THE SOLUTİON OF THE PROBLEM OF LOCATING THE DISTRUBITION OF

GOODS BY A SHIPPING AGENT

İbrahim İŞİT
Abstract
Transportation is a significant cost factor that affects the total cost of the product. The
problem of vehicle locomotion a long time ago came to reduce the cost of transporting
commodities. There are many approaches and models that are recommended to solve this
problem. In this study, a saving algorithm was proposed to solve the problem of agent
transportation in İstanbul. We will consider an establishment in Istanbul that uses its own
inventories and owns a marine vessel to supply goods to the vessels that need supplies or
supplies. In this study, we will use as an example, a company providing services to the vessels
in the anchorage areas of Istanbul. This company is constantly replenishing the ships that have
come to certain anchorage areas. We will only evaluate supplies made by sea.
Key Words: Saving Algorithm, Vehicle Rotation Problem, Maritime
INTRODUCTION
Logistics is generally defined as the transport of a product or service from a supply point
to various demand points. Both distribution and procurement procedures require effective
transport management. A good distribution management can lead to a significant reduction in
the overall distribution costs of the operator [1].
The vehicle routing problem (VRP) is a combinatorial optimization and integer
programming problem which asks "What is the optimal set of routes for a fleet of vehicles to
traverse in order to deliver to a given set of customers?" [7].
Car locating problems are generally concerned with the distribution of goods and
services between specific points in a network. In today's product distribution, the problems of
transporting goods and people are increasing [2]. For example, distribution of products and
services from one or more depots to various customer locations, Inventory planning and
shipment of products to sales locations, Trading of goods by airway companies, Distribution of
beverages in restaurants and cafés, money distribution, Petrol distribution, Milk distribution
and collection, Delivery of online purchases, DVD movie rental service, Garbage collection
and relocation, Distribution of store items in the main warehouse, There are many common
problems in daily life such as postal services, Distribution costs correspond to 1/3 - 2/3 of
logistics costs.
Multiple and often conflicting purposes can be taken into account in VRP. The most common
goals are: minimizing the total cost of carriage of vehicles based on total distance (or total travel
time) and fixed costs of the vehicles used (and the driver concerned), Minimizing the total
number of vehicles (or number of drivers), balancing the journeys in terms of travel time and
vehicle load, Minimizing the penalties resulting from fragmented distribution to customers [3].
Distribution costs correspond to 1/3 - 2/3 of logistics costs. Therefore, the efficient and
efficient use of distribution equipment and personnel has become an area of interest for business
managers. Developments in the transportation sector will affect the cost of the product. The
Vehicle Routing Problem (VRP) was first applied 60 years ago and it was aimed to improve
the product transportation. In order to reduce the cost of distribution, finding a route that gives
the shortest time or the shortest distance on a particular route is the most important factor [4].
Many methods have been proposed for the solution of VRP. VRP solution methods are
divided into precise solution methods and heuristic methods [2]. The exact solution methods
are divided into two; Minimum Used Vehicle Method and Multi - Faceted Approach. Intuitive
methods are also divided into two groups; Classical Heuristic Methods and Meta Heuristic
Methods. As an example of classical heuristic methods, savings algorithm, sweep method, two
step method and improved petal heuristics can be shown. Examples of meta-heuristic methods
include annealing simulation, artificial neural networks, tabu search and ant algorithm.
In a standard vehicle rotating problem, it is attempted to meet the demands of customer
points located at different points by means of vehicles in storage. In doing so, the objective is
to efficiently and efficiently determine the route that meets customer needs in the shortest
possible time, on the shortest route, and at the least cost. The following factors must be taken
into account when making vehicle rotation [9]:
- The demands of the customers in the network must be fully met.
- Each destination in the network must be visited only once by a single vehicle.
- The route begin at the store and end at the store again.
- The total requests for customers on the track shouldn’t exceed the total capacity of the
transport vehicle.
- Each vehicle must only operate on just one route.
- Vehicle routing should be the minimization of the total distance that the basic aim vehicles
will take.
There are numerous solutions to the problem of vehicle rotating problems. The problem
can be solved in two different ways in cases where there is only one origin point and a large
number of distribution points. These include: single-vehicle tool tipping and multi-vehicle
tooltip. However, this study deals with a problem that distributes with only one vehicle with a
large number of distribution points moving from one origin point to another.
LITERATURE REVIEW
Problems with vehicle routing are long-running subject (about 50 years). There are
hundreds of different models and algorithms in the literature. It is an issue that many researchers
have been dealing with because of the multiplicity of the application area and the interest of the
problems [8].
The main objective of the problem is to minimize the cost function, to provide all the
constraints, to Reduce the number of vehicles to be used the most, and to decrease the total
distance or time to a minimum. The secondary purpose is to maximize customer satisfaction.
 VRP has many applications in the industry. VRP optimization software can provide cost
savings of up to 5% [6], since transport costs are generally a significant part of a product's cost
(10%) [5].
In the real life of VRP bring with it lots of limitations. Constraints are collected in 3
different groups [2]:
1 ) Vehicle-related constraints
- Vehicle capacity boundary (heft or volume)
- Total time limitation
- Legal limitations for the operating hours of the driver
2) Customer - related restrictions
- Each customer requests a product or distributes a specific product type; Logistics companies,
for example.
Having certain periods for distribution
3) Another constraints
- Redistribution on the same day as a vehicle, after returning to the warehouse
- One turn is longer than a day
- Too many warehouses
In the VRP, distribution rods must meet the following conditions [2]:
- All customers' requests must be done.
- Every customer has to be in one vehicle tracking route.
- The total customer requirements placed on a distribution route has to be lower than the
capacity of the vehicle on that route.
- Every route must start at the warehouse and after than it must finish at the warehouse.
- The total distance traveled on any given course should not exceed the predetermined
maximum course distance.
- In some VRP varieties, x number of vehicles is fixed, and in some variants it is variable.
In this study, the savings algorithm was used because there is only one warehouse and
a large number of distribution centers and there is no ambiguity. Saving algorithm is usually
used in the literature to solve simple problems quickly and easily. In this study, it is achieved
with simple MS Excel applications without any software program required.
SAVING ALGORTHYM
Clarke and Wright developed the saving algorithm in 1964 to solve the problem of
vehicle rotating. In this algorithm, we have one warehouse and one more than distribution site.
Starting from the greatest savings value, routes are identified and combined. The saving
algorithm is shown in figure below.

i. i.
Bostancı oa Bostancı oa

. .
B e fo r e la n n in g A ft e r la n n in g

Figure 1. Saving Algorithm

From figure 1 we find the amount of savings (𝑠𝑖𝑗) with equation 1. Equation 2 is the
formula for the amount of savings. In Equation 2, C0𝑖 travel cost to the warehouse 𝑖 customer,
C0𝑗 travel cost to the customer, and 𝑐𝑖𝑗𝑖 represents the cost of travel to the customer.

𝑠𝑖𝑗=(𝑐0𝑖+𝑐0𝑖+𝑐0𝑗+𝑐0𝑗)−(𝑐0𝑖+𝑐0𝑗+𝑐𝑖𝑗) (1)

𝑠𝑖𝑗= 𝑐0𝑖+𝑐0𝑗−𝑐𝑖𝑗 (2)

To reach the end, 𝑐𝑖𝑗 = 𝑐𝑗𝑖 will be assumed. On this assumption, one-way streets will
not be considered. In this study, the saving algorithm is solved with the help of MS Excel and
no other software is used
APLICATION
In this application, the maritime agency located in the southern side of Istanbul provides
the supply to the ships in the sea using the sea route. Supply refers to food, deck and machinery
equipment, and various equipment requirements. Transportation costs prior to the sea route will
not be considered. The one boat will be used for distribution. On the boat, EUR 6 type pallets
will be used to carry the materials. The carrying capacity of these pallets is 500 kilograms. The
boat has a total carrying capacity of 16 pallets. The boat’s machine power is high and will not
be affected by ship traffic. This study used real anchorage zones, but the agency operation is
fiction. Departure and return distances are considered equal for anchorage areas and Bostancı
Quay. The distance values given in the tables were calculated from the nautical miles.
It is designed as a starting point Bostanci Quay, which is used as a warehouse in the
theory of saving algorithm. The anchorage areas (excluding the quarantine area) and the
anchorage area of the Ambarlı port were used according to the instructions of "Istanbul Harbor
Anchorage Usage Guidelines" used in the south of İstanbul. In all of the anchorage areas there
are areas reserved for ships. In this application, the distances are calculated while the middle
region of the area of the anchorage sites is considered.
 Anchorage zone’s descriptions [10,11]:
1- B Anchorage Zone: Free-standing anchorage for harbor shields and long-stay ships
2- C Anchorage Zone: Free anchorage of ships carrying dangerous goods
3- Küçükçekmece Anchorage Zone: For carr ing dangerous cargo ships
4- AmbarlıAnchorage Zone: For ships coming Ambarlı ort
5- Kartal Achorage Zone; For vessels not carrying hazardous cargo that will pass
through the Bosphorus from Istanbul to the Sea of Marmara and İzmit Gulf and dangerous cargo
vessels operating in the cabotage line

All anchorage areas and Bostancı ua numbered on the tables.

Table 1. Anchorage area and demands

Anchorage areas
Supply Number of
Numbers and Bostancı Codes
Demands Required Pallets
Quay

Bostancı Quay 0 - -
0
B Anchorage
1 3,7 ton 8
1 Area
C Anchorage
2 4,1 ton 9
2 Area
Küçükçekmece
3 0,5 ton 1
3 Anchoarge Area
Ambarlı
4 3 ton 6
4 Anchorage Area
Kartal Anchorage
5 2 ton 4
5 Area
 Table 2 gives the distance matrix of the anchorage areas to be distributed to each other
and to the Bostancı ua , which is considered as the main warehouse. This distance matrix
represents the distance between two points. Table 2 was measured using Google Earth. When
the measurement was made, the ship separation line in Marmara region was taken into
consideration and it was tried to draw the closest possible route. Therefore, the distance matrix
contains real data. Saving quantities (𝑠𝑖𝑗) can be calculated from Table 2. For example, the
amount of savings between zone B and zone C is calculated as follows: 𝑠12 = 𝑐01 + 𝑐02-𝑐12
In this formula, 𝑐01 indicates the distance from zone B to Bostanci Quay, ie 9.2 miles, distance
c02 to zone berth, . The value of C12 gives the distance between B and C, which is 3,4 nautical
miles. As a result, the savings between B and C is s12 = 9,2 + 11,7-3,4 = 17,5.

Table 2 Distance matrix

0 1 2 3 4 5

1 9,16

2 11,7 3,4

3 17,6 8,5 5,4

4 26 17,2 13,7 8,5

5 7,4 15,7 17,3 22,7 30,4

Table 3 Calculated saving matrix

1 2 3 4 5

2 17,46

3 18,26 23,9

4 17,96 24 35,1

5 0,86 1,8 2,3 3

All savings in Table 3 are sorted from largest to smallest. The calculation process starts
with the largest value, and then the other values are processed sequentially.
 Table 4. Saving treatments
Saving Number of
Location 1 Location 2 Decision
Amount Required Pallets
35,1 3 4 7 3//4
24 2 4 16 2//3//4
23,9 2 3 16 Same route
18,26 1 3 24 Over capacity
17,96 1 4 24 Over capacity
17,46 1 2 24 Over capacity
3 4 5 20 Over capacity
2,3 3 5 20 Over capacity
1,8 2 5 20 Over capacity
0,86 1 5 12 1 // 5

As shown in Table 4, the largest savings is the path between the savings value of 34.1
and the areas 3 and 4. If the number of pallets required for replenishment requests of the 3rd
and 4th regions are collected, 1 + 6 = 7 pallets are required. When the track's pallet capacity is
16, this route does not exceed the capacity and is suitable for deployment. The number of pallets
required for the distribution of 2nd and 4th regions with the saving value of 24 in the 2nd place
is 9 + 6 = 15. If this route is combined with the previous route, the total pallet requirement for
route 2-3-4 will be 9 + 1 + 6 = 16 and is suitable for distribution.
Region 1 at the saving value of 18,26 can not join the distribution route. Because 8 more
pallets will be needed and therefore 24 pallets will be needed, the capacity will be exceeded.
The 5th zone with 3 saving values can not join the distribution route because 4 more
pallets will be needed, and 20 pallets will be exceeded due to the need for pallets and
distribution will not be possible in that case.
If the required pallet is more than 16 requests, the decision is "Over capacity" and if the
same rotations are "same routes".

RESULTS

In this study, appropriate routes have been determined for the distribution of an agency
workstation with the help of the savings algorithm. If the agency makes an unscheduled
distribution, it will be able to make 5 different trips if all the anchorage areas are irregular,
which will result in high fuel cost. The agency boat will travel a total of 143.7 miles if it goes
back and forth separately for distribution to each anchorage areas.
As a result of this study, two different routes were determined to provide the maximum
savings in distribution work. These are roots 2-3-4 and 1-5. Distribution is done twice a day.
The total distance covered is 83.9 sea miles.
As a result, the total distance traveled decreased by 59.8 nautical miles. This resulted in
a gain of approximately 40% from fuel and time lost in the journey.
 REFERANCES
1. Kosif, b., & ekmekçi, i. (2012). Araç rotalama sistemleri ve tasarruf algoritması
u gulaması.
2. Düzakın, e., & demircioğlu, m. (2009). Araç rotalama problemleri ve çözüm
öntemleri. Çukurova üniversitesi iktisadi ve idari bilimler fakültesi dergisi, 13
3. Atmaca, e. (2012). Bir kargo şirketinde araç rotalama problemi. Tübav bilim
dergisi, 5(2), 12-27.
4. Ulutaş, a., ba rakçıl, a. O., & kutlu, m. B. Araç rotalama probleminin tasarruf
algoritması ile çözümü: Sivas’ta bir ekmek fırını için u gulama. Cumhuriyet
üniversitesi iktisadi ve idari bilimler dergisi, 18(1), 185-197.
5. Comtois, claude; slack, brian; rodrigue, jean-paul (2013). The geography of transport
systems (third edition. Bas.). London: routledge, taylor & francis group. Isbn 978-0-
415-82254-1.
6. Geir hasle, knut-andreas lie, ewald quak,; kloster, o (2007). Geometric modelling,
numerical simulation, and optimization applied mathematics at sıntef ([online-ausg.].
Bas.). Berlin: springer verlag.
7. URL-1: http://www.wiki-zero.net/ vehicle routing problem
8. Dantzig, g.b., ramser, .h., ‘‘the truck dispatching problem’’, management science, 6,
80-91 (1959)
9. Tan, k.c., 2000, a framework of supply chain management literature, europan journal of
purchasing supply management
10. Instructıons for use of Istanbul harbor anchorage locatıons
11. URL-2: northmaritime.com/ports-info/turkish-straits-instanbul-region/31-ports-of-
ambarli.html