Exercise 6:

Question 1:

The Floyd-Warshall algorithm is an efficient method for finding the shortest paths between all
pairs of vertices in a weighted graph. It works by iteratively improving the estimate of the
shortest path between two vertices, considering all possible intermediate vertices.
The algorithm initializes a distance matrix D with the following values:
 D[i][i] = 0 for all vertices i (distance to self is zero)
 D[i][j] = weight of edge (i,j) if it exists
 D[i][j] = ∞ (infinity) if there's no direct edge between i and j
This process systematically considers each vertex k as a potential intermediate point on a path
from i to j. If going through k results in a shorter path than the current best known path from i to
j, the distance is updated.
The algorithm's time complexity is Θ(n³), where n is the number of vertices9. Despite this cubic
complexity, Floyd-Warshall is often preferred for its simplicity and ability to handle negative
edge weights (though not negative cycles).
One of the key advantages of Floyd-Warshall is that it can find shortest paths between all pairs
of vertices in a single execution, making it particularly useful for dense graphs or when multiple
shortest paths are needed.

Question 2:

The Floyd-Warshall algorithm can be adapted to calculate the transitive closure of a relation by
modifying its initialization and update steps. Here's how it works:
1. Initialization:
 Set R[i][j] = 1 if there's a direct edge from i to j in the original relation
 Set R[i][j] = 0 if there's no direct edge from i to j
 Set R[i][i] = 1 for all vertices i (reflexive property)1
2. Main algorithm:
The core of the algorithm remains similar to the shortest path version, but with boolean
operations:
This step checks if there's a path from i to j either directly or through an intermediate vertex k.
3. Result:
After the algorithm completes, the matrix R represents the transitive closure:
 R[i][j] = 1 if there's a path from i to j in the transitive closure
 R[i][j] = 0 if there's no path from i to j
The key difference from the shortest path version is that instead of finding minimum distances,
we're establishing the existence of paths. The logical OR operation (|) replaces the minimum
comparison, and the logical AND operation (&) replaces addition.
This adaptation effectively transforms the Floyd-Warshall algorithm into Warshall's algorithm
for transitive closure, maintaining the same O(n³) time complexity.

Question 3:

A real-life problem that can be modeled using a graph is airline route networks. In an airline
network, airports are represented as vertices, and direct flights between them are represented as
edges. An important centrality measure that is not commonly discussed in basic network theory
courses is Effective Resistance Centrality (also called Kirchhoff Index Centrality).
Effective Resistance Centrality
This measure is based on electrical network theory and is calculated using the concept of
effective resistance in a resistor network. It considers how easily a node (airport) can be reached
from all other nodes in the network. Nodes with low effective resistance centrality are well-
connected in a way that ensures multiple efficient paths exist to reach them, making them more
reliable in case of disruptions.
Interpretation of Effective Resistance Centrality in Airline Networks
 Airports with low effective resistance centrality act as robust hubs, meaning that even if
some flights are canceled, passengers can still efficiently reach their destinations through
alternative routes.
 Airports with high effective resistance centrality are more vulnerable to isolation,
meaning disruptions at these airports can cause major travel delays.
Example with 6 Airports
Consider a small airline network with the following airports and direct flight connections:
 A connects to B, C
 B connects to A, D, E
 C connects to A, D
 D connects to B, C, E, F
 E connects to B, D, F
 F connects to D, E
Using Effective Resistance Centrality, airport D has the lowest resistance, meaning it is the best
hub for connecting flights. If an airline needs to prioritize investments in infrastructure, D should
be a focus since it provides strong connectivity. However, airport A has higher resistance,
meaning disruptions there can cause more significant delays.
Sources
1. Klein, D. J., & Randić, M. (1993). Resistance distance. Journal of Mathematical
Chemistry, 12(1), 81-95.
2. Wu, J., Barahona, M., Tan, Y., & Deng, H. (2018). Spectral measure of structural
robustness in complex networks. IEEE Transactions on Network Science and
Engineering, 5(3), 205-216.
3. Chat-GPT

Question 4:

(a) The reasoning for this part is logically sound. The explanation uses induction to prove that
σ(s, v) correctly represents the number of shortest paths from s to v when v is extracted from the
queue. Here's why it's correct:
1. Base case: For the initial vertex s, σ(s,s) = 1, which is correct as there's only one path
from a vertex to itself.
2. Inductive step: Assuming correctness for all vertices with distance less than d from s,
when a vertex v at distance d is dequeued, its σ(s,v) value is calculated as the sum of
σ(s,u) for all predecessors u of v. This is correct because:
 All predecessors u are at distance d-1 and have been processed (due to BFS
order).
 Each shortest path to v must go through one of its predecessors.
 The total number of paths to v is the sum of paths to all its predecessors.
The inductive reasoning effectively proves the correctness of σ(s,v) for all vertices v.
(b) The explanation for this part is also logically sound, though it correctly points out a reversal
of symbols v and u in the question compared to the answer. The reasoning is as follows:
 σ(s,u) represents the total number of shortest paths from s to u.
 If v is a predecessor of u, then σ(s,u) includes all paths that go through v, plus potentially
other paths.
 Therefore, σ(s,u) must be greater than or equal to σ(s,v).
This logic aligns with the nature of the Brandes algorithm, where the number of shortest paths to
a vertex is always at least as large as the number of paths to any of its predecessors.
In conclusion, both parts (a) and (b) demonstrate sound logical reasoning in explaining key
aspects of the Brandes algorithm's correctness.

Question 5:

a) The heuristic is a function that estimates the cost from the current point in the path to the
end goal. It guides A* algorithm to search in the direction of the goal, improving
efficiency by not exploring less promising path.
Example: In video game development, the A* algorithm uses heuristic to find the shortest
path for an AI character to reach a target, considering the terrain and obstacles.
 b) To get A* from Dijkstra’s algorithm, which only accounts for the actual cost from the
start node to current node introduce a heuristics function that estimates the cost to the
goal.
For A* to be corrected, the heuristic must be admissible, meaning it never overestimate the
actual cost to reach the goal. If the heuristic is admissible, A* is guaranteed to find the
shortest path. If not, the algorithm might fail to do so, and the prove would no longer be true.

Question 6:

a)
One way to find bridges in a graph is using Depth-First Search (DFS) with discovery and low
values. This approach runs in O(V + E) time and is based on Tarjan’s Bridge-Finding
Algorithm.
Algorithm Steps:
1. Initialize discovery[u] and low[u] for each vertex uuu:
o discovery[u] is the time when uuu is first visited.
o low[u] is the earliest visited vertex reachable from uuu.
2. Perform a DFS traversal of the graph.
3. If a child vvv of uuu satisfies low[v] > discovery[u], then (u,v)(u, v)(u,v) is a bridge.
o This means vvv and its descendants cannot reach an ancestor of uuu without using
(u,v)(u, v)(u,v), so removing it disconnects the graph.

Source: Tarjan, R. (1974). A note on finding the bridges of a graph. Information Processing
Letters, 160-161.

b) A cycle graph has no bridges because every vertex belongs to at least one cycle, ensuring
multiple paths exist.

c) It’s a tree. A tree is an acylic connected graph since removing any edge will disconnect
the graph.