27

Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

Optimization Modeling of a Poultry Industry

Supply Chain Network
Mohammad Khadem #1, Ahm Shamsuzzoha#2, Sujan Piya #3
#
Department of Mechanical and Industrial Engineering
Sultan Qaboos University, Muscat, Oman
1
[email protected]
2
[email protected]
3
[email protected]

Abstract— Supply chain management is an asset to supply chain members and improving the network
every industry globally due to its positive outcomes [18]. Supply chain entails the progression of
such as faster response time, reduction of unwanted organization, their facilities and operations that are
inventory and lower sales costs with enhanced included in manufacturing and delivering of a
customer service. It is therefore important to focus on
product or service. The progression starts with
improving the supply chain network of any industry.
primal suppliers of raw materials and ends to the
In this research a deterministic linear programming
(LP) model has been developed for a supply chain costumer. It includes all operations and facilities
network for the poultry Industry in Oman. The study that are accomplished to provide a product or
analyzes the existing supply chain network within a service [2], [3], [16]. Operations involve inventory
poultry industry and recommended its improvement management, forecasting, purchasing, scheduling,
based on the identified factors while giving more production, quality control, etc. Facilities involve
emphasis on the routing and distribution network factories, warehouses, processing department,
aspects of supply chain. The recommendation, in the distribution department, offices, etc. There are two
form of optimization model, is verified and validated
types of movements in the supply chain system: (i)
using Lingo optimization software. Also, heuristic
movement of physical material in the same
method is proposed and tested to overcome the
complexity of optimization model. direction of the chain, and (ii) interchange of
Keywords— Supply and distribution network, routing, information that transfers in both directions of the
case study, Poultry model, Oman
chain.

Supply chain management (SCM) is very

1. Introduction
important in business community nowadays due to
The word supply chain was coined by management the complexity of the systems overlapping each
consultants in 1980s [11], [12]. In 1990’s supply other. Therefore, the number and type of
chain management (SCM) has stimulated much organizations within a supply chain network
enthusiasm among researchers and the industrial depend on the complexity and category of the
practitioners alike which is evident from the rapid organizations, whether it is manufacturing or
growth of the number of articles published in this service oriented. Obviously, manufacturing
domain. SCM is basically a system of individuals, organization is more complicated than service
organizations and the flow of resources, services, organization. The performance of the supply chain
activities and information between them [14]. It depends on the policies and activities that are
implies that it is an integration of all the processes carried out during the production process, which is
across the supply chain network and the located in the middle of the chain [2]. Therefore,
management of this integration. Ref. [6] and [19] the production affects the flow of material and
emphasized on the effective integration of supply information. It also determines policy and rules that
chain members and maintaining close relationships are implemented to coordinate all stations together
with both suppliers and customers. At present, the in the chain.
focus of most of the researchers in this domain is to
The purpose of SCM is to improve efficiency and
understand the long-term relationships between the
effectiveness of companies, which directly affect
profits and minimize costs. Company uses various
______________________________________________________________
International Journal of Supply Chain Management strategies related to the procurement of raw
IJSCM, ISSN: 2050-7399 (Online), 2051-3771 (Print) materials, transportation of materials to and from
Copyright © ExcelingTech Pub, UK (http://excelingtech.co.uk/)
 28
Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

the company, manufacturing operation and there are several studies, which proved that the
distribution of the product to the customer to rapid increase of consumers’ awareness against
improve the performance of whole supply chain their daily food and the regular changes of global
[17]. Supply chain management is very critical to economics decrease the attractiveness of poultry
have successful business [4], [10]. It deals with industry. Other factors such as competition of
major issues including global expansion and rivalry, power of buyers and also higher threats
sourcing, price fluctuation, rapid growth of from substitutes are some other concern of
multinational companies, etc. The task of company. There are three important responsibilities
implementing comprehensive idea of supply chain of the case industry:
management is not easy at all. However, if it has
been put into place, every component within the 1) Constructing the distribution network in the
chain greatly gains benefit [7], [15]. country that enables it to make overall control
of distribution process.
The objective of this research is to design a supply 2) Operations, which include several methods of
chain network of poultry oriented industry with an managing inventories and other methods that
aim of increasing benefit by reduce operating cost. help to support the overall value of the process.
According to ref. [9] designing an efficient supply 3) Marketing sector, which increases the
chain network is a challenge task in terms of even competitiveness of the company.
single economic performance measure. Apart from
economic benefit the network should also Evan though the case company is facing many
maximize value creation among the partners in the challenges as discussed above, the focus of this
network [9]. To reach this objective we studied the study will be on improving the distribution
current supply chain network used in the case network. This is due to the reason that flow of
industry and evaluate it. Thereafter, we identify the goods as customer order and then going through
constraints and the challenges that affect the supply supply, production and finally delivering it back to
chain network and then design the best routes for customer as an end product requires logistic and
distribution network. Further we implement and distribution capabilities robust enough to adapt to
evaluate the obtained network using optimization constantly changing situation [5].
software.
3. Mathematical formulation to design
The rest of the paper is structured as follows. the supply network
Section 2 describes in brief the case company under
the study. Section 3 presents and discusses Designing a supply chain network has generated
proposed mathematical model. Section 4 is closer attention in the past decades from both
dedicated to the numerical analysis. Finally, section academia and practitioner [13]. Company puts
5 gives the concluding remarks and more emphasizes on the design issue to manage
recommendations for future research works. product movements from sources to demand points
efficiently [20]. In this research the objective is to
2. Description of the case industry minimize the total transportation distance covered
by truck in the distribution network while
The case company is a big poultry industry situated distributing products to the outlets. For developing
in Oman. The industry is breeding poultry a model, it is assumed that all trucks are identical in
products, such as chicken, turkey, duck, geese for term of capacity and the planning horizon of
meat, egg and other related products. It has a delivery is for a week. Following are the index and
production capacity of 17,000 million tons of meat parameters used in the optimization model.
products annually and enjoy 25% share of Oman’s
poultry consumption. It produces and delivers the Notations
products to many wholesale and retail outlets n = number of clusters
throughout Oman. The meat items are either k = truck index (k = 1, ……….., NV )
delivered fresh or in a frozen form. One major issue
facing the company is to identify the best Parameters
distribution network based on available trucks and NV = number of trucks
with respect to fluctuating demand. Apart from that Ck= capacity of truck k
 29
Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

Qj= demand quantity at point j 4. Numerical analysis

dij = shortest distance from point i to point j 4.1 Data collection
pj= processing time at point j The data was collected from the case industry,
tij= transportation time from point i to point j
survey and through observation. The collected data
Tk= maximum time allowed for truck k
unfolds the issues and the significance of the
S = Subset of nodes
distribution network that is implemented. Survey
has been done about the case industry delivery
Binary Variable
1, if truck ݇ travel from point ݅
service in Muscat and nearby region. This survey

‫ = ݆݇݅ݔ‬൝ to point ݆.
has focused on medium hypermarkets and stores

0, otherwise.
outlets where same truck can be used to satisfy
more than one outlet due to medium demand. All
Objective function, trucks that are used by the case industry are
n n NV monitored and tracked using GPS. In addition, the
min ∑∑∑
i =1 j =1 k = 1
d ij x ijk (1 ) case industry uses a barcode system in order to
monitor the sales and the inventory level of each
Subjected to: product. This kind of information is used by the
n NV market analysis in the company and shared with
∑∑x k
ij =1 j = 2 ,....., n (2) managers and other representatives. Table 1 shows
i = 2 k =1
n NV
different types of trucks, number of trucks and
∑∑x k
ij =1 i = 2 ,....., n (3) average monthly demand in each region.
j = 2 k =1
n NV Table 1. Number and types of trucks and demand
∑x −∑x
i =1
k
ip
k =1
k
pj =0 p = 1,...., n Branches Type of Number Demand
trucks of trucks (kgs)/month
k = 1,....., NV (4) Muscat 3.5 ton 22
n n 883,152
∑∑Q x j
k
ij ≤ Ck k = 1,....., NV (5 ) 5 ton 1
i =1 j =1
n n n n 18 ton 2
∑∑ p j x ijk + ∑∑t ij x ijk ≤ T k Saham 3.5 ton 4 145,137
i =1 j =1 i =1 j =1 Salalah 3.5 ton 4 186,727
k = 1,.....,NV (6) Nizwa 3.5 ton 3 106,820
∑∑x k
ij ≤ S −1 (7 ) Al-Kamil 3.5 ton 3 107,058
i∈s j∈s

x ijk ∈ {0 ,1}
We used Google map to assign places of shops that
(8 )
need the case industry products to be supplied. In
addition, we calculated distances between all these
The objective function is as shown in Equation (1).
shops by using Google map. The time required
It minimizes the total travel distance. In the
equation if the nodes i and j is selected for truck k between each shop to another has been assumed
then the value of ‫ ݆݇݅ݔ‬will be 1, otherwise it will be taking in consideration traffic and road condition.
zero. In the model node 1 is assumed to be a The state of the roads in the region under
warehouse/ depot from where the truck will start consideration has been studied in order to assume
journey and come back after delivery. Equations the time accurately. In the initial phase we
(2) and (3) define that only one truck can reach at considered only one cluster consisting of ten shops.
each outlet in the route to satisfy demand and then These shops were selected based on their need and
come back to the main warehouse. Constraints (4) locations. A survey was conducted and data
establish that the truck entering into any node must collected in order to understand the demand and
leave it and goes to another node or the warehouse. needs of each of these shops (shown in Table 2).
Equation (5) imposes a constraint on tuck capacity.
Table 3 and Table 4 displays the distance and travel
Equation (6) shows that the sum of total processing
time respectively between shops, represented here
and transportation time should be less than or equal
to the maximum allowed time for each truck in as nodes. In the table node 1 represents depot.
each round. Constraint (7) prevent any sub tour
between nodes. Finally, Constraints (8) define the
binary requirement for the variable.
 30
Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

Table 2: Demand in the different shops in Muscat

region

No. Shop Demand

kgs
1 Warehouse 0
2 Al Gadir 600
3 Al Hadir 600
4 Masri mosque 400
5 Rawabi Magyoth Trade. &Cont. 300
6 Modern Muscat Shopping Center 700
7 Umm Ammar 600
8 Yanabea Al Ataa Modern 400
9 Al Bader 900
10 Khoudh 6th 800
11 Al Maqar 600

Table 3. Distances between Nodes (kilometer)

1 2 3 4 5 6 7 8 9 10 11
1 22.6 23 22.3 22.9 22.5 22.7 23 26.5 26.1 27.6
2 22 1.7 1.6 0.45 2.8 2 2.1 9.3 10.1 8.9
3 22.2 1.7 1.3 1.5 1.8 0.28 0.8 8.3 8.5 8
4 21.5 1.6 1.3 1.4 2 1.4 1.8 8.8 9 8.4
5 22.1 0.45 1.5 1.4 3.2 1.8 2.2 8.9 9.1 8.6
6 20.7 3.4 2.1 2 3.2 1.5 2.4 8.3 8.5 7.9
7 21.9 2 0.54 1.4 1.8 1.5 0.85 7.3 8 7
8 22.5 2.1 0.85 1.8 2.2 2.4 0.75 8.2 8.4 7.9
9 24.9 9.3 7.6 9.3 8.8 8.2 7.2 7.6 0.5 2
10 24.5 10.1 8.4 10 9.7 8.5 7.9 8.4 0.5 2.1
11 26.1 8.9 7.3 9 8.5 7.9 6.8 7.3 2 2.1

Table 4. Travel time between Nodes (minute)

1 2 3 4 5 6 7 8 9 10 11
1 32 31 29 31 27 30 32 26 25 29
2 31 7 7 2 7 8 8 16 17 15
3 30 7 4 6 5 2 2 14 14 13
4 28 7 4 6 6 5 7 15 16 14
5 29 2 6 6 8 7 8 15 16 14
6 26 9 5 6 8 4 7 14 15 13
7 29 8 2 5 7 4 3 15 15 13
8 31 8 2 7 8 7 3 14 15 13
9 26 16 14 15 15 14 15 14 2 5
10 25 17 14 16 16 15 15 15 2 5
11 29 15 13 14 14 13 13 13 5 5
 31
Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

4.2 Results Discussion Table 6. Result for 10 outlets using heuristic

method
The collected information’s was analyzed using a
Cluster Distance # iteration Run Route
mathematical model developed in section 3. For the (KM) time Truck 1 Truck 2
analysis Lingo optimization software was used to Cluster 1 90.3 33892 0.16 1-4-3-2-5-1 1-6-1
obtain the optimal route for distributing the case Cluster 2 100 19356 0.11 1-11-9-10-1 1-8-7-1
industry product. Even though 11 nodes (10 shops
+ the warehouse) were considered as discussed in
The results from the heuristic methods show that
4.1, the version of Lingo that was used could
the company needs 4 trucks to cover a total
generate optimal solution only up to 7 nodes as
distance of 190.3 kilometers. A couple of trucks are
shown in Table 5. To resolve this issue, we
assigned to 5 outlets (4, 3, 2, 5 and 6) and the rest
suggested a heuristic method.
to another 5 outlets (11, 9, 10, 8 and 7). The
Table 5. Lingo result for the test case delivery route for each truck is as shown in the last
two columns of Table 6. The route shows that the
# of # Run Num. of Total Obj. truck will return to the main warehouse after the
nodes trucks Time iterations Num. function
of (KM) delivery. This route ensures that all the demand of
variables outlet being considered in the analysis is satisfied.
3 3 0.0 162 30 46.5
6 0.0 264 48 46.5 Next, we check the percentage reduction in the
4 3 0.0 757 56 47.1 movement of trucks by the heuristic method as
6 0.01 2305 92 47.1 compared to the existing practice. Heuristic
5 3 0.01 3943 90 46.85
method found that 4 trucks are needed to cover 190
6 0.01 4700 150 46.85
km. Assuming that the truck needs at least 15
minutes to load, and the truck can only leave the
6 3 0.07 93276 132 90.05
warehouse twice, a truck will cover on average 95
6 0.11 82615 222 90.05 km per day [(190/4)*2=95]. Assuming that the
7 3 0.13 134118 182 90.8 company works 6 days a week and the truck
6 1.16 581709 308 90.8 operates 26 days per month, in total a truck will
8 3 infinity infinity 240 None cover 2470 km on average per month. Therefore,
6 infinity infinity 408 None
% reduction = [(3609-2470)/3609]*100% = 31.5%

In the above equation 3609 shows the total km

4.2.1 Heuristic Method truck travels per month on average at present. This
shows that by using the proposed method the
As the problem we are solving is NP hard in its
strongest sense, we propose a heuristic method to reduction in the distance by truck will be 31.5%
solve complex problem in a reasonable timing. which is very significant. Thus, such a significant
Divide and conquer algorithm is one type of reduction will have considerable impact on fuel
heuristic algorithms. This algorithm splits the main costs, maintenance costs, and cost of depreciation.
problem into smaller sub problems that can be
solved easily. Solutions of these small problems are
then combined to a solution of the original one. For 5. Conclusions
the data as shown in Table 3 we divide the problem
into two sub problems with each problem In this research we have developed a deterministic
consisting of 6 nodes (5 outlets and a warehouse).
linear programming (LP) model for a supply chain
In the process of dividing, the geographical
information of the outlets was considered. It means network for the poultry Industry in Oman. The
that the outlets with the least distance between developed model is tested with respect to the real
them were clustered into the same group. The result data collected from the case company, as well as,
of 11 nodes using proposed heuristic method is as through market survey. Test using Lingo
shown in Table 6. optimization software shows that medium and big
size problem cannot be solved within reasonable
computational timing. Therefore, to solve large
scale industrial problem a heuristic method is
proposed which divides the overall problem into
 32
Int. J Sup. Chain. Mgt Vol. 6, No. 2, June 2017

multiple-sub problems. The test of heuristic method Journal of Operational Research, 203(2), 283-
for medium size problem generates satisfactory 293, 2010.
result. Numerical result shows that the proposed [9] Klibi, W., & Martel, A., “The design of robust
method can reduce travel distance of up to 31.5% value-creating supply chain networks”, OR
with respect to the existing system. spectrum, 35(4), 867-903, 2013.
[10] Kumar, S. & Nigmatullin, A., “A system
The research needs further testing of heuristic dynamics analysis of food supply chains –
method for solving big size problem. Also, there is case study with non-perishable products”,
a need to validate the result of algorithm with Simulation Modelling Practice and Theory,
19(2), 2151-2168, 2011.
respect to optimal solution obtained from the
[11] Lambert, D. M., & Cooper, M. C., “Issues in
mathematical model.
supply chain management”, Industrial
marketing management, 29(1), 65-83, 2000.
References [12] Lamming, R., “Squaring lean supply with
supply chain management”, International
Journal of Operations & Production
[1] Beamon, B. M., “Supply chain design and
Management, 16(2), 183-196, 1996.
analysis: Models and methods”, International
Journal of Production Economics, 55(3), 281- [13] Levi, D. S., Kaminsky, P., & Levi, E. S.,
294, 1998. “Designing and managing the supply chain:
[2] Beamon, B. M., “Measuring supply chain Concepts, strategies, and case studies”. 3rd
edition, McGraw-Hill, 2008.
performance”, International Journal of
Operations & Production Management, 19(3), [14] Lu, D., “Fundamentals of supply chain
275 – 292, 1999. management”, Bookboon, 2011.
[3] Farahani, R.Z., Rezapour, S., Drezner, T. & [15] Mutha, A. & Pokharel, S., “Strategic network
Fallah, S., “Competitive supply chain network design for reverse logistics and
design: An overview of classifications, remanufacturing using new and old product
models, solution techniques and modules”, Computers & Industrial
applications”, Omega, 45, 92-118, 2014. Engineering, 56(1), 334-346, 2009.
[4] Georgiadis, P., Vlachos, D. & Iakovou, E., “A [16] O’Keeffe, M., “Establishing supply chain
system dynamics modeling framework for the partnerships: lessons from Australian
strategic supply chain management of food agribusiness”, Supply Chain Management:
chains”, Journal of Food Engineering, 70(3), An International Journal, 3(1), 5-9, 1998.
351-364, 2005. [17] Othman, A. A., Hassan, M. G., Ismail, M. A.,
[5] Gligor, D. M., & Holcomb, M. C., & Mohd Sharif, K. I., “Exploring the
“Understanding the role of logistics Moderating Effect of Quality Management
capabilities in achieving supply chain agility: Practices towards Supply Chain Performance
a systematic literature review”, Supply Chain in Malaysia Automotive Industry”,
Management: An International Journal, 17(4), International Journal of Supply Chain
438-453, 2012. Management, 5(4), 108-113, 2016.
[6] Gosling, J., Purvis, L., & Naim, M. M., [18] Scott, C., Lundgren, H., & Thompson, P.,
“Supply chain flexibility as a determinant of “Guide to supply chain management”.
supplier selection”, International Journal of Springer Science & Business Media, 2011.
Production Economics, 128(1), 11-21, 2010. [19] Tolone, W. J., “Virtual situation rooms:
[7] Gunasekaran, A., Patel, C. & McGaughey, connecting people across enterprises for
R.E., “A framework for supply chain supply-chain agility”, Computer-Aided
performance measurement”, International Design, 32(2), 109-117, 2000.
Journal of Production Economics, 87(3), 333 [20] Varsei, M., & Polyakovskiy, S., “Sustainable
– 347, 2004. supply chain network design: A case of the
[8] Klibi, W., Martel, A., & Guitouni, A., “The wine industry in Australia”, Omega, 66, 236-
design of robust value-creating supply chain 247, 2016.
networks: a critical review”, European