On The Optimization of Dijkstra’s Algorithm

Seifedine Kadry, Ayman Abdallah, Chibli Joumaa

American University oft he Middle East, Egaila, Kuwait

{seifedine.kadry, Ayman.abdallah, Chibli.joumaa}@acm.edu.kw

Abstract. In this paper, we propose some amendment on Dijkstra’s algorithm

in order to optimize it by reducing the number of iterations. The main idea is to
solve the problem where more than one node satisfies the condition of the
second step in the traditional Dijkstra’s algorithm. After application of the
proposed modifications, the maximum number of iterations of Dijkstra’s
algorithm is less than the number of the graph’s nodes.

Keywords: Dijkstra’s algorithm, directed graph, shortest path

1 Introduction

The shortest-route problem determines a route of minimum weight connecting two

specified vertices, source and destination, in a weight graph (digraph) in a
transportation network. Other situations can be represented by the same model like
VLSI design, equipment replacement and others. There are different types of shortest
path algorithm [5] to find the shortest path of any graph. Most frequently encountered
are the following:
• Shortest path between two specified vertices
• Shortest path between all pairs of vertices
• Shortest path from a specified vertex to all others
The most efficient algorithm used to find the shortest path between two known
vertices is Dijkstra’s algorithm [1]. Some improvements on Dijkstra algorithm are
done in terms of efficient implementation [3] and cost matrix [4]. In this paper, we
propose some improvement in order to reduce the number of iterations and to find
easily and quickly the shortest path.

2 Dijkstra’s Algorithm

The problem of finding the shortest path from a specified vertex s to mother t can be
stated as follows:
A simple weighted digraph G of n vertices is described by a n by n matrix D=[dij],
where, dij = length (or distance or weight) of the directed edge from vertex i to vertex
j:
 > 0, if i ≠ j

d ij = = 0, if i = j
= ∞ (large number), if there is no edge from i to j


Dijkstra’s algorithm labels the vertices of the given digraph, At each stage in the
algorithm some vertices have permanent labels and others temporary labels. The
algorithm begins by assigning a permanent label 0 to the starting vertex s, and
temporary label infinity to the remaining n-1 vertices. From then on in each iteration,
another vertex sets a permanent label, according to the following rules:
a. Every vertex j that is not yet permanently labeled gets a new temporary
label whose value is given by min[old label of j, (old label of i + dij)],
where i is the latest vertex permanently labeled, in the previous iteration,
and dij is the direct distance between vertices i and j. If i and j are not
joined by an edge, then dij=infinity.
b. The smallest value among all the temporary labels is found, and this
becomes the permanent label of the corresponding vertex. In case of more
than one shortest path, select any one of the candidates for permanent
labeling. Steps a and b are repeated alternately until the destination vertex t
gets a permanent label. The first vertex to be permanently labeled is at a
distance of zero from s. The second vertex to get a permanent label (out of
the remaining n-1 vertices) is the vertex closest to s From the remaining n-
2 vertices, the next one to be permanently labeled is the second closest
vertex to s. And so on. The permanent label of each vertex is the shortest
distance of that vertex from s. This statement can be proved by induction.

3 Modified Dijkstra’s Algorithm

In this section, we will present a modified version of Dijkstra's Algorithm in order

to reduce the total number of iteration by optimized the situation of many shortest
paths:
a. Every vertex j that is not yet permanently labeled gets a new temporary label
whose value is given by min[old label of j, (old label of i + dij)], where i is
the latest vertex permanently labeled, in the previous iteration, and dij is the
direct distance between vertices i and j. If i and j are not joined by an edge,
then dij=infinity [6].
b. The smallest value among all the temporary labels is found, and this
becomes the permanent label of the corresponding vertex. In case of more
than one shortest path, select all of them for permanent labeling.
c. Steps (a) and (b) are repeated alternately maximum n-1 times until the
destination vertex t gets a permanent label.
d. Le T=[tij] is the shortest path matrix that we can build it form D, as follows
0, if d ji = 0 or d ji = ∞
tij = 
d ji , otherwise
3 Illustrative Example

The network in the following figure (figure 1) gives the distances in miles between
pairs of cities 1,2,… and 8. For efficiency purpose, we will find the shortest route
between cities 1 and 8 using the traditional and modified version of Dijkstra
algorithm:

Fig. 1: Network of 8 nodes

Dijkstra algorithm:
The number of iterations to find the shortest route between cities 1 and 8 by using
Dijkstra algorithm is 8 (figure 2), this is the result of TORA [1] software:

Fig. 2: Iterations 1 to 8 of Dijkstra algorithm

Modified Dijkstra algorithm:
By using the modified version of Dijkstra algorithm, 5 iterations only are required to
get the shortest route between cities 1 and 8:
  0 1 2 ∞ ∞ ∞ ∞ ∞
Initially  1 0 1 5 2 ∞ ∞ ∞
 
 2 1 0 2 1 4 ∞ ∞
 
∞ 5 2 0 3 6 8 ∞
D=
∞ 2 1 3 0 3 7 ∞
 
∞ ∞ 4 6 3 0 5 2 
∞ ∞ ∞ 8 7 5 0 6 
 
∞ ∞ ∞ ∞ ∞ 2 6 0 
and the set of permanent vertices is P={1}.
Iteration 1: { } 1j {
P = 1,2 , d = 0,1,2,4, ∞, ∞, ∞, ∞ }
Iteration 2: P = {1,2,3}, d1 j = {0,1,2,4,3,6, ∞, ∞}
Iteration 3: P = {1,2,3,5}, d1 j = {0,1,2,4,3,6,10, ∞}
Iteration 4: P = {1,2,3,4,5,6}, d1 j = {0,1,2,4,3,6,10,8}
Iteration 5: P = {1,2,3,4,5,6,7,8}, d1 j = {0,1,2,4,3,6,10,8}

0 1 0 0 0 0 0 0
0 0 1 0 2 0 0 0

0 0 0 2 0 4 0 0
 
0 0 0 0 0 0 0 0
T=
0 0 0 0 0 0 7 0
 
0 0 0 0 0 0 0 2
0 0 0 0 0 0 0 0
 
0 0 0 0 0 0 0 0

Fig. 3: Tree of shortest route

Based on the previous tree (figure 3), the shortest route and distance between cities 1
and 8 are: 1-2-3-6-8 and the distance is 8 miles.

4 Conclusion

In this paper an optimization of Dijkstra algorithm is presented. The main idea is to

deal with the second stage of the traditional version of this algorithm where the
shortest routes from a node to its successors are many and to reduce the number of
iterations. The main feature of the proposed amendment is the remarkable reduction
in the number of iterations and the ease in finding the shortest route from any city to
any city. In the future work, we will implement the proposed modifications then to
compare it with the existing one in terms of complexity.
References

1. Taha, H. Operations research an introduction, ninth edition. Pearson publisher, 2011.

2. Nar, D. Graph theory with applications to engineering and computer science, Prentice Hall,
1997.
3. Sneyers J., Schrijvers, T. and Demoen B. Dijkstra's Algorithm with Fibonacci
Heaps: An Executable Description in CHR. 20th Workshop on Logic
Programming (WLP'06), Vienna, Austria, February 2006.

4. Brendan, H., Gordon, J. Generalizing Dijkstra's Algorithm and Gaussian Elimination for
Solving MDPs. Conference: International Conference on Automated Planning and
Scheduling/Artificial Intelligence Planning Systems - ICAPS(AIPS) , pp. 151-160, 2005.
5. Ravi, N., Sireesha, V. Using Modified Dijkstra’s Algorithm for Critical Path Method in a
Project Network. International Journal of Computational and Applied Mathematics. Volume
5 number 2 pp 217-225. 2010.
6. Sudhakar, T.D.,Vadivoo, N.S., Slochanal, S.M.R., Ravichandran, S. Supply restoration in
distribution networks using Dijkstra's algorithm. International Conference on Power System
Technology, Signapure. 2004 (PowerCon 2004)