ISSN:2321-1156

International Journal of Innovative Research in Technology & Science(IJIRTS)

Application of Operation Research in Logistics and
Warehouse Optimization
Hrithik Shukla, Jainam Chhadva, Jasika Arora, Karan Sheth, Kartik Malhotra

Abstract
Logistics and warehouse problems are ill structured material are subjected, but also process of packaging,
and real- world problems. The main reason behind sorting, labelling and many others.
conducting this research is the high degree of Operations Research is also used to determine the
application of operation research required to solve density of terminal networks at warehouses, size and
these disordered problems. Operation research has a capacity of warehouses, and others.
long tradition in reducing costs and improving
solutions. Logistics and warehouse optimization is Network Optimization Problems:
commonly required in most of the industries like
Network optimization [1] is a model in operations
automobile, FMCG, container terminals, defence,
research which works like a special type of linear
etc. This paper will provide an overview of the
programming model. Networks are an imperative
inbound and outbound transportation in
subclass of linear programs that are intuitive, easy to
warehouses, problematic issues faced, the
solve, and have good integrity properties. Networks
mathematical methods used by the firms to minimize
provide a useful way to solve problems even if there are
the transportation cost, application of queueing
additional constraints or variables that prevent use of
theory, as well as the limitations of the current
networks for modelling the entire problem.
system. The first part of the paper will deal with
application of network optimization network and However, Network models have three main advantages
queueing theory to predict the waiting time of the in general:
trucks for loading/unloading purposes whereas, the
second part would deal with problems like selecting 1. They can be solved faster than linear programming
the transport route to minimize the time and cost of problems. Using Network Optimization, problems
transportation. whose linear program would have 1000 rows and
30,000 columns can be solved in a matter of seconds.
Keywords: Queueing theory, Network optimization This allows network models to be implemented in
problems and simulation. many applications (such as real-time decision making)
for which linear programming would be unsuitable.

Introduction 2. They have naturally integer solutions. By identifying

that a problem can be framed as a network program, it is
Operation Research incorporates a wide range of possible to solve special types of integer programs
mathematical problem-solving techniques and methods, without resorting to the unproductive and time-
applied in the pursuit of better decision making and consuming integer programming algorithms.
efficiency, such as simulation, mathematical
optimization and other. There is a sizeable distance 3. They are intuitive. Network models provide a
between the manufacturing location and the demand language for talking about problems that is much more
areas due to availability of factors of production such as instinctive than the variables, objective, and constraints
raw material, labour, water. So the good from various used in the language of linear and integer programming.
locations are collected and warehoused or assembled till
There are drawbacks of network models as well such as
the final product for the consumer is ready.
the fact that they cannot formulate the wide range of
Warehouse optimization is essential to transfer the models that linear and integer programs can. However,
material from the manufacturer to end user in the they occur often enough that they form a significant tool
shortest possible way, in shortest possible time in for real decision making.
requested quality and quantity at minimum costs. Thus,
A part of Network Optimization problems is the
Operations research methods can help identify and help
Minimum Cost Flow Method which is
reduce the costs and best line of implementing supply
an optimization and decision problem solving
chains. In warehouse optimization, not only the process
related to income and expenses established with the
[1] http://www.4er.org/CourseNotes/Book%20B/B-
IV.pdf

1
Application of Operation Research in Logistics and Warehouse Optimization
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
methodology used to find the cheapest possible way of A characteristic Logistics Network consists of the
sending a certain amount of flow through a flow following:
network. An example of a flow network is given later.
A typical application of this problem involves finding 1. Facilities: -
the best delivery route from a factory to a warehouse Plants/Vendors
where the road network has some capacity and cost Ports
associated. The minimum cost flow method is one of Warehouse
the most fundamental amid all flow and circulation Retailers/Distribution Centers
problem solving methods because most other problems Customers
can be cast as a minimum cost flow problem and also 2. Raw materials and finished products that flow
that it can also be solved efficiently using the network between the facilities.
simplex algorithm. Another method of solving An example of a typical logistics configuration is
minimum cost flow problems can be solved by linear provided in the diagram given below:
programming, since we optimize a linear function, and
all constraints are linear.

Logistics Network Optimization

[2]
:
The field of Operations research plays a significant role
Logistics Network Optimization. The optimization
model which aims to obtain the desired results based on
mathematical inputs and assumptions is derived from
the precise mathematical procedures gained from
operations research, offer the best or optimum solution
based on a given implemented formula. This model is
based on mathematical and statistical models only. A
strategic supply chain network with the perceptibility,
competence and flexibility to execute more Strategic Decisions to be taken while designing a
competitively results in a number of efficiencies. The typical Logistics Network, to make such decisions
primary driver for assessing a logistic network design is
companies take the help of mathematical and statistical
sometimes an effort to reduce cost, or an effort to
improve service, and often, or both. The right network model.
design optimizes transportation, distribution operations, From the perspective of warehousing, one has to
and inventory investment to strike the ideal balance of consider constraints and benefits in order to truly come
cost and service. up with an optimized logistics network.
Assuming that plants and retailer locations are fixed, we
Some benefits of Logistics Network Optimization are concentrate on the following strategic decisions in terms
given as follows: of warehouses.
1. Deciding the optimal number, location, and size of
 Reduced logistics costs warehouses,
 Improved distribution network and asset utilization 2. Determining optimum sourcing strategy; which
 Increased inventory turns plant/vendor should produce which product,
 Improved customer delivery compliance 3. Determining best distribution channels; which
 Increased carrier capacity and collaboration warehouses should service which retailers etc.?
 Improved cost to service
 Reduced cross functional waste The objective is to design or reconfigure the logistics
 Increased visibility to out of plan/network activity to network so as to minimize annual system-wide costs,
promote root cause awareness and corrective action including production/ purchasing costs, Inventory
 Supply chain resiliency – ability to survive, adapt, carrying costs, facility costs (handling and fixed costs)
and grow during turbulent change and Transportation costs.
Let us try to understand this by going through the That is, we would like to find a minimal-annual-cost
process of designing a Logistics Network for cost and configuration of the distribution network that satisfies
time optimization: product demands at specified customer service levels.

[2] An Example of a typical trade-off involved in this

http://mat.gsia.cmu.edu/classes/QUANT/NOTES/chap1 decision-making process:
1.pdf

2
INTERNATIONAL JOURNAL OF INNOVATIVE RESEARCH IN TECHNOLOGY & SCIENCE | VOLUME 5, NUMBER 5,SEPTEMBER, 2017
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
Balancing the costs of opening new warehouses with Let’s see how the Queueing theory [3] of operation
the benefits of being near the customer, quicker supply research is applied to maximize the working capacity at
advantage and healthier inventory handling advantages a given cost or minimize the warehousing cost.
that are availed by increasing the number of Queueing theory is the mathematical study of waiting
warehouses. lines or queues. This theory is basically used to predict
waiting time based on queues and to calculate the
This constraint is explained graphically in figure resources required to minimize the waiting time. When
number 2: The optimal number of warehouses is at the the requirements enter the system, it is then serviced by
lowest point in the total cost curve which represents the service lines or servers and if the lines are busy, then the
maximum number of warehouses that can be allocated requirement has to enter into the queue and will be
at the lowest costs. serviced in the order of the queue

In case of warehousing, the resources required would

be labour, conveyor belt points, offloading bays, e-
points, etc. The requirements that would have to wait in
queue would be trucks at the gates and goods at various
service points in the warehouse.

The three important components of queueing theory are:

 Arrival process
 Service Mechanism
 Queue characteristics

Arrival Process: Arrival process is how the

requirements arrive at the operation place, i.e., singly or
in groups. It is also concerned with the time intervals at
which the requirements reach the operation place.

Optimization of Warehouse processes Service Mechanism: Service mechanism involves the

service time required to perform particular task by the
Once the goods are produced, the manufacturing unit server on per unit basis. It also covers information about
sends the goods to the warehouse for the storage as well the resources required to initiate the service, the No. of
as packaging purposes. There are various operation service lines available, whether the lines are parallel or
research theories such as queueing theory, inventory in series.
model, etc. which are used in optimizing the warehouse
processes. The typical operations performed by Queue characteristics: This component determines the
warehouses are as under: way in which the customers or requirements waiting in
the line will be choose to serve next. There are
following disciplines to choose the requirement:

Receiving Put Away  First In First Out- Under this method, the
customer with the longest waiting will be
served first.
 Last In First Out- Under this method, the
customer with the shortest waiting will be
served first.
Order  Priority- The customers with high priority are
Storing served at first.
Picking

[3]
Jaroslav Masek, Juraj Camaj and Eva Nedeliakova.,
2015. “Application the Queuing Theory in the
Warehouse Optimization.” International Journal of
Packaging Dispatching Social, Behavioural, Educational, Economic, Business
and Industrial Engineering, Vol: 9.
URL:http://waset.org/publications/10002701/applicatio
n-the-queuing-theory-in-the-warehouse-optimization

3
Application of Operation Research in Logistics and Warehouse Optimization
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
 Processor sharing- Service capacity is shared goods done is also the service line of the queueing
equally between the requirements. model. The warehousing operation ends as soon as the
 Shortest job first- The upcoming job to be goods are dispatched.
served is of the smallest size.
Thus, Queueing system is suitable in the detection of
 Shortest remaining processing time- The next
information about:
job to be served is the one with the smallest
remaining processing requirement.  No. of items in stacks closer to the gates
 Average waiting time over to get serve at
In case of warehousing, the operation starts from the
each service line
point of arrival of trucks loaded with goods. The arrival
process is continuous in nature as the trucks keeps on  No. of services lines available are either
coming in the warehouses. The waiting time of the sufficient or not.
trucks depends upon the number of offloading bays the  What would happen if there is breakdown in
warehouse contains and the time required by bays in any of the service lines?
offloading each truck. The queueing characteristics Based on this information, it is possible to choose
generally followed by warehouses is FIFO which the optimal number of service lines as well as area of
next truck with the longest waiting time will be served warehouse which, in turn, would provide number of
first. Around 10% [4] of the warehousing costs are workers to be employed.
incurred at receiving stage.

Service lines in warehouse includes the offloading bays Vehicle Routing Problems
where the goods are offloaded, the E-points where the Once the goods are dispatched from the warehouse, the
details of the goods like weight, category, and other transporter or logistics department of company has to
characteristics are recorded. After the details are select the shortest or the optimum route in order to
recorded, the goods are put away by putting it on the provide deliver in the shortest possible time and at the
conveyor belts which directs the goods to the storage minimum cost. In order to select such route, vehicle
area. The goods are then again serviced by the labours routing problems are used.
at the storage area as they to classify the goods on the
Vehicle routing problems are difficult problems to find
basis order frequency before stacking. The goods with
the most feasible option, applying in areas like
maximum order frequency, i.e., utmost priority are put
transportation, logistics, communication and
near to the outgoing gates in order to minimize the
manufacturing. The most common problem called the
waiting time for loading. Therefore, in warehousing, the
capacitated vehicle routing problem (CVRP) requires
service lines are in series as next operation takes place
finding a solution to a simplified transport problem
at the end of previous operation.
where a number of identical vehicles are delivering to
When the goods are in demand, the pick-up list is given the customers from central spot, and there is a
to the employees to find the required goods from the restriction on the number of customers a vehicle can
pallets and pack the goods. The picking and packaging deliver to. The CVRP is an extension of the traveling
operations alone costs around 55% of the warehouse salesman problem.
costs as it includes small processes like searching,
These problems can be distinguished into the following
travelling, paperwork, etc. Once the goods are packed,
types:
they are put to another service line where their details
are recorded again and is sorted for identification by 1. Assigning vehicles to customers and making routes
putting barcodes, and RFID tags. when there are multiple depots, a heterogeneous set
of vehicles, multiple periods, in which customers
While loading goods, the waiting time for vehicles
have to be served more than once and split deliveries
depends upon the type of equipment used for loading
where customer is served one or more vehicles.
like hydra or is done by labours. Here, the loading of
2. When the researcher has to make sequences,
including backhauls and pickup-and-delivery where
deliveries are made customers and goods are picked
[4] Jan Karasek. “An Overview of Warehouse
up from particular spots, and orders can be received
Optimization.” European Journal of Operational
dynamically, depending on the constraints such that
Research, Vol: 203, No. 3.
goods should be delivered before others, and also
URL:http://citeseerx.ist.psu.edu/viewdoc/download?doi
where there are multiple trips, in which vehicles can
=10.1.1.665.2431&rep=rep1&type=pdf
depart from and return to the central spot more than
once.

4
INTERNATIONAL JOURNAL OF INNOVATIVE RESEARCH IN TECHNOLOGY & SCIENCE | VOLUME 5, NUMBER 5,SEPTEMBER, 2017
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
3. Problems where there are fixed sequences, like time such as direct resources (used in physical transportation
frames in which each customer has to get the goods of freight from one destination to another) and indirect
or networks with time dependent features such as resources (used in storing, handling and consolidating at
costs or travelling time. various terminals). However, as time is passing, the
conventional L&T systems that guarantee on time
VRP has many solution methods which are exact and saving, less damage, and fast delivery to customers have
heuristic in nature, even after methods are tailored for become extremely complicated and result in increasing
solving a specific problem, exact methods for VRP can costs and decision support systems. Thus descriptive
only solve problems with a constraint having customers modeling of L&T is becoming more popular in many
up to 200 only. companies.

There are three main different approaches to modelling L&T problems appropriate for simulation can be
the VRP: [5] categorised as New Design (comprising of network
designing, least cost transportation modes and route
1. Vehicle flow formulations- Using integer variables, planning), Evaluation of Alternatives Design
the solution is found. This is for basic VRP’s and (comprising of transportation mode alternatives and
it’s suitable for problems where cost can be service performance) and Refinement and Redesign of
expressed as the sum of costs related to deliveries. Existing Operations (for operational performance
This method cannot handle many practical analysis)
applications. Simulation models in L&T domains are built for
2. Commodity flow formulations- Additional integer strategic planning, tactical planning, traffic control and
variables are used representing the flow of dispatching.
commodities along the paths travelled by the
vehicles. This has only recently been used to find an The major L&T problem domains are Distribution and
exact solution. warehousing, Trucking Operations, Truck docks and
3. Set partitioning problem- These have an Ramp operations in air cargo hub, the major and most
exponential number of binary variables which are frequently arsing problem being Distribution and
each associated with a different feasible circuit. The warehousing.
VRP is then instead formulated as a set partitioning
problem which asks what is the collection of circuits Previously, distribution has been regarded as a source of
with minimum cost that satisfy the VRP constraints. cost. Then steps have been taken in order to improve the
This allows for very general route costs. This operating efficiency and cost minimisation. This has led
method required exponential number of binary to an improvement in vehicle utilisation and material
variables which are associated with different
handling systems. Production shall be planned in such a
feasibility. The problem is adjusted as a set partition
way that it takes into account marketing, distribution
problem which is solved for minimum cost
satisfying the constraints. and other activities. Logistics represents an action to
bring distribution and other related tasks to the main
stream of planning. Thus, planning effective distribution
Role of Simulation in designing strategies has become an important factor for growing
Logistics Model companies.
Due to immense competition in today’s business world,
when new products with a short life cycle are Logistics and Distribution
introduced in the markets, it becomes important for Logistics is the most important factor that affects
enterprises to invest some attention in their logistics decisions regarding distribution. The integration of
systems. Thus, it is important to see how simulation can transport, warehousing and inventory management
be used for designing, analyzing and managing logistics under the banner of distribution management is a
and transportation systems. frequent consequence of taking a logistics-based
approach. Logistics implies the need for vertical
Current Approaches integration with supply chain resulting in marketing
An L&T (Logistics and Transportation) consists of hubs regarding distribution. This implies a completely
connected by a set of traffic lanes. Thus the networks different distribution system. [6]
form hub-and-spoke arrangements between
destinations. L&T systems utilize several resources

[5] The vehicle routing problem. Philadelphia: Soc. for [6] http://www.uh.edu/~lcr3600/simulation/app-lt.html
Industrial and Applied Mathematics. 2002. ISBN 0-
89871-579-2.

5
Application of Operation Research in Logistics and Warehouse Optimization
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
Several changes in technology and demand for products Conclusion
have also lead to markets being very volatile. This has
led to problems for logistics management. This has In the process of conducting the research our group
created substantial problems for logistics management. visited some of India’s leading logistics companies. It is
Short life periods means shorter lead times, which is important to note that these companies have an
defined as the elapsed period of time from receipt of enormous amount of data to organise and have to
customer order to delivery. Also because production execute several managerial decisions in order to
lead times are shortening, through the use of new maximise productivity and minimize costs. Both of
technology, this in effect means transport and storage these processes incorporate tools of Operational
account for a greater proportion of the total Research to a heavy extent. Operations research makes
procurement-to-deliver lead-time. Thus skilful use of a number of scientific methods combined with
integration of logistics and distribution is extremely reasonable planning to solve real life logistics and
important to maximise value through customer service. warehousing problems.

As discussed in this project Logistics management

Data Requirements for a Warehouse Simulation includes planning, organization, coordination and
Model managerial decision making all of which can be
The input data required is: achieved using mathematical models that we have
 Number of plants discussed. For example: A Logistics company will need
 Number and locations of warehouses to choose from various routes to conduct the delivery of
 Number of customers a product from its manufacturing stage right to the
 Customer demand to warehouses hands of the customer; minimization of costs and
 Part numbers produced at different plants execution under the time constraint becomes of
 Bill of materials paramount importance. Functions studied under
 Transportation times Operations research such as network designing not only
help in these designing alternatives but also provide a
Thus most of this data vary over time, hence correspond simple yet effective solution to the company which can
to some probability distribution generated over time by be implemented on the ground level.
information collected over a period.
Through this project we saw numerous applications of
Simulation Outputs and Responses models from the discipline of operations research and
how they can be implemented effectively to optimally
The list of responses or outputs that a logistics user is solve logistics and warehousing problems. We started
interested in knowing about is: with Network Optimization Problems which provide a
 Average utilization simpler way of solving real time problems using
Warehouse variables and constraints used in the language of linear
Trucks and integer programming. To elucidate we discussed
Airplanes Minimum Cost Flow Method which
 Inventory levels is an optimization and decision problem solving
Production plants methodology used to find the cheapest possible way of
Warehouses
sending a certain amount of flow through a flow
 Transportation delays
network.
Between plant and warehouse
Between warehouse and customers We moved on to comprehensively summarize
 Customer orders application of the Queuing theory in the process of
Average waiting times at a warehouse warehousing process optimization. We discussed the
Number waiting at a warehouse
Arrival process, Service Mechanism and Queue
Characteristic as components of the Queuing theory and
As industries are growing, the need for freight studied how various disciplines and decisions impact
transportation will continue to increase and the L&T our decision-making process. Then we used vehicle
systems will become more complex. In order to build a routing problems to show how the optimum transport
strong system of distribution and transportation, L&T can be selected. It was followed by the simulation
companies will have to invest their time and money for method with which one easily design the whole logistics
a manageable, cost effective, scientific and clear system.
strategy. This means that applications of mathematical
modelling and numerical solution techniques such as In order to support India’s fast-paced economy, growth
simulation will continue to grow in L&T companies. of logistics industry is very essential. It is estimated that

6
INTERNATIONAL JOURNAL OF INNOVATIVE RESEARCH IN TECHNOLOGY & SCIENCE | VOLUME 5, NUMBER 5,SEPTEMBER, 2017
 ISSN:2321-1156
International Journal of Innovative Research in Technology & Science(IJIRTS)
the Indian logistics industry will continue to show 3) The vehicle routing problem. Philadelphia:
robust growth of 10-15% annually, leading the pace of Soc. for Industrial and Applied Mathematics.
growth of the economy at large. 2002. ISBN 0-89871-579-2
4) Xiang Li, 2014. “Operations Management of
References Logistics and Supply Chain: Issues and
Directions.” Discrete Dynamics in Nature and
1) Alexander Pokahr, Lars Braubach, Jan Society, Vol: 2014.
Sudeikat, Wolfgang Renz, Winfried URL:http://dx.doi.org/10.1155/2014/701938
Lamersdorf. “Simulation and implementation 5) Jan Karasek. “An Overview of Warehouse
of logistics system based on agent technology.” Optimization.” European Journal of
Link: https://vsis-www.informatik.uni- Operational Research, Vol: 203, No. 3.
hamburg.de/getDoc.php/publications/329/HIC URL:http://citeseerx.ist.psu.edu/viewdoc/down
L_paper.pdf load?doi=10.1.1.665.2431&rep=rep1&type=pd
2) Jaroslav Masek, Juraj Camaj and Eva f
Nedeliakova., 2015. “Application the Queuing
Theory in the Warehouse Optimization.” Websites:
International Journal of Social, Behavioural, 1) http://www.uh.edu/~lcr3600/simulation/app-
Educational, Economic, Business and lt.html
Industrial Engineering, Vol: 9.
2) http://www.4er.org/CourseNotes/Book%20B/B
URL:http://waset.org/publications/10002701/a -IV.pdf
pplication-the-queuing-theory-in-the-
warehouse-optimization
3) http://mat.gsia.cmu.edu/classes/QUANT/NOT
ES/chap11.pdf

7
Application of Operation Research in Logistics and Warehouse Optimization