertices and with


links between computers as edges of capacity set to the price of disconnect-
ing such a link. The source of the network is set to vertex 1 and the sink
at vertex N . Find a max flow in such a network and the corresponding min
cut. Disconnect the edges which cross the cut in the direction of the parti-
tion containing 1 to the partition containing N .

2. You have n warehouses and n shops. At each warehouse, a truck is loaded


with enough goods to supply one shop. There are m roads, each going from
a warehouse to a shop, and driving along the ith road takes di hours, where
di is an integer. Design a polynomial time algorithm to send the trucks to
the shops, minimising the time until all shops are supplied. (25 pts)

Hint: Combine a binary search with a max flow. By sorting you can assume
that di form an increasing sequence. Fix i and consider only roads which
take ≤ di hours to travel from a warehouse to the corresponding shop and
use max flow to see if they are enough to obtain a matching of warehouses
with shops which is of size n. Use a binary search on i to find the smallest
di which meets the requirements.

Solution: First sort time travel distances di in an increasing order; thus, we


can assume that di+1 ≥ di . Consider a value di for some i and construct a

1
bipartite graph Gi with warehouses wj as the left side of the partition and
with shops sj (1 ≤ j ≤ n). Connect all warehouses with all shops which are
within travel distance times di . Use max flow to see if such bipartite graph
has a perfect maximum matching of size n. Use a binary search to find
the smallest i such that graph Gi has a matching of size n, i.e., a matching
in which every warehouse has been matched with a shop, so that different
warehouses are assigned different shops.

3. A large group of children has assembled to play a modified version of hop-


scotch. The squares appear in a line, numbered from 0 to n, and a child is
successful if they start at square 0 and make a sequence of jumps to reach
square n. However, each child can only jump at most k < n squares at a
time, i.e., from square i they can reach squares i + 1, i + 2, . . . , i + k but
not i + k + 1, and a child cannot clear the entire game in one jump. An
additional rule of the game specifies an array A[1, . . . , n − 1], where A[i] is
the maximum number of children who can jump on square i. Assuming the
children co-operate, what is the largest number of children who can success-
fully complete the game?(25 pts)

Hint: Connect every square i with squares i + 1, . . . , i + k with a directed edge


of infinite capacity. Now limit the capacity of each square appropriately and
use max flow.
Solution: Construct a flow network with vertices 0, 1, 2, . . . , n such that for
all i > 0 vertex i corresponds to hopscotch square i. Connect the additional
entry vertex 0 with vertices corresponding to squares 1 to k and every vertex
corresponding to square i with squares i+1, . . . , i+k, all with edges of infinite
capacity. Set the capacity of each vertex 0 < i < n to A[i]. Set 0 vertex as
a source and vertex n as a sink and find max flow in such a network. The
number of kids which can reach square n is equal to the max flow in such a
network,

4. Use max flow algorithm to solve the following problem. You are given an
n×n chess board with k white bishops on the board at the given cells (ai , bi ),
(1 ≤ ai , bi ≤ n, 1 ≤ i ≤ k). You have to determine the largest number of
black rooks which you can place on the board so that no two rooks are in
the same row or in the same column or are under the attack of any of the k
bishops (recall that bishops go diagonally).(25 pts)
Solution: Construct a bipartite graph with n left side vertices representing
rows of the board and n right side vertices representing columns. Connect
with directed edges of capacity one every left side vertex i with all right side

2
vertices j which satisfy the property that the cell (i, j) of the board is not
under attack from one of the bishops. Add a super source and connect it
with all left vertices with edges of capacity one; add a super sink and connect
every right side vertex with such super sink with edges of capacity one. Find
max flow; if the flow equals n then the problem has a solution. The rooks
should be placed on cells (i, j) which correspond to edges with flow in them.

You might also like