The Problem

Given a list of friendships, calculate the degrees of separation between each person listed and the coolest guy around Quinn

Step 1: Data Representation

  • Often the hardest part of solving any problem
  • Very cruicial to the rest of your approach
  • Should be well thought out
  • The more problems you do, the easier this becomes

Undirected Graphs

  • We will represent the data given to us as an undirected graph
  • What is a graph?
    • Not your typical line or bar graph
    • Graphs in math are defined as a 2-tuple (V, E)
    • A set of vertices connected by a set of edges
  • Each person will be a vertex in our graph and each friendship will be an edge
  • Undirected just means that each edge is bidirectional (just like the friendship relation)

Step 2: Choosing an Algorithm

  • Now that we have decided on a way to represent our data, we must choose an algorithm to process it
  • Since we have decided to use a graph, it would only make sense to consider graph algorithms
  • Breadth first search is a commonly used algorithm for traversing graphs
  • Useful for this problem because it traverses the graph in a level order manner
    • It first handles the root (Quinn)
    • Then moves to Quinn’s direct neighbours (QDist level 1)
    • Continuing down until each level has been processed

Step 3: Implementation

  • Now that we have our data structure and algorithm mapped out, we can begin implementing them

Graph Implementation

  • Most languages do not include a graph data structure in the standard library
  • Thankfully they are relatively simple, and most languages do provide the tools nessesary to do so
  • Generally implemented using a map
  • We map verticies to a list of their corresponding adjacent verticies

How I Did it

  • Not the only way

Breadth First Search and Cycle Detection

  • Breadth first search is also relatively easy to implement
  • One thing to keep in mind is that cycles may appear in the graph
  • For instance:

Breadth First Search and Cycle Detection

  • To handle this we simply need to make sure we do not visit any vertex more than once
  • Each time a node is visited add it to a set of visited nodes
  • Do not process any node that has already been visited

How I Did It

Putting it all together

  • We now have a map that maps each person to their QDist value
  • Basically done at this point, we just need to handle the uncool people
  • Since there does not exist a path in the graph from them to Quinn, they will not appear in the visited map
  • This can be solved many different ways
  • You could store all the names at the very beginning and then check which do not appear in the visited map