SIMULATION AS A TOOL FOR LOGISTIC PROBLEMS

SOLUTION

RADEK DYL

Institute of Transport, VŠB-TU Ostrava

[e-mail]

Streszczenie Publikacja poświęcona jest wykorzystaniu symulacji
komputerowej jako możliwego narzędzia dla rozwiązywania
logistycznych problemów. W tekście są omówione korzyści zasto-
sowania komputerowych symulacji i modelowania. W artykule
omówiono możliwe przykłady odpowiedniego softwarowego śro-
dowiska, które wytwarza podstawę własnej symulacji i opracowana
wstępów dla wytwarzania własnego modelu i wyjścia, które umoż-
liwia symulacja.

Słowa kluczowe: Symulacja, logistyka, modelowanie.

1. INTRODUCTION
Nowadays, if an organization that is interested in logistics has to be

successful, then it would adapt to continually changing conditions on
concrete market to be competitive. It must creatively and in time react on
needs of customers, increases its productivity, innovates technologies and
improves organization.

For so long logistical systems for these purposes had been evaluating,

optimizing and designing via classic mathematical methods. However,
these methods had generally been solving a given problem only locally
without consideration to many factors that own explored system
influence.

Compared to these days for solving of logistical problems simulation

methods are used.

They enable to companies easier and faster problems identification

including its solution.

Between main contributions of simulation methods, there is the model

behavior tracking instead of tracking of concrete object’s dynamic
behavior its reactions tracking on organization and technical changes. By
this access models of non-existing systems can be created and during
projection a designed system behavior can already ensures required

notions. By simulation results we can minimize a risk that springs from
changes of organizational and technological processes.

It should be ensured successful solutions’ findings of simulated

projects and innovations by accurate costs verification. It can be summary
said that greatness savings of time, human resources, energies and
financial means become by using of simulation tools.

2. PROBLEMS SOLUTION WITH COMPUTER

SIMULATION SUPPORT

The computer simulation is a tool that we use during planning,

projecting and optimizing of company processes. Using of computer
simulation we can solve complicated systems that we couldn’t solve by
using of analytical methods or the using of analytical solution would be
very simplified and so received solutions would be very inaccurate and
hence also unusable.

Using of computer simulation enables us to track a behavior of

surveyed system in real time, also in speed-up or slow-down time, it
means we can simulate in short time for example assembly-lines supplies
for one year forward. Further we can investigate individual variants of
given solution to minimize risks of wrong decisions and we can be ready
for exceptional and unexpected situations like e.g. system failures.
Simulation dramatically simplifies decision making between single
solutions and often it detects situations and states that are very badly
expectable and they would be reasons of large financial loss.

Solution of different logistical problems by using of computer

simulation offers us much more larger view of surveyed process and so it
leads to better understanding of a real system.

2.1 Realization of simulated systems
If computer simulation would be use for solution of concrete

problem, then it is naturally necessary to create a computer model of real
system. This model would be as realistic as possible, an abstraction rate
should be as little as possible and veracity of created model has to
respond to reality. Generally it can be said that a computer model’s
creation becomes from detailed analysis of surveyed system. Base on this
analysis it happens mathematical evaluating and processing of gained data
about real system. This mathematical processing is necessary for correct
implementation of gained and processed data into computer model.

For correct mathematical processing it is necessary to determine what

type of system is.

Always we go out from behavior of real system, possibly from its

components. If parameters of this behavior change in time (or by another
variable), we handle with dynamic systems. In case when it doesn’t
change in time, we handle with static systems. According to behavior
properties of real system, they could be deterministic systems. We can
exactly determine values of their state variables in each time.
Characteristic mark for deterministic systems is that by same initial
conditions the simulation results will be identical. But if real systems act
randomly, we handle with stochastic systems. These systems are always
evaluated by probability distribution.

The characteristic mark of these systems is that for same initial

conditions the simulation results wouldn’t be generally identical.
Dynamic systems can be also divided by point of view on their
parameters’ changes on time. If parameters change discreetly then we
handle with discrete systems. If parameters change continuously, we
handle with continuous systems.

Figure No. 1: Systems division

If we know what system is, then we can begin with own realization of

computer model.

It is necessary to say that it is especially suitable to use computer

simulation for dynamic systems modeling because for static systems
solution much more suitable is using of easier method that are using
classic tools of numerical mathematics.

In principle, a model creation proceeds in two stages. First we create

a conceptual model based on analysis of real system. Second we
implement it into software environment and so by this manner we create a
computer model.

It is necessary to devote big attention on acquisition and processing of

needed data because we want to have described a real system as accurate
as possible, and the believability of created model should correspond to
reality. Analysis of real system isn’t often easy, especially when we
handle systems with stochastic character. For these systems it is needed to
determine proper probability distribution that really describes a real state.
To determine of proper probability distribution for gained statistical set it
is necessary to test this distribution if it can be used for acquired data. By
suitably chosen test we obtain required parameters which are necessary
for implementation into a chosen simulation environment as is shown in
Figure 2.

Figure No. 2: Possible choosing method of probability distribution

If there are found out required data for a real system and there was

made their analysis, then follows a part in which comes to own creation
of simulation model and its implementation into computer environment.
During creation of computer model it can’t be omitted all elements of real

system and relations that are created between these elements. After
creation of simulation model it becomes the testing of this model. First
type of testing is a model verifying (validation) where we judge how well
a created model represents a real system. It means what an agreement
degree is between data, which are generating by simulation model, and
data, which are acquiring by real system. Another model testing is a
truthfulness verifying of model (verification) where we judge if a created
model satisfies specified requirements and if we are able to realize it.

Another part is a preparation of individual simulation experiments,

after this own experimenting will following. It consists in pointed changes
of individual parameters’ values to reach targets we determined during
model designing. Further an analysis is executed from outputs of
individual simulation experiments. For this we usually use methods of
mathematical statistics. After an analysis of results the last part follows
and it is a realization of an optimal variant of solution.

2.2 Using of suitable simulation environment
For solving of concrete logistical problems there are very wide-

ranging possibilities of simulation use. But in case when a company has
decided to use this possibility, then there are two solution variants. Both
variants come out from company costs that are reserved for solving of
concrete problem. Solution of this problem has to be effective from
financial point of view.

The first variant how to solve this problem is the giving this problem

to a specialized firm.

It is given preference to this possibility if it is some one-shot problem

because for this case it isn’t effective to buy some simulation software.
The second variant is a purchasing of needed simulation software that is
usually very expensive and it pays only in the case when a company will
use this software “yearly” by own employees, possibly by external
collaborators.

For problems solving from own costs, it is then needed a proper

simulation tool. Nowadays there are many products that are oriented for
concrete problems solving, or they able to solve whole spectrum of
logistical problems like e.g. Witness software. From a general point of
view it can be said that Witness environment is dedicated for creation and
visualization of simulation models. It deals with classis window
application in which a requirements’ motion of simulated system is
displayed dynamically (materials or customers), further states of indivi-
dual elements, operations and actual resources’ use are depict. Generally

we can say that Witness software provides data for models creation that is
indispensable for behavior evaluation of given system in accordance with
our chosen criteria (see Figure 3).

Figure No. 3: Graphical outputs from simulation environment Witness

3. POSSIBLE USING OF SIMULATION IN LOGISTICS
Using of computer simulation has a very wide span, so it is evident its

use also in logistics, or during designing and optimizing of logistical
systems. Logistical system is arbitrary system that has connection with
production, supplies, distribution and services offering to its customers.

We can’t forget transportation, stock-keeping, packing and other parts

of logistical chain.

The main benefit of simulation method is risk elimination during a

creation of logistical system, also during a system implementation and
during changes executing of already existing system.

Concretely we can say that the computer simulation is invoked in

transport area, especially for transport simulation in single workplaces or
in whole company. Another use is in planning and product control where
it is dealing with orders assignment to individual product parts with
respect for actual situations in workshops, including its control. It also
applies by product systems projecting or their innovation where it deals
with a design of dispositional organization of individual elements or with
requirements supply of needed capacity. There is also a use in perfection
of individual logistical conceptions (Figure 4) or in stock and resources
minimizing, in dimensioning of assembly lines, in findings of
appointment fitness of robotized systems and in personnel

implementation. Important functions are a reduction of semi-finished
production and through periods, determination of product and transport
assessments, including realignment of delivery supplies of primary
commodities and semi-finished products with production.

Figure No. 4: Model of logistical conception on resources distribution in store

4. CONCLUSION
This contribution deals with using of computer simulation in

logistics. There is introduced simulation as a possible tool for solving of
logistical problems that can provide us believable information about
possible consequences that are executed in individual logistical areas.
Solution, obtained with simulation, markedly simplifies a decision-
making between single suggested variants and often it uncovers some
situations and states that are not much predictable and they could be
reasons of great financial losses. So it is clear the using of computer
simulation for companies, interested in logistical activities, becomes
necessity, because they have to flexibly respond to hard market
requirements and so in this way, they could succeed in business.

ABSTRACT

This contribution deals with using of computer
simulation as a possible tool for logistic problems
solution. In the text there are discussed advantages
which the using of computer simulation and modelling
bring. Further it concentrates on some possible
examples of suitable development environment using
that simulation's own ground is. At least it also solves
field of inputs processing for models creation, and
outputs which simulation offers.

REFERENCES

[1] Dyl, R.: Využití simulačních metod v logistice. Ostrava, VŠB-TUO,

Workshop 2005, ISBN 80-248-0750-5

[2] Kůs, Z., Glombíková, V., Halasová, A.: Simulace výrobních

systémů. Liberec, Technická univerzita v Liberci, 2002, ISBN 80-
7083-642-3

[3] Křivý, I., Kindler, E: Simulace a modelování. Ostrava, Ostravská

univerzita, 2001, ISBN 80-7042809-0

[4] Tiller, P.: Řešení rozložení zásob ve skladu při respektování

návaznosti logistických systémů. Ostrava, VŠB-TUO, Fakulta
strojní, Institut dopravy, 2002 [Diplomová práce]

[5] http://humusoft.cz