Bulletin of Faculty of Science ,Zagazig University (BFSZU) 2023

Bulletin of Faculty of Science, Zagazig University (BFSZU)))

e-ISSN: 1110-1555
Volume-2023, Issue-2, pp-1-5
https://bfszu.journals.ekb.eg/journal
Research Paper DOI: 10.21608/bfszu.2023.160904.1180
================================================================

A Data Analytics Approach for Solving Vehicle Routing Problem

Using Real Time Data

MOHAMED GAMAL a*, AMAL F. A. EL-GAWADa, AMANY ABDELSAMEAb

a
Faculty of Computers and Informatics, Zagazig University, Sharkia 44519, Egypt
b
Electronics Research Institute, Cairo 12622, Egypt

Corresponding author: [email protected]

ABSTRACT : This paper presents a new method for solving the traditional vehicle routing problem based
on real-time data. The proposed approach starts with collecting and analyzing real time data from
established sensors or data entered manually to make sure this data represent the real time status for
routes and edges then selecting the most influential factors (route width, car flow, car speed and average
number of cars that move in this rout in specific slot of time) that effect directly in classifying the
available routes. Based on this selection, a comparison is made between different routes and the best one
is selected accordingly. The approach is an iterative process that is repeated to obtain the global
optimum. Experimental results show that the proposed approach has the potential to obtain global
solutions that are optimal or more accurate because it affected by the real time data, so it reflects the real
status of the rout in real time than other evolutionary methods.
KEYWORDS: Real Time Data, Data Analytics, Vehicle Routing Problem.
----------------------------------------------------------------------------------------------------------------------------- --------
Date of Submission: 2022-09-05 Date of acceptance: 2023-04-09
----------------------------------------------------------------------------------------------------------------------------- ----------
I. INTRODUCTION
The Vehicle Routing Problem (VRP) is one of the most studied routing problems which has a wide range of
applications. Many problems in practice can be expressed in the form of VRP forms, such as, computer cabling,
robot control, traffic routing and crystallography. Most of Travelling Salesman Problems (TSP) and Vehicle
Routing Problem describe the situation in which a single vehicle or salesman completes all the routes in the
shortest possible time. The size of these problems can be rather large. For example, according to recent reports
on TSP applications, the circuit board boring problem was equivalent to a TSP of about 17,000 cities [1], while
the VLSI matrix construction problem was comparable to a TSP of 1.2 million cities [2]. With low complexity,
medium accuracy and low computational time the capacitated vehicle routing problem has been successfully
solved using metaheuristic algorithms such as simulated annealing [3]. There are algorithms available to give us
exact solutions, but they are inadequate [4].
The simplest way to solve these difficulties is to formulate them as a travelling salesman problem. The
majority of applications are based on real-world issues and hence appear to be of great relevance [10]. However,
in all of these studies, solutions offered based on static or approximated data. It is possible that these solutions
are not based on rapid changes in the factors affecting the results, in other words, solutions are not based on real
time data. In this paper, we present a way to solve this problem by applying the nearest neighbor algorithm on
real collected traffic data.

https://bfszu.journals.ekb.eg/journal Page 1
Bulletin of Faculty of Science ,Zagazig University (BFSZU) 2023
II. MATERIALS AND METHODS

Data Analytics and Nearest Neighbor Algorithm

2.1. Data Analytics: Data analytics help us to understand the blind view of data so we can use it in our case to
find the main characteristics that directly affect the weight of routes. Our data analytics approach consists of five
main steps as discussed below [11].
2.2. Nearest Neighbor Algorithm for VRP: The Nearest Neighbor Approach (NN) is a rough algorithm for
locating a suboptimal VRP solution. This algorithm begins with a single city as a starting point and continues to
visit all cities until they have all been visited exactly once. It seeks out the shortest route as shown in Fig.1 [16]
Step 1: Begin with any random vertex, which you'll refer to as current vertex.
Step 2: Find an edge called V that provides the shortest distance between the present
vertex and an unvisited vertex.
Step 3: Change the current vertex's state to unvisited vertex V and mark it as visited.
Step 4: If all of the vertices are visited at least once, the condition is terminated.
Step 5: Now proceed to step 2

2.3 Modifies Nearest Neighbor Algorithm based on real time data Steps:
Step 1: Begin with any random vertex, which you'll refer to as current vertex.
Step 2: extract the real time data from sensors as data.
Step 2: recalculate the edges weight based on the new real time data.
(The edge weight = avg speed on edge / edge length)
Step 3: Find an edge called E that provides the shortest distance between the present vertex and an
unvisited vertex.
Step 4: Change the current vertex's state to unvisited vertex V and mark it as visited.
Step 5: If all the vertices are visited at least once, the condition is terminated.
Step 6: Now proceed to step 2

Fig.1 Possible routes between visited locations.

https://bfszu.journals.ekb.eg/journal Page 2
Bulletin of Faculty of Science ,Zagazig University (BFSZU) 2023

Algorithm 1: Modifies Nearest Neighbor Algorithm based on real time data

// the input of this algorithm is Vertex, Edges and Edges weights

// the weights of edges are updated in each step
Data: Vertex, Edges and Edges weights
// the output is the best way between source and destination based on real time data
Result: Return the shortest path between source and destination
Function Nearest Neighbor (Vertex, Edges , Edges weights):
create vertex set Q
// Initialization
// loading the real time data for each round
Edges_data = Load_sensor_data(YYYY,MM,DD)
// calculate the edge weight based on new data
Edge_weight = avg_speed /edge_length
For Each vertex v in Graph: // initial distance from source to vertex v is set to ∞
dist(v) ← INFINITY // Previous node in optimal path from source
prev(v) ← UNDEFINED
add v to Q
end
dist(source) ← 0 // main loop
while Q is not empty:
U ← vertex in Q with min dist(u)
remove U from Q
S ← empty sequence
U← target // Do something only if the vertex is reachable
if prev(u) is defined or u = source: // Construct the shortest path with a stack
S
while u is defined: // Push the vertex onto the stack
u at the beginning of S // Traverse from target to source
u ← prev(u)
end
end
end

III. RESULTS

In this section, we will present some results from practical experiments for which data are
entered in different formats (estimated data, real fixed data, and real time data which
updated after each round), using this we will compare those results and we must know that
the main goal of using this data is real experiment, even if it does not have the best Practical

https://bfszu.journals.ekb.eg/journal Page 3
Bulletin of Faculty of Science ,Zagazig University (BFSZU) 2023
results.

This experiment is executed with estimated data, real fixed data, and real time data which
updated after each round, the particles in the initial population are selected at random. The
calculating results are shown as table 1:

The results of table 1 show that: our algorithm is quite efficient, and it was able to obtain
fairly ideal results

Type of data Number of best result worst result average result

executions
estimated data 25 35 53 44
real fixed data 25 30 46 38
real time data 25 33 60 46.5
Table.1 The comparisons of the results

Although the average results of this method are not the best, but it is the closest to reality
because it is based on real data and up to date.

REFERENCES

[1] Litke, J.D.: An Improved Solution to the Traveling Salesman Problem with Thousands of Nodes.
Communications of the ACM 27(12), 1227–1236 (1984)
[2] Korte, B.: Applications of Combinatorial Optimization. Talk at the 13th International Mathematical
Programming Symposium, Tokyo (1988)
[3] T. Vidal, T.G. Crainic, M. Gendreau, C. Prins, Heuristics for multi-attribute vehicle routing problems: a
survey and synthesis, Eur. J. Oper. Res. 231 (2013) 1–21.
[4] G. Laporte, The vehicle routing problem: an overview of exact and approximate algorithms, Eur. J. Oper.
Res. 59 (1991) 345–358.
[5] G. Clarke, J.W. Wright, Scheduling of vehicles from a central depot to a number of delivery points, Oper.
Res. 12 (4) (1964) 568–581.
[6] B. Gillett, L. Miller, A heuristic algorithm for the vehicle-dispatch problem, Oper. Res. 22 (2) (1974) 340–
349.
[7] M.L. Fisher, R. Jaikumar, A generalized assignment heuristic for vehicle routing, Networks 11 (1981)
109–124.
[8] R. Holmes, R. Parker, A vehicle scheduling procedure based upon savings and a solution perturbation
scheme, Oper. Res. 27 (1976) 83–92.
[9] Laporte, G. and S. Martello, THE SELECTIVE TRAVELLING SALESMAN PROBLEM. 2019.
[10] LENSTRA Mathematisch Centrum, J. K., &G Rinnooy Kan, A. H. (2019). Some Simple Applications of the
Travelling Salesman Problem. 26, 717-73.
[11] Li, B., Yang, J., & Hu, D. (2020). Dam monitoring data analysis methods: A literature review.

[12]Feng, Y., Duives, D., Daamen, W., & Hoogendoorn, S. (2021). Data collection methods for studying
pedestrian behaviour: A systematic review. 187. doi: 10.1016/j.buildenv.2020.107329
[13]Xu, X., Lei, Y., & Li, Z. (2020). An Incorrect Data Detection Method for Big Data Cleaning of Machinery
Condition Monitoring. 67, 2326-2336.
[14] Farouk, A., & Zhen, D. (2019). Big data analysis techniques for intelligent systems. 37, 3067-3071.
[15] Qin, X., Luo, Y., Tang, N., & Li, G. (2020). Making data visualization more efficient and effective: a
survey. 29, 93-117
[16] Cheikhrouhou, O., & Khoufi, I. (2021). A comprehensive survey on the Multiple Traveling Salesman

https://bfszu.journals.ekb.eg/journal Page 4
Bulletin of Faculty of Science ,Zagazig University (BFSZU) 2023
Problem: Applications, approaches and taxonomy. 40
[17] Ariyaluran Habeeb, R. A., Nasaruddin, F., Gani, A., Targio Hashem, I. A., Ahmed, E., & Imran, M.
(2019). Real-time big data processing for anomaly detection: A Survey. 45, 289-307.
[18] Krishnamurthi, R., Kumar, A., Gopinathan, D., Nayyar, A., & Qureshi, B. (2020). An overview of iot
sensor data processing, fusion, and analysis techniques. 20, 1-23.
[19] Madyatmadja, E. D., Nindito, H., Bhaskoro, R. A., Verasius, A., Sano, D., & Sianipar, C. P. M. (2021).
Algorithm To Find Tourism Place Shortest Route: A Systematic Literature Review. 28.
[20] Boyko, N. (2021). An Overview of Existing Methods for Solving the Travelling Salesman Problem in
Order to Find the Optimal Solution for Economic Problems. 1(7), 85-102.
[21] Aneiros, G., Cao, R., Fraiman, R., Genest, C., & Vieu, P. (2019). Recent advances in functional data
analysis and high-dimensional statistics. Journal of Multivariate Analysis, 170, 3-9.

https://bfszu.journals.ekb.eg/journal Page 5