Proceedings,16th IFAC Symposium on

Proceedings,16th
Proceedings,16th
Information ControlIFAC
IFAC Symposium
Symposium
Problems on
on
in Manufacturing
Proceedings,16th
Information
Information Control
ControlIFAC Symposium
Problems
Problems in
in Available online at www.sciencedirect.com
on
Manufacturing
Manufacturing
Bergamo, Italy,
Proceedings,16th June 11-13,
IFAC 2018
Symposium on
Information
Bergamo, Control
Bergamo, Italy,
Italy, June
JuneProblems
11-13,
11-13, in Manufacturing
2018
2018
Information Control
Bergamo, Italy, JuneProblems in Manufacturing
11-13, 2018
Bergamo, Italy, June 11-13, 2018 ScienceDirect
IFAC PapersOnLine 51-11 (2018) 1428–1433
The
The use
use of
of simulation
simulation in
in the
the design
design of
of Milk-Run
Milk-Run
The use of simulation in the
intralogistics
The use of simulation in the
intralogistics design
design of Milk-Run
systems
systems of Milk-Run
intralogistics systems
intralogistics systems
Kamila Kluska* Paweł Pawlewski*
Kamila
Kamila Kluska*Kluska* Paweł Paweł
 Pawlewski*
Pawlewski*
Kamila Kluska* Paweł 
 Pawlewski*
*Poznan University Kamila of Technology,
Kluska* ul.Paweł
Strzelecka 11, 60-965 Poznań, Poland
Pawlewski*
Strzelecka
*Poznan University
*Poznan University of of Technology,
Technology, ul.
ul. Strzelecka 11,
11, 60-965 Poznań, Poland
60-965 Poznań,
 [email protected]) Poland
(e-mail:
*Poznan [email protected],
University of Technology, ul. Strzelecka 11, 60-965 Poznań, Poland
(e-mail:
(e-mail: [email protected],
[email protected], [email protected])
[email protected])
*Poznan University of Technology, ul. Strzelecka 11, 60-965 Poznań, Poland
(e-mail: [email protected], [email protected])
(e-mail: [email protected], [email protected])
Abstract: This paper describes methodology of milk-run system design and implementation. The
Abstract: This
Abstract: This paper
paper describes
describes methodology
methodology of of milk-run
milk-run system
system design design and and implementation.
implementation. The The
methodology
Abstract: consists
This paper of seven basic
describes steps and of
methodology includes
milk-runthe use
system of simulation
design andtechnologyimplementation. as a toolThe to
methodology
methodology
Abstract: consists
consists
This paper of
of seven
seven
describes basic
basic steps
steps
methodology and
and includes
includes
of milk-runthe
the use
use
system of
of simulation
simulation
design and technology
technology
implementation. as
as aa tool
toolTheto
to
support
methodology design, analysis,
consists of verification
seven basic and
steps implementation
and includes theof usemilk-run
of concept.
simulation Exemplary
technology as simulation
a tool to
support
support
methodology design,
design, analysis,
analysis,
consists of verification
verification
seven basic and
and
steps implementation
implementation
and includes theof
of usemilk-run
milk-run
of concept.
concept.
simulation Exemplary
Exemplary
technology as simulation
simulation
a tool to
model
support ofdesign,
milk-run system verification
analysis, is based on and hybrid modelling approach,
implementation of usingconcept.
milk-run discrete-eventExemplary and agent-based
simulation
model
support of
model ofdesign, milk-run
milk-run system is based
system verification
analysis, is based on hybrid
on and
hybrid modelling approach,
modellingsimulation
implementation approach,
of milk-run using discrete-event
usingconcept.
discrete-eventExemplary and agent-based
andaagent-based
simulation
simulations.
model The
of milk-run approach
system is is implemented
based on hybrid in FlexSim
modellingsimulation software.
approach,software. Paper
using discrete-event contains description
andaaagent-based
simulations.
simulations. The approach
Theshowing
approach is implemented
is most
implemented in FlexSim
inelements
FlexSim Paper contains description
model
and of milk-run
illustrations
simulations. The
systemthe
approach
is
is
based on hybrid
important
implemented in
modelling
FlexSim ofsimulation
approach,
the milk-run
simulation
software.
using
system
software.
Paper
discrete-event contains
in simulation
Paper contains
and description
agent-based
model.
a description
and
and illustrations
illustrations showing
showing the
the most
most important
important elements
elements of
of the
the milk-run
milk-run
simulations. The approach is implemented in FlexSim simulation software. Paper contains a description system
system in
in simulation
simulation model.
model.
© 2018,
and
Keywords:
and IFAC
illustrations(International
showing
modelling,
illustrations showing
theFederation
simulation,most important
the most hybrid, of Automatic
important
elements
discrete
elements Control)
event the Hosting
ofsystem,
ofsystem,
milk-run
the milk-run bysystem
agent-based Elsevier
system
inLtd.
simulation, All rights
simulation
in simulation reserved.
model.
logistics.
model.
Keywords:
Keywords: modelling,
modelling, simulation,
simulation, hybrid,
hybrid, discrete
discrete event
event agent-based
system, agent-based simulation,
simulation, logistics.
logistics.
Keywords: modelling, simulation, hybrid, discrete event system, agent-based simulation, logistics.
Keywords: modelling, simulation, hybrid, discrete event system, agent-based simulation, logistics.
 Section 3 describes methodology of milk-run system design,
1. INTRODUCTION  Section
Section 33 describes
describes methodology
methodology of
of milk-run
milk-run system
system design,
1. INTRODUCTION
1. INTRODUCTION  composed
Section 3 of 7 basing
describes steps. Section
methodology of 4 discusses
milk-run system the design,
hybrid
1. INTRODUCTION composed
composed
Section
modelling
of
of 7
3 describes
approach.
basing
7 basing steps.
methodology
Section
Section
steps. Section5
4
ofpresents discusses
4 discusses
milk-run ansystem the design,
the
example
hybrid
hybrid
design, of
The key activity in most production
1. INTRODUCTION systems is handling and composed
modelling of 7
approach.basing steps.
Section Section
55 presents4 discusses
an the hybrid
The
The key
key activity
activity in
in most
most production
production systems
systems is
is handling
handling and
and modelling
composed
application approach.
of
of 7 basing
methodology Section
steps. Section
and presents
4
describes an example
discusses
frameworkexample
the hybrid
which
of
of
movement
The of materials.
key activity in most ItItproduction
involves supplying
systems is raw raw materials
handling modelling
and application approach.
of methodology Section and 5describes
presents framework
an example which of
movement
movement
The key of
of materials.
materials.
activity in most It involves
involves
production supplying
supplying
systemsproducts raw materials
materials
is handlingfrom application
modelling
supports
and supports the of methodology
approach.
model Section
building and 5
process.describes
presents
Section framework
an6 example
conclude which of
and
to workstations, collecting finished application the of methodology and describes framework which
movement
to
to of materials.
workstations,
workstations,
movement of materials.
It involves
collecting
collecting
It involves
supplying
finished
finished
supplying
raw materials
products
products
raw from
from
materials supports
application
discuss theofmodel
future model building
building
methodology
research
process.
process.
and
directions.
Section
Section
describes 66 conclude
conclude
framework whichand
and
workstations
to and transporting
workstations, collectingthem them to storage
finished locations,from
products and discuss
supportsfuture the model
research building process. Section 6 conclude and
directions.
workstations
workstations
to and
and
workstations,
collecting and
transporting
transporting
collecting
supplying them to
to storage
finished
work-in-progress storage locations,
locations,
products
materials to and discuss
and
from
next supportsfuture
discuss future
research
the model
research
directions.
building process. Section 6 conclude and
directions.
workstations
collecting
collecting and andand transporting
supplying
supplying them to storage
work-in-progress
work-in-progress locations,
materials
materials to and discuss future research directions.
to next
next
workstations and
for transporting
further them to In-plant
processing. storage locations,
logistics and
also 2. MILK-RUN LOGISTICS SYSTEMS
collecting
workstations
workstations andfor supplying
further work-in-progress
processing. In-plantmaterials
logistics to next
also 2.
2. MILK-RUN
MILK-RUN LOGISTICS LOGISTICS SYSTEMS SYSTEMS
collecting
includes andfor
collecting further
supplying processing.
and work-in-progress
returning empty In-plant logistics
materials
containers also
to from
next 2. MILK-RUN LOGISTICS SYSTEMS
workstations
includes
includes for
collecting
collecting further
and
and processing.
returning
returning In-plant
empty
empty logistics
containers
containers also The concept
from
from 2. of
MILK-RUN a milk-run is
LOGISTICS derived from the delivery
SYSTEMS
workstations
work-stations for further processing.
to designated place of empty In-plantTo
storage. logistics
make these also The The concept
concept of
of a
a milk-run
milk-run is
is derived
derived from
from the the delivery
delivery
includes
work-stations
work-stations collecting
to
to and
designated
designated returning
place
place of
of storage.
storage. containers
To
To make
make from
these
these method used in the dairy industry where tankers collect milk
includes
flows collecting
possible and and returning
efficient, many empty containers
decisions should be from method
made The concept used inofthea dairy
milk-run is derived
industry where from the
tankers collectdelivery
milk
work-stations
flows
flows possible
possible toand
anddesignated
efficient,
efficient, place
many
many of storage.
decisions
decisions To
should
should make be
be these
made
made The
from concept
several of
farmers;a milk-run
i.e., oneis derived
tanker from
combines the
loads delivery
from
work-stations
regarding what,toand designated
how much, by placewho, ofwhere,
storage. To make
when andbehow theseto frommethod used in
several the dairy
farmers; i.e.,industry
one tanker wherecombines
tankers collect
loads frommilk
flows possible howefficient, many decisions should made method
different used in
farmers thetodairy
supply industry where tankerstocollect milk
regarding
regarding
flows
move,
what,
what,and
possible
considering
how much,
much,
efficient,
the
by
by
material manywho,
who, where,
where, when
decisions
requirements when
should
in
and
and
use behow
how
made
points
to
to different
from several farmers;
farmers to i.e., oneaa tanker
supply
full
full
truckload
combinesto
truckload
the dairy
loads
the from
dairy
regarding
move,
move, what,
considering
considering how much,
the
the by
material
material who, where,
requirements
requirements when in and
use
in and
use howhow
points
points to from
producer.several The farmers;
use of i.e.,
milk-runone tanker
system combines
minimizes loadsthe from
total
regarding
and what, of
availability how much,
materials by who,
in source where,
locations.when different farmers
to producer. The useto of supply
milk-run a full
system truckload
minimizes to the the dairy
total
move,
and considering
availability
and availability of the material
materials
of materials in requirements
source
in source locations. in use points
locations. in use points distance different
distance farmers
travelled to
and supply
maximizes a full
the truckload
utilization to
of the
vehiclesdairy in
move, considering the material requirements producer.travelledThe useand of maximizes
milk-run system minimizes
the utilization the total
of vehicles in
and availability
It significantly of materials
affect in
overall sourcesystemlocations.
performance and order producer.
order to meetThe use
demand of milk-run
requirements. systemThus, minimizes
the cost is the total
reduced
and
It availability ofaffect
significantly materials in source locations. distance travelled
to meet demand and maximizes
requirements. theThus,
utilization of vehicles
the cost is reduced in
It
It
significantlycost,
manufacturing
significantly so itoverall
affect
affect
overall
is important
overall
system
systemissue
system
performance
to identify and
performance
performance
and
the cost
and
distance
cost
order and
to
and
travelled
transportation
meet demand
transportation
and maximizes
requirements.
the utilization
is more effective.
is more effective. Thus, Another
the
of vehicles
cost
Another
derivation
is reduced
derivation
in
manufacturing
manufacturing
It significantly cost,
cost, so
affect it is
is important
so service
itoverall important
system issue
issue to
to identify
identify itand
performance the
the order
of the to meet
milk-run demand
concept requirements.
is the practice Thus,
of the cost is
delivering reduced
bottles of
best option
manufacturing in terms
cost, ofof
so service and
it is important total cost. There-fore is costtheand transportation
concept is is more effective. Another derivation
best
best option
option of
manufacturing
the subject
in terms
in many
terms of
so service
cost,studiesit isand and
and
important total issue
total
improvement
cost. to identify it
There-fore
cost.projects.
issue There-fore
to identify itthe
the
is of
is cost
milk andmilk-run
to transportation
households. The is the practice
more
daily effective.of delivering
Another
route ofofdelivering
bottles of
the milkderivation
supplier
best
the option in terms of service and total cost. There-fore it is of
milkthe milk-run
to households. concept The is the practice
daily route of the milk bottles
supplier of
bestsubject
the option of
subject in many
of termsstudies
many studies
of serviceand
and and improvement
total cost.projects.
improvement There-fore it is of
projects. the milk-run
involves two
milk to households.
concept
types ofThe
is the practice
activities
daily route- of delivering
delivering
of the full full bottles
bottles
milkbottles
supplier
of
the
Theresubject
are of
many many studies
approaches and
for improvement
design and projects.
organize in-plant involves
milk to two types
households. ofThe activities
daily - delivering
route of the milk supplier of
the subject of many studies and improvement to homes and collecting
projects.in-plant involves two types of activities - delivering full bottles ofempty bottles.
There
There
logisticsare
are many
many approaches
approaches for
for design
design and
and organize
organize in-plant milk to homes and collecting empty bottles.
There
logisticsaresystems. For lean production
many approaches systems,
for design systems,and organize the milk-run
in-plant involvesmilk to homes
two types of activities - delivering full bottles of
and systems
collectingwere empty bottles.
There areissystems.
logistics
system systems.
many
commonly
For
For lean
lean
approaches
used
production
production
for design systems,
because andal-lows
it
the
thefor
organize milk-run
milk-run
in-plant
many Initially,
milk to milk-run
homes and systems
collectingwere empty used as a means to connect
bottles.
logistics systems. For lean production systems, the milk-run Initially, milk-run used as
system
system
logistics
benefits,
is
is commonly
commonly
systems.
such For
as lowering
used
used
lean because
because
production
the because
it
it al-lows
al-lows
systems,
cost of purchase the
and for
for
for many
many
milk-run
service suppliers
of suppliers and
Initially, milk-run were only
systems used outside
wereoutside
used as ofaathe means
means
to connect
factory. Klenk
to connect
system
benefits,
benefits, is
such
suchcommonly
as
as lowering
lowering used the
the cost
cost of
of it
purchase
purchaseal-lowsand
and service
service many of
of Initially, and
milk-run were only
systems used
were used of
as a the
meansfactory.
to Klenk
connect
system
modes is
ofsuchcommonly
transport, used
and make because
people's it al-lows
work more for many
efficient and Galka
suppliers (2012),
and BusinessDictionary.com
wereBusinessDictionary.com
only used outside of the (2017),
factory.define
Klenk in-
benefits,
modes
modes of
ofsuch as lowering
transport,
transport, and the cost
make
andapproach
make of purchase
people's
people's work
work and service
more
more efficient
efficient of and Galka (2012), (2017), define in-
benefits,
and safer. A as lowering
milk-run the cost of purchase
involves and service
handling multiple of suppliers
plant
and Galkamilk-runand were
(2012), systems onlyasused a outside
means
BusinessDictionary.com to of the
transport factory.
(2017), materialsKlenk
define in- to
modes
and
and of transport,
safer.
safer. A
A milk-run
milk-run andapproach
make people's
approach involves
involves work more efficient
handling
handling multiple
multiple plant
and milk-run
Galka (2012), systems asinathemeans
BusinessDictionary.com to transport(2017), materials
define in- to
modes
materials of (from
transport,
many and makeinpeople's
orders) one work
transport, more
which efficient
is why various points
plant milk-run of end use factory in one run. Typically,
and safer.(from
materials
materials A milk-run
many approach
orders) in
in oneinvolves handling
transport, whichmultiple
which is why various
is why points ofsystems
endon useaas
asin
athe
means
factory to transport
in one run.materials
Typically, to
and
it safer.(from
is difficult many and
Atomilk-run
design orders)
approach
to verify. one transport,
involves handling multiple plant
milkmilk-run
avarious run
points
systems
operates
of end use in
a means
fixed
the route,
factory
tobuttransport
in in some
one run.
materials
cases it to
Typically, is
materials
it is (from
is difficult to many
to design orders) in one transport, which is why a milk run operates on a fixed route, but in some cases it is
it difficult
materials (from many and
design and to
to verify.
orders) in one transport, which is why various
verify. based
a milkon onpoints
run a fixed ofschedule.
operates
end use in(Knez
on a fixed
the factory
and in one2015)
Gajsek, run. Typically,
it
Theis difficult
objectiveto design
of this and
paper to isverify.
to provide a methodology of based a milk run a fixed
operates schedule.
on a fixed (Knezroute, but in some
and Gajsek,
route, but in some2015)cases it is
cases it is
it is difficult
The objective to ofdesign
this and
paper to isverify.
to provide a methodology of based on a fixed schedule. (Knez andmain
Gajsek, 2015) The first
The objective
milk-run system of this
designpaper is to provide
andisimplementation. a methodology
An important of A milk-run
based on a fixedsystem consists
schedule. of
(Knez fourandmain
Gajsek, elements.
2015)
The objective of design
this paper to provide a methodology of A milk-run system
milk-run
milk-run
The
part
system
system
objective
of the of design
this paper
methodology
and
andisisimplementation.
implementation.
to provide ausing
verification
An important
An simulation
important
methodology of element
A milk-run consists
system of consists
consists
of four
a set of algorithms
of four mainthatelements.
define theThe
elements. The
first
overall
first
milk-run
part
part of system
the
of the design
methodology
methodology and is implementation.
verification
is implementation.
verification using
using An important
simulation element consists of a set of algorithms that define the overall
milk-run
technology. system design and
Simulation modelling is commonly An simulation
important
used in A milk-run
structure
element of system
the work,
consists of
consists
a including
set of
of fourrules
algorithms
main
and
that
elements.
methods
define the
Theorder
for first
overall
part of the Simulation
technology.
technology. methodologymodelling
Simulation is verification
modelling is
is using simulation
commonly
commonly used
used in
in structure
element of picking
the work,
consists of a including
set of rules
algorithms and
that methods
define theforoverall
order
part
projectsof the methodology
and studies in order is verification
to analyse, using simulation
verify and evaluate collection,
structure of picking
the work,and and delivery
including processes,
rules and methods logistics train
for order
technology.
projects and Simulation
and studies in modelling
in order to
to analyse, is commonly
verify and used in collection,
and evaluate delivery processes, logistics train
projects
technology.
the concepts studies
Simulation
of logistics order
modelling
system analyse, verify
is commonly
design. Simulation evaluate
used
enables in structure
formation,
collection,
ofetc.
the The
picking
work,
and
including
second element
delivery
rules is and
the methods
processes, logistics
logistics
for order
train, the
train
projects
the and
concepts
the concepts studies
of
ofthe in
logistics
logisticsorder
system
system to analyse,
design. verify and
Simulation
design.inverify
Simulation evaluate
enables formation,
enables collection, etc. The second
picking transportation, element
and delivery processes, is the logistics
logisticstrain, the
train
projects and studies
understanding in order
dynamics toinherent
analyse, most and evaluate
production means
formation,for material that consists of a tugger
the concepts ofthe
understanding
understanding logistics
dynamicssystem design.in
inherent inSimulation
most enables means
production for etc. The second
material element that
transportation, is the logistics
consists train,
of results
a tuggerthe
the concepts
systems, ofthe
i.e. how dynamics
logistics
system system
behaviour inherent
design.
changes most time.
Simulation
over production
enables formation,and
means a number
for
etc. ofThe
material
second
trailers. The element
transportation, use ofthat is the
logistics logistics
consiststrainsof
train,
a
the
tugger in
understanding
systems,
systems, i.e.
i.e. how
how the dynamics
system
system behaviour
behaviour inherent
changes
changes in most
over
over production
time.
time. and
means a number
for of
material trailers. The
transportation, use of logistics
that consiststrainsof results
a tugger in
understanding the dynamics inherent in most production increased and a number transport
of trailers.volumes, simultaneous
The usesimultaneous
of logistics trains delivery and
resultsand in
systems,
The paperi.e.is how systemasbehaviour
organized follows. Sectionchanges2over definestime. milk-run increased and a number transport volumes, delivery
systems, i.e. how systemasbehaviour changes2over time. collection of of
increased transport goodstrailers.
andThe
volumes,
use of logistics
containers,
simultaneous
trains
and better delivery
results
handling
and
in
The
The paper
paper is
is organized
organized as follows.
follows. Section
Section 2 defines
defines milk-run
milk-run collection of goods and containers, and better handling
logistics
The systems
papersystems
is organizedin general and
as follows. discusses related
Section 2related literature.
definesliterature. increased
performance,
milk-run performance, transport
as volumes,
compared with simultaneous
more delivery
traditional and
in-plant
logistics
The papersystems
logistics is organizedin
in general
general and
and discusses
as follows. discusses
Section 2related milk-run collection of
definesliterature. of as
goods and containers,
compared with moreand better handling
traditional in-plant
logistics systems in general and discusses related literature. collection goods and containers, and better handling
logistics systems in general and discusses related literature. performance, as compared with more traditional in-plant
performance, as compared with more traditional
in-plant
2405-8963 © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Copyright © 2018 IFAC 1460
Peer review©
Copyright
Copyright ©under
2018 responsibility
2018 IFAC
IFAC of International Federation of Automatic
1460Control.
1460
10.1016/j.ifacol.2018.08.314
Copyright © 2018 IFAC 1460
Copyright © 2018 IFAC 1460
 IFAC INCOM 2018
Bergamo, Italy, June 11-13, 2018 Kamila Kluska et al. / IFAC PapersOnLine 51-11 (2018) 1428–1433 1429

logistics systems. The third element is an operator who is a from storage areas and delivering them to
multi-purpose worker, named Mizusumashi (Water Strider or production areas,
Water Spider). In addition to just driving the train, the
operator loads and unloads material onto and from the train,  Collecting finished goods from production areas;
provides safe movement through the transport network, and i.e., picking up materials from production areas and
information flow. The fourth element is the transport delivering it to finished product storage areas,
network. It consists of many connected, one-way paths for
 Supplying work-in-progress materials from
logistics trains. These paths run between all distribution and
production areas to other production areas for
collection points as well as the parking and service areas for
further processing,
logistics trains.
Milk-run logistics systems have many advantages. You and  Returning empty containers that are used to
Jiao (2014) note that implementing the milk-run concept in transport materials to their storage areas.
the express delivery industry results in improved load factors, The network complexity and numerous flows make milk-run
more efficient use of time, shorter shipment distances, systems difficult to design, model, and implement, especially
reduced logistics transportation costs, and better fulfilment of in terms of planning the transport routes and developing
transport demand. One of the most important advantages is operational rules. Chee, Chong, and Chin (2012) and Klenk
the reduction in transportation costs due to consolidated and Gal-ka (2012) identify operating rules for milk-run
distribution and collection of goods. Combining these systems, including:
activities results in shorter travel distances, lower incidences
of collisions, fewer workers performing non-value-added  Each delivery and warehouse operation is carried
transport operations, and better inventory control. Toshinori out only by a designated employee from the
et al. (2010) note that milk-run systems enable increased warehousing department (production workers are
vehicle loading rates and low levels of inventory. not involved in these processes).
Another advantage of milk-run systems is more accurate  Empty material containers from production are
goods delivery in JIT systems due to internal transport collected by logistics trains and transported to the
synchronization - Banyai (2011). The availability of the right warehouse.
amount of the right materials maximizes the efficiency of
continuous-flow manufacturing cells. It also reduces the  Kanban cards must be collected at regular intervals
space and investment required for road networks, thus and returned to the Supermarket.
reducing the total production area. Figure 1 illustrates the
flow of material among elements of a logistics system before  All materials for delivery to production are collected
and after the implementation of the milk-run concept. from a Supermarket based on the rule “take only
what you need”. Each material must be delivered to
a production cell. Materials cannot be returned to the
Supermarket.

 Kanban signals determine the delivery processes.

 All supply routes must be clearly defined and

standardized.

 Points of use, racks and work-in-process buffers

must be located along the supply routes in order to
enable for efficient deliveries.

 Each material and component supplied to the

Fig. 1. Flow of materials between elements of logistics workstations must be ready to use. Boxes and totes
system before and after implementing a milk-run system should be opened and packaging material removed.
(based on Piatkowski, 2017).
 Each material must have a specific location on a
Milk-run logistics systems are more proactive than traditional rack or storage place. Material should not be stored
approaches because they respond to and handle many on the floor.
material requests from many production areas in one
transport. While offering many operational benefits, the  All efforts to improve internal transport efficiency
approach is more complex. and to implement a milk-run system (with the use of
simulation) should be undertaken to minimize
Greenwood, Kluska, and Pawlewski (2017) identify a series forklift traffic between the warehouse and
of flows in intralogistics systems: production areas.
 Supplying production areas with materials that are 3. METHODOLOGY OF MILK-RUN SYSTEM DESIGN
needed for production; i.e., picking up raw materials AND IMPLEMENTATION

1461
IFAC INCOM 2018
1430
Bergamo, Italy, June 11-13, 2018 Kamila Kluska et al. / IFAC PapersOnLine 51-11 (2018) 1428–1433

As described in the previous section, the design and Description Material name (e.g., frame, bolt, nut,
implementation of a milk-run system is a significant yoke)
challenge. A major challenge in milk-run systems is Daily usage Maximum amount of material used in
coordinating and meeting the requirements of both a Material a day through the entire plant
Requirements Planning (MRP) system and the production Usage location Process/areas where the material is
processes. Milk-run systems must be designed to ensure used (e.g., Cell 14)
punctual deliveries of materials to every workstation in a Storage location Address (location) where the material
production area. is stored
Order frequency Frequency with which the material is
Based on experiences of authors the methodology of Milk
ordered from the supplier (e.g., daily,
Run system design and implementation is defined. This
weekly, monthly, as required)
methodology is composed of seven steps presented at Figure
Container type Packaging type of the container (e.g.,
2.
cardboard box, reusable tote, wire
basket)
Container weight Weight of an empty container
Single part Weight of one unit of material
weight
Total package Weight of a full container of material
weight
Container length Length or depth of the container
Container width Width of the container
Container height Height of the container
Usage per Number of parts required for one
assembly finished product
Hourly usage Maximum number of pieces used per
hour
Standard The number of material units in one
container quantity container
Containers used Maximum number of containers with
per hour parts required per hour
Shipment size Size of a standard shipment in days
(1 week shipment = 5 days)
# of cards in loop Number of pull signals that are in the
system
Fig. 2. Main steps of milk-run design methodology.
The second step in the process is to design the supply routes
First step in the design and implementation of a milk-run
(distribution and col-lection routes). This includes defining
system is to analyse the main production process. This
the path networks, selecting and describing the method for
includes describing of current operations in terms of tasks,
transportation, defining the number of locations of the use
responsibilities, cycle time, levelling (Heijunka process),
points and stops on the delivery routes. Racks for storage of
scheduling approach, and layout of production and
materials at each point of use must be specified and located
warehouse areas including the supermarket. An important
near to the delivery routes.
activity is the analysis of materials and components that are
transported in terms of their transport requirements and The third step in designing a milk-run system concerns the
packaging methods. A useful tool for the analysis is a “Plan information flow and management. There are two basic
For Every Part” (PFEP) database – (Harris et al. 2003). The approaches. The first uses a fixed schedule for the logistic
tool requires extensive knowledge about each of the materials train. In this case, there is a fixed time between deliveries and
and components that are transported, such as the location of varying delivery quantity. The logistics train operator collects
their use, demand quantity, external sources of supply, Kanban cards while delivering material, retrieves the needed
transport and storage requirements, and transport container amount of materials at the supermarket, and then de-livers
size and their capacity. The Table 1 shows the most common this materials to the points of use. The second type of
categories of parts information for a PFEP necessary for delivery method is based on “pull” signals from points of use
milk-run design. which results in varying times between deliveries but a fixed
quantity of transported materials. The logistics train operator
Table 1. PFEP data elements necessary for milk-run begins transport in response to inventory levels falling below
design - based on (Harris et al. 2017) trigger points and transporting certain fixed amounts of
Category Description materials to several locations. Depending on the method
Part ID Number used to identify the material chosen, a schedule for logistic train movement is defined in
in the facility

1462
 IFAC INCOM 2018
Bergamo, Italy, June 11-13, 2018 Kamila Kluska et al. / IFAC PapersOnLine 51-11 (2018) 1428–1433 1431

terms of frequency and quantity of material needed at each intelligence means the ability to complete changeable task
use point. lists (the deliveries list). In this case, an operator must have
the ability to receive and send messages related to the
The fourth step in the design milk-run system is to plan the adoption of task list, and also to send a message about the
timetable for logistics trains. It requires to standardize of all end of implementation of the task list. In the literature this
operations occurring during the work and determining the approach is also referred to as Task Driven (Beaverstock et
time of their implementation. It requires to calculate times of al. 2011).
deliveries for all parts.
ABS is a simulation technique that models the overall
The fifth step involves validation; i.e., before a milk-run behaviour of a system through the use of autonomous system
system is implemented, the project should be validated components (also referred to as agents) that communicate
through the use of simulation. These issue is crucial, so it is with each other (North and Macal, 2007). The behaviour
the subject of next section of paper. incorporated into an agent determines its role in the
The sixth step is implementing the system in the factory. environment, its interaction with other agents, its response to
messages from other agents, and indeed whether its own
The seventh, and final step, involves maintenance and behaviour is adaptable (Mustafee and Bishoff, 2011).
improvement of the sys-tem. This includes audits at the Greenwood (2016) distinguished many reasons why agent-
supermarket, along delivery routes, and at use points. based simulation approach is not widely applied in
manufacturing systems modelling. The most important is that
4. A HYBRID MODELLING APPROACH FOR manufacturing systems are complex, but composed of well-
SIMULATING MILK-RUN LOGISTICS SYSTEMS defined processes, so Discrete Event Simulation is good
approach for modelling in this case. However, some aspects
As mentioned earlier the verification with the use simulation of such systems, for example material handling and in-plant
model is crucial for finding the best solution for Milk-Run logistics operations, allows for use other modelling
project. To build the simulation model of production process paradigms, like agent-based modelling.
with deliveries based on Milk-Run, the use of simulation pro-
grams is necessary. On the market are available many Macal and North (2014) define agent-based systems as “a
software packages, but it is not possible to find the package computational frame-work for simulating dynamic processes
which support milk-run system. Therefore, the goal is to that involve autonomous agents. An autonomous agent acts
build tools which provide support for quick building of the on its own without external direction in response to situations
simulation model of milk-run system. Based on performed the agent encounters during the simulation.” Agent
projects and experiences the hybrid modelling approach was technology is necessary and applicable in modelling the
chosen for simulation milk run logistics systems. Hybrid operations of milk-run systems due to the complexity and
modelling means combination of the discrete event (DES) coordination of activities among multiple objects. Discrete
and agent based simulation (ABS). Event Simulation allows for modelling such interactions, but
an agent-based approach is more straightforward and
DES has been the main way for the process simulation of intuitive.
manufacturing and logistics for about four decades. This is
adequate for problems that consist of queuing simulations, Greenwood (2016) and Pawlewski (2015) provide examples
and a variability is represented through stochastic of combining the Agent Based Simulation and Discrete Event
distributions (Siebers et al. 2010). This approach is applicable System approaches in modelling material-handling
in simulating the manufacturing and supply chain processes. operations.
DES models are characterized by (Korytkowski et al. 2013) a
pro-cess oriented approach (the focus is on modelling the Summarizing, to model main processes (production
system in detail, not the entities). They are based on a top- processes) DES approach is recommended and for some
down modelling approach and have one thread of control logistics processes which support main production pro-cesses
(centralized). They contain passive entities (i.e. something is ABS approach is used.
done to the entities while they move through the system) and For the projects the FlexSim is commonly used, due to the
intelligence (e.g. decision making) is modelled as part of the following features (Beaverstock et al, 2011):
system. In DES, queues are the crucial element. A flow of
entities through a system is defined, macro behaviour is  ease of use of objects in a real size with drag and
modelled and input distributions are often based on drop technology,
collected/measured (objective) data. In case of Milk Run
system the process approach is insufficient.  loading an .dwg file from the layout directly to a
model,
Operator and logistics train are task executers. It means that
they have the list of tasks to do. The operator decides what he  availability of objects ASRS vehicle, Crane, Robot,
will do next based on list. So operator is an agent. Therefore Elevator,
approach based on ABS (Agent Based Systems) is used. ABS
modelling seems to be useful for modelling operators, and  extended possibilities to model conveyors,
logistics train which have their own "intelligence”, where the

1463
IFAC INCOM 2018
1432
Bergamo, Italy, June 11-13, 2018 Kamila Kluska et al. / IFAC PapersOnLine 51-11 (2018) 1428–1433

 fitting the shape of trucks and their parameters – in

real values,

 built-in experimenter tool integrated with OptQuest,

 including task sequence technology.

Flexsim Simulation Software is a powerful analytical tool,
which allows building three-dimensional computer models of
systems as well as studying and analysing their performance
at a significantly lower cost than in case of real-time
simulations. Moreover constructed model must be easy to use
and understandable for its future users, because it takes into
account all their requirements for the verification of the
designed process and also helps them to understand it.
Fig. 4. Train movement at a single milk-run path in
5. SIMULATION MODELLING OF MILK-RUN simulation model.
LOGISTICS SYSTEMS
In order to systematize and simplify the process of milk-run
The implementation of hybrid simulation in Flexsim Discrete system modelling (Greenwood et al., 2017) proposed a
Event System environment with the use of agent logic for comprehensive framework. This framework sets the
modelling work of logistics train operator in milk-run system requirements for the features of system elements that must be
is presented in Figure 3. Both a close-up of a ready trains on included into model’s representation of milk-run logistics
train-forming area and the overall system are shown. The system. It consists of three levels of modelling and captures
overall view shows a single milk-run path, trailer and tugger the: (1) basic physical objects of the system, including their
parking and the train-forming area. The proposed approach key characteristics, attributes and relationships, (2) important
for modelling milk-run logistics system leverages the operational activities that define the work executed by the
available constructs and modelling capability in FlexSim. Big objects in the system, and (3) principal decision processes
advantage of this approach is does no need to use special that drive its performance.
agent-focused software and sophisticated algorithms written
At first level the framework is considered static and defines
in other programming languages, such as C++.
relationships necessary for interactions between elements. It
also considers three classes of physical elements: the
materials that are moved between locations, by material
mover named logistics train, that travel on logistics networks.
These elements are defined based on attributes related to
location, capacity, speed, size, reliability, availability, etc. At
second level presented framework defines operational
activities that creates the sys-tem dynamics. The five primary
activities are: the travel of train on paths, train positioning,
material loading and unloading, resource availability, and the
generation of supply and demand for materials in the system.
The loading activity execution is shown at Figure 5.

Fig. 3. Simulation model of milk-run system with the use of

hybrid approach.

The material mover is a key component in any material

handling system. In a milk-run system the material mover is a
logistics train. The train consists of one tugger, one or two
operators responsible for driving the tugger, loading and
unload-ing operations, and variable number of trailers used
for carrying materials. However, the train is the most
important element in milk-run logistics system. The logistic
train in simulation model is shown in Figure 4.

Fig. 5. The loading of materials at logistics train in simulation

model.

1464
 IFAC INCOM 2018
Bergamo, Italy, June 11-13, 2018 Kamila Kluska et al. / IFAC PapersOnLine 51-11 (2018) 1428–1433 1433

The third and highest level consists of decision variables that Greenwood, A. (2016). An Approach to Represent Material
can be tactical or strategic. Decision variables drives the Handlers as Agents in Discrete-Event Simulation
system’s behaviour and performance. It is featured in terms Models, Highlights of Practical Applications of Scalable
of sophistication and complexity and includes decision- Multi-Agent Systems, pp. 98-109.
making processes. The decision variables creates the Greenwood, A., Kluska, K., Pawlewski, P. (2017). A Multi-
foundation of a transport order - the operational scheme for level Framework for Simulating Milk-Run In-Plant
moving a specified number of materials between several Logistics Operations, Bayo J., Vale Z., Hallerborg K.,
locations over a specified time span. Nathieu P., High-lights in Cyber Physical Multi-Agent
Systems, The PAAMS Collection – international
6. CONCLUSIONS AND FURTHER RESEARCH Workshop of PAAMS 2017 (2017).
Harris, R., Harris, C., Wilson, E., (2003). Making Materials
This paper presents the methodology of Milk Run system Flow: A lean material-Handling Guide for Operations,
design and implementation, consists of seven basing steps. Production-Control, and Engineering Professionals.
First step in the design and implementation of a milk-run Lean Enterprises Inst Inc.
system is to analyse the main production process with the use Klenk, E., Galka, S. (2012). Analysis of Parameters
of PFEP. The second step is to design the supply routes on Influencing In-plant Milk-run Design for Production
layout of production area. The third step in designing a milk- Supply, Technische Universität München, 12th
run system concerns the information flow and management. International Material Handling Research Colloquium.
The fourth step is to plan the timetable for logistics trains. Knez, M, Gajsek, B. (2015). Implementation of In-plant
The fifth step involves validation with the use of hybrid Milk-run Systems for Material Supply in Lean
modelling approach. The sixth step is implementing the Automotive Parts Manufacturing, International
system in the factory. The seventh and final step, involves Conference on Logistics and Sustainable Transport.
maintenance and improvement of the system. Korytkowski, P., Wisniewski, T., Rymaszewski, S. (2013).
An example of implementation of described methodology is An evolutionary simulation-based optimization approach
simulation model created with use of the hybrid approach, for dispatching scheduling, Simulation Modelling
including discrete-event and agent-based modelling. Model Practice and Theory, 35, 69–85.
was built based on the framework proposed by Greenwood et Macal, C.M., North, M.J. (2014). Introductory tutorial:
al. (2017) with the use FlexSim simulation software. The Agent-Based Modeling and Simulation, Winter
basic task executer objects (for example operator of logistics Simulation Conference, pp. 6-20.
train) are extended to include agent-based capabilities. It Mustafee, N., Bischoff, E.E. (2011). A Multi-Methodology
means that they have the list of tasks to do. The task executer Agent-Based Approach For Container Loading, Jain, S.,
decides what he will do next based on his list. Creasey, R.R., Himmelspach, J., White, K.P., Fu, M.,
Proceedings of the 2011 Winter Simulation Conference.
Since this paper provides methodology in general, future North, M.J., Macal, C.M. (2007). Managing business
work will further develop and expand the tools used in complexity: Discovering strategic solutions with agent-
several steps and verify it in project in real production based modeling and simulation, Oxford University Press:
system. New York.
Pawlewski, P. (2015). DES/ABS Approach to Simulate
Acknowledgments: The work was carried out as part of the Warehouse Operations, Highlights of Practical
POIR.01.01.01-00-0485/17 project, "Development of a new type of Applications of Agents, Multi-Agent Systems, and
logistic trolley and methods of collision-free and deadlock-free
implementation of intralogistics processes", financed by NCBiR.
Sustainability - The PAAMS Collection
Communications in Computer and Information Science,
Volume 524, pp. 115-125.
REFERENCES
Piatkowski, M. (2017). Milk run tugger deliveries marek on
Banyai, A. (2011). Comissioning Process Design of Pick-To- lean.pdf, http://thebooksout.com, 22 April 2017.
Sequence Milk Run in the Case of Just-In_Sequence Siebers, P.O., Macal, C.M., Garnett, J., Buxton, D. and Pidd,
Production, Advanced Logistic Systems, Vol.5. issue 1, M. (2010). Discrete-Event Simulation is Dead, Long
p 119-126. Live Agent-Based Simulation!. Journal of Simulation,
Beaverstock ,M., Greenwood, A., Lavery, E., Nordgren, W. 4(3), pp. 204-210.
(2011). Applied Simulation. Modeling and Analysis Toshinori Nemoto, Katsuhiko Hayashi and Masataka
using Flexsim, Flexsim Software Products, Inc., Canyon Hashimoto, (2010). Milk-Run Logistics by Japanese
Park Technology Center, Orem, USA. Automobile Manufacturers in Thailand, Procedia Social
BusinessDisctionary.com 2017. WebFinance, Inc. and Behavioral Sciences, Vol. 2, No. 3, pp 5980-5989.
www.businessdictionary.com/definitions/milk-run.html, You Zhenlai, Jiao Yang, (2014). Development and
January 31 2017. Application of Milk-Run Distribution Systems in the
Chee, S.L., Chong, M.Y., Chin, J.F. (2012). Milk-run kanban Express Industry Based on Saving Algorithm,
system for raw printed circuit board withdrawal to Mathematical Problems in Engineering, Vol. 2014, No.
surface-mounted equipment, Journal of Industrial 536459, pp 1-6.
Engineering and Management JIEM, 2012 – 5(2):382-
405, ISSN: 2013-8423.

1465