mafulechka
mafulechkaAug. 15, 2019, 4:22 a.m.

Dijkstra's algorithm

Dijkstra's algorithm allows us to find the shortest path between any two vertices in a graph.

It differs from a minimum spanning tree in that the shortest distance between two vertices may not include all vertices in the graph.


How Dijkstra's algorithm works

Dijkstra's algorithm works on the basis that any subpath B -> D of the shortest path A -> D between vertices A and D is also the shortest path between vertices B and D.

Dijkstra used this property in the opposite direction, i.e. we re-estimate the distance of each vertex from the initial vertex. We then visit each node and its neighbors to find the shortest subpath to those neighbors.

The algorithm uses a "greedy" approach in the sense that we find the next best solution, hoping that the end result is the best solution for the whole problem.

An example of Dijkstra's algorithm

It's easier to start with an example and then think about the algorithm.

Dijkstra's algorithm. Pseudocode.

We need to store the path distance of each vertex. We can store it in an array of size v, where v is the number of vertices.

We would also like to get the shortest path, not just know its length. To do this, we map each vertex to the last updated path length.

Once the algorithm is finished, we can return from the destination node to the source node to find the path.

A minimum priority queue can be used to efficiently obtain the vertex with the smallest path distance.

function dijkstra(G, S)
    for each vertex V in G
        distance[V] <- infinite
        previous[V] <- NULL
        If V != S, add V to Priority Queue Q
    distance[S] <- 0

    while Q IS NOT EMPTY
        U <- Extract MIN from Q
        for each unvisited neighbour V of U
            tempDistance <- distance[U] + edge_weight(U, V)
            if tempDistance < distance[V]
                distance[V] <- tempDistance
                previous[V] <- U
    return distance[], previous[]

Code for Dijkstra's algorithm

An implementation of Dijkstra's algorithm in C++ is given below. The complexity of the code can be improved, but abstractions are good for relating code to an algorithm.

#include <iostream>
#include <vector>
#define INT_MAX 10000000
using namespace std;
void DijkstrasTest();
int main()
{
    DijkstrasTest();
    return 0;
}
class Node;
class Edge;
void Dijkstras();
vector<Node*>* AdjacentRemainingNodes(Node* node);
Node* ExtractSmallest(vector<Node*>& nodes);
int Distance(Node* node1, Node* node2);
bool Contains(vector<Node*>& nodes, Node* node);
void PrintShortestRouteTo(Node* destination);
vector<Node*> nodes;
vector<Edge*> edges;
class Node
{
public:
    Node(char id) 
        : id(id), previous(NULL), distanceFromStart(INT_MAX)
    {
        nodes.push_back(this);
    }
public:
    char id;
    Node* previous;
    int distanceFromStart;
};
class Edge
{
public:
    Edge(Node* node1, Node* node2, int distance) 
        : node1(node1), node2(node2), distance(distance)
    {
        edges.push_back(this);
    }
    bool Connects(Node* node1, Node* node2)
    {
        return (
            (node1 == this->node1 &&
            node2 == this->node2) ||
            (node1 == this->node2 && 
            node2 == this->node1));
    }
public:
    Node* node1;
    Node* node2;
    int distance;
};
///////////////////
void DijkstrasTest()
{
    Node* a = new Node('a');
    Node* b = new Node('b');
    Node* c = new Node('c');
    Node* d = new Node('d');
    Node* e = new Node('e');
    Node* f = new Node('f');
    Node* g = new Node('g');
    Edge* e1 = new Edge(a, c, 1);
    Edge* e2 = new Edge(a, d, 2);
    Edge* e3 = new Edge(b, c, 2);
    Edge* e4 = new Edge(c, d, 1);
    Edge* e5 = new Edge(b, f, 3);
    Edge* e6 = new Edge(c, e, 3);
    Edge* e7 = new Edge(e, f, 2);
    Edge* e8 = new Edge(d, g, 1);
    Edge* e9 = new Edge(g, f, 1);
    a->distanceFromStart = 0; // set start node
    Dijkstras();
    PrintShortestRouteTo(f);
}
///////////////////
void Dijkstras()
{
    while (nodes.size() > 0)
    {
        Node* smallest = ExtractSmallest(nodes);
        vector<Node*>* adjacentNodes = 
            AdjacentRemainingNodes(smallest);
        const int size = adjacentNodes->size();
        for (int i=0; i<size; ++i)
        {
            Node* adjacent = adjacentNodes->at(i);
            int distance = Distance(smallest, adjacent) +
                smallest->distanceFromStart;

            if (distance < adjacent->distanceFromStart)
            {
                adjacent->distanceFromStart = distance;
                adjacent->previous = smallest;
            }
        }
        delete adjacentNodes;
    }
}
// Find the node with the smallest distance,
// remove it, and return it.
Node* ExtractSmallest(vector<Node*>& nodes)
{
    int size = nodes.size();
    if (size == 0) return NULL;
    int smallestPosition = 0;
    Node* smallest = nodes.at(0);
    for (int i=1; i<size; ++i)
    {
        Node* current = nodes.at(i);
        if (current->distanceFromStart <
            smallest->distanceFromStart)
        {
            smallest = current;
            smallestPosition = i;
        }
    }
    nodes.erase(nodes.begin() + smallestPosition);
    return smallest;
}
// Return all nodes adjacent to 'node' which are still
// in the 'nodes' collection.
vector<Node*>* AdjacentRemainingNodes(Node* node)
{
    vector<Node*>* adjacentNodes = new vector<Node*>();
    const int size = edges.size();
    for(int i=0; i<size; ++i)
    {
        Edge* edge = edges.at(i);
        Node* adjacent = NULL;
        if (edge->node1 == node)
        {
            adjacent = edge->node2;
        }
        else if (edge->node2 == node)
        {
            adjacent = edge->node1;
        }
        if (adjacent && Contains(nodes, adjacent))
        {
            adjacentNodes->push_back(adjacent);
        }
    }
    return adjacentNodes;
}
// Return distance between two connected nodes
int Distance(Node* node1, Node* node2)
{
    const int size = edges.size();
    for(int i=0; i<size; ++i)
    {
        Edge* edge = edges.at(i);
        if (edge->Connects(node1, node2))
        {
            return edge->distance;
        }
    }
    return -1; // should never happen
}
// Does the 'nodes' vector contain 'node'
bool Contains(vector<Node*>& nodes, Node* node)
{
    const int size = nodes.size();
    for(int i=0; i<size; ++i)
    {
        if (node == nodes.at(i))
        {
            return true;
        }
    }
    return false;
}
///////////////////
void PrintShortestRouteTo(Node* destination)
{
    Node* previous = destination;
    cout << "Distance from start: " 
        << destination->distanceFromStart << endl;
    while (previous)
    {
        cout << previous->id << " ";
        previous = previous->previous;
    }
    cout << endl;
}
// these two not needed
vector<Edge*>* AdjacentEdges(vector<Edge*>& Edges, Node* node);
void RemoveEdge(vector<Edge*>& Edges, Edge* edge);
vector<Edge*>* AdjacentEdges(vector<Edge*>& edges, Node* node)
{
    vector<Edge*>* adjacentEdges = new vector<Edge*>();
    const int size = edges.size();
    for(int i=0; i<size; ++i)
    {
        Edge* edge = edges.at(i);
        if (edge->node1 == node)
        {
            cout << "adjacent: " << edge->node2->id << endl;
            adjacentEdges->push_back(edge);
        }
        else if (edge->node2 == node)
        {
            cout << "adjacent: " << edge->node1->id << endl;
            adjacentEdges->push_back(edge);
        }
    }
    return adjacentEdges;
}
void RemoveEdge(vector<Edge*>& edges, Edge* edge)
{
    vector<Edge*>::iterator it;
    for (it=edges.begin(); it<edges.end(); ++it)
    {
        if (*it == edge)
        {
            edges.erase(it);
            return;
        }
    }
}

We recommend hosting TIMEWEB
We recommend hosting TIMEWEB
Stable hosting, on which the social network EVILEG is located. For projects on Django we recommend VDS hosting.

Do you like it? Share on social networks!

Comments

Only authorized users can post comments.
Please, Log in or Sign up
AD

C ++ - Test 004. Pointers, Arrays and Loops

  • Result:50points,
  • Rating points-4
m

C ++ - Test 004. Pointers, Arrays and Loops

  • Result:80points,
  • Rating points4
m

C ++ - Test 004. Pointers, Arrays and Loops

  • Result:20points,
  • Rating points-10
Last comments
ИМ
Игорь МаксимовNov. 22, 2024, 8:51 a.m.
Django - Tutorial 017. Customize the login page to Django Добрый вечер Евгений! Я сделал себе авторизацию аналогичную вашей, все работает, кроме возврата к предидущей странице. Редеректит всегда на главную, хотя в логах сервера вижу запросы на правильн…
Evgenii Legotckoi
Evgenii LegotckoiOct. 31, 2024, 11:37 a.m.
Django - Lesson 064. How to write a Python Markdown extension Добрый день. Да, можно. Либо через такие же плагины, либо с постобработкой через python библиотеку Beautiful Soup
A
ALO1ZEOct. 19, 2024, 5:19 a.m.
Fb3 file reader on Qt Creator Подскажите как это запустить? Я не шарю в программировании и кодинге. Скачал и установаил Qt, но куча ошибок выдается и не запустить. А очень надо fb3 переконвертировать в html
ИМ
Игорь МаксимовOct. 5, 2024, 4:51 a.m.
Django - Lesson 064. How to write a Python Markdown extension Приветствую Евгений! У меня вопрос. Можно ли вставлять свои классы в разметку редактора markdown? Допустим имея стандартную разметку: <ul> <li></li> <li></l…
d
dblas5July 5, 2024, 8:02 a.m.
QML - Lesson 016. SQLite database and the working with it in QML Qt Здравствуйте, возникает такая проблема (я новичок): ApplicationWindow неизвестный элемент. (М300) для TextField и Button аналогично. Могу предположить, что из-за более новой верси…
Now discuss on the forum
m
moogoNov. 22, 2024, 4:17 a.m.
Mosquito Spray System Effective Mosquito Systems for Backyard | Eco-Friendly Misting Control Device & Repellent Spray - Moogo ; Upgrade your backyard with our mosquito-repellent device! Our misters conce…
Evgenii Legotckoi
Evgenii LegotckoiJune 24, 2024, 12:11 p.m.
добавить qlineseries в функции Я тут. Работы оень много. Отправил его в бан.
t
tonypeachey1Nov. 15, 2024, 3:04 a.m.
google domain [url=https://google.com/]domain[/url] domain [http://www.example.com link title]
NSProject
NSProjectJune 4, 2022, 12:49 a.m.
Всё ещё разбираюсь с кешем. В следствии прочтения данной статьи. Я принял для себя решение сделать кеширование свойств менеджера модели LikeDislike. И так как установка evileg_core для меня не была возможна, ибо он писался…

Follow us in social networks