Professional Documents
Culture Documents
DSS Test2
DSS Test2
odluka (VJEBE)
Veba 1:
Menadersko donoenje odluka
Hamurabi
1.
pokrenuti igru na
www.hamurabigame.com
(postoje verzije za iPod,
iPhone, ...)
upoznati pravila (ciljevi,
resursi, alternative)
"vladati" bar jedan ciklus
(10 godina)
pitanja:
2.
3.
4.
(Solver Add-In)
Primer:
Transportni problem
Problem optimalne organizacije
transporta robe do potroaa
Shipping Routes
S1
S2
c11
c12 c21
c13
c22
P1
P2
c23
P3
x11 =? x12=? x13 =?
x21 =? x22=? x23 =?
Hamurabi?
Problem
Organizovati transport proizvoda od proizvoaa do distributivnih
centara u blizini gradova, tako da se minimizuje cijena transporta.
Pri tome ne treba premaiti poznate mogunosti svakog od
proizvoaa i iskazane potrebe distributivnih centara svakog od
gradova.
U MS Excel primjeru:
proizvoai: S. Carolina, Tennessee, Arizona
gradovi:San Fran, Denver, Chicago, Dallas, New York
Novi problem
Tools/Solver
Unos podataka o problemu
promjenljive - (Changing cells)
ogranienja (izrazi)
ciljna promjenljiva (Target Cell)
x2
podruje
rjeenja
uslov
x1
Matrini oblik: trai se maksimum ciljne funkcije F(x1,...,xn)= cTx, uz niz uslova Axb
Trai se rjeenje, ako postoji (region dopustivih rjeenja)
Dualni problem: reformulacija problema, tako da se umjesto maksimuma funkcije trai
minimum
Jo neki pojmovi
Optimizacija - traenje optimalnog rjeenja, maksimuma ili minimuma
odreene veliine koju nazivamo cilj, koja zavisi od konanog broja
ulaznih promjenljivih, meusobno nezavisnih ili povezanih preko
jednog ili vie ogranienja.
Literatura
1. http://en.wikipedia.org/wiki/Hammurabi
2. http://www.hamurabigame.com
3. Cvetkovi D., Simi S., "Linearno programiranje", Odabrana
KRAJ
Poslovna inteligencija*
(Business Intelligence, BI)
*Izraz poslovna inteligencija je najee korieni prevod engleskog pojma business
intelligence, iako se koriste i termini poslovno izvjetavanje, poslovno istraivanje i
upravljanje poslovnim informacijama.
BI - definicije
Poslovna inteligencija predstavlja korienje svih potencijala
podataka i informacija u preduzeu radi donoenja boljih
poslovnih odluka i, u skladu sa tim, identifikaciju novih poslovnih
mogunosti.
Poslovna inteligencija kao rjeenje sadri tehnologije i
proizvode iji je cilj da obezbijede informacionu podrku kada
treba donijeti operativne i strateke poslovne odluke.
2/32
3/32
Izvjetajne aplikacije
Ad hoc upiti i izvjetavanje
Multidimenzionalna analiza
Statistike analize i Data Mining
Planiranje
BI Poslovna inteligencija
4/32
5/32
skladitenje
agregiranih,
ekstrahovanih i
filtriranih podataka
sa mogunou slojevitog,
multidimenzionalnog pristupa podacima,
radi donoenja odluka najvieg stratekog nivoa
6/32
BI Poslovna inteligencija
ETL
7/32
Transformacija
razliiti formati podataka, netane vrijednosti podataka,
nekonzistentnost primarnih kljueva, problem sinonima i
homonima, skrivena procesna logika
BI Poslovna inteligencija
OLAP
9/32
10/32
14/02/97
931359
25779
25779
899000
148000
455800
248500
46700
6580
1650
736550
736550
463000
150700
76400
45000
1450
28150
122100
11250
Poslovnica 1
Poslovnica 2
Poslovnica
3
..
..
..
..
.
Poslovnica n
Konsolidacija tabela
OLAP kocka
Poslovnice
Vrijeme
Podatak
Dvodimenzionalne
tabele
Podaci iz DW-OLTP
KljUEVI
Slog #1
Slog #2
Slog #3
Slog #4
Slog #5
Slog #6
Slog #7
Slog #8
Slog #9
12/32
Proizvod
Film
Soiva
Kamere
Film
Soiva
Film
Kamere
Soiva
Film
Region
Istok
Jug
Sever
Jug
Istok
Jug
Sever
Jug
Istok
DIMENZIJE
Mesec
Dec-01
Jan-02
Feb-02
Mar-02
Apr-02
Maj-02
Jun-02
Jul-02
Avg-02
Prodaja
240
250
690
425
300
500
125
400
800
Region
Sever
Jug
Istok
Mesec
Jan
Feb
Mar
Proizvod
Film
Soiva
PRODAJA
Kamere
Anegdota iz ivota...
Klasini primjer inteligencije dat je u anegdoti iz akog
ivota Karl Friedrich Gauss-a.
Kao prvokolac je dobio "nerjeiv" zadatak da sabere
brojeve od jedan do sto.
Gauss je vrlo brzo i elegantno doao do rjeenja:
- uvidio je da je zbir prvog i zadnjeg broja 101, drugog i
predzadnjeg takoe 101 i tako sve do zbira zadnjeg
para brojeva 50 i 51.
- s ozirom da parova ima 50, proizvod 50 puta 101 daje
rjeenje 5050.
Tehnologije u razvoju
informacionih sistema
Prof. dr Lazo Rolji
NetBeans
NetBeans je razvojno okruenje koje se uveliko oslanja na Java
programski jezik.
Osnovna prednost koju ima nad Visual Studio razvojnim
okruenjem je injenica da je besplatno i open-source.
Prua podrku za razvoj Java dekstop, mobilnih i web aplikacija kao
i HTML5 aplikacija. Prisutni su i alati koji omoguavaju koritenje
PHP-a i C/C++ programskog jezika.
Eclipse (1)
Eclipse je alternativa razvojnom okruenju
NetBeans.
Posljednja verzija Eclipse razvojnog okruenja je
Eclipse Standard 4.3.1.
Eclipse zajednica se fokusira na izgradnju otvorene
razvojne platforme sastavljene od proirivih
framework-a.
Eclipse (2)
Eclipse podrava razvoj slijedeih softverskih
rjeenja:
Java EE i web aplikacije,
C/C++ aplikacije,
Jave EE alati za razvoj aplikacija sa report
podrkom,
Eclipse Modeling alati,
Eclipse (3)
Koritenjem razliitih ekstenzija, Eclipse se moe proiri da prui
podrku za razvoj aplikacija koritenjem sljedeih programskih jezika:
C/C++,
COBOL,
Fortran,
Haskell,
Perl,
Python,
Ruby.
Softver za podrku
na stratekom nivou odluivanja
Prof. dr Lazo Rolji
informacioni sistem,
informacija,
upravljanje,
znanje,
menaderske odluke,
inteligencija,
sistem,
rjeenja.
Da mu prenese sistematsko znanje o tome ta kreator
(dizajner) SPO treba da zna.
2. ta je donoenje odluka ?
To je proces izbora izmeu dvije ili vie alternativa
odluke.
DSS je:
- interaktivni raunarom podrani sistem ili podsistem
- namijenjen da pomae donosiocu odluka da,
odluivanja.
SPO je opti pojam (termin) za svaku primjenu raunara koja
pojedincu ili grupi omoguuje da pripremi odluku.
Dokumente,
Znanje i / ili modele,
rjeenje.
izbor/dizajn modela,
rjeavanje modela
izbor odluke i
implementacija.
SPO (DSS)
ES ekspertski IS
Poslovna inteligencija
taktiki, i
strateki.
baza podataka,
generator sistema za podrku odluivanju i
korisnik.
ta su ekspertski SPO?
Ekspertski sistemi su raunarski sistemi-softver koji
od proraunavanja.
ta je poslovna inteligencija?
Poslovna inteligencija predstavlja korienje svih potencijala
poslovnih mogunosti.
Izvjetajne aplikacije
5. Planiranje
39/32
40/32
ta je OLAP?
On-Line Analytical Processing (OLAP) je proces koji
analitiarima i menaderima omoguava brz, postojan,
interaktivan pristup pogledu na informacije koje su
Anegdota iz ivota...
Klasini primjer prirodne inteligencije dat je u anegdoti iz
akog ivota Karl Friedrich Gauss-a.
Kao prvokolac je dobio "nerjeiv" zadatak da sabere brojeve
od jedan do sto.
Gauss je vrlo brzo i elegantno doao do rjeenja:
Hvala na panji !
Sadraj:
Enterprise serveri
Kreiranje BP
Normalizacija BP i ER dijagrami
Data Warehouse
OLAP
Proces odluivanja
....
Personalni raunari
Klijent / server
Raunarstvo u oblaku
Enterprise Server
(1) Enterprise Server naziva se raunar koji slui kolektivnim potrebama
jednog preduzea a ne samo jednog pojedinca, odjeljenja ili specijalizovane
aplikacije.
Istorijski, sve do nedavno i Mainframe raunari su bili Enterprise serveri, iako
se oni nisu nazivali serverima.
Kao manji, obino serveri bazirani na UNIX-u i Wintel raunari postali su bri i
bili su snabdjeveni sa programima za upravljanje u preduzeu. Oni su se
takoe nazivali serverima preduzea.
U ovoj upotrebi, Enterprise Server je i raunarski hardver i njegov glavni
softver, operativni sistem.
Primjeri su Sun Microsystem- raunari sa njihovim UNIX baziranim Solaris ili
Linux sistemima, Hewlett-Packard ( HP ), sistemi za zavretkom na Windows
2000 sistemom, i IBM-a iSeries sistemi (od kojih je najvei na zSeries 900
ranije nazivanih S / 390 ).
Enterprise server
(2) Neke kompanije koriste naziv Enterprise Server da njime opiu
"superprogram" koji radi pod operativnim sistemom na raunaru i prua
usluge za administratora sistema i za poslovne aplikacione programe i vie
specijalizovanih servera koji se pokreu na raunaru.
Prije porijekla ove upotrebe, takve usluge su se ponekad smatrale dijelom
operativnog sistema ili su dolazile u posebnim softverskim paketima.
Prvobitno, mnoge usluge koje se pruaju putem Enterprise Servera bile su
dostupne samo na IBM ili slian mejnfrejm raunarima, dok su se na manje
monim raunarima pokretale specijalizovane aplikacije.
Kako su ovi manji "server" raunari (kao to su oni iz Sun Microsistems i HP)
bivali bolje prilagoeni za poslovne (i za internet) aplikacije, paket usluga
potrebnih za upravljanje irom kompanije preimenovan je u "The
Enterprise Server. "
Vie specijalizovanih servera danas su Web server , firevall server, server
baze podataka, itd.
In the particular business modeled here, an order can have more than one
part but each part is provided by only one supplier. If we build a relation
called ORDER with all the fields included here, we would have to repeat the
name and
address of the supplier for every part on the order, even though the order
is for
parts from a single supplier. This relationship contains what are called
repeating
data groups because there can be many parts on a single order to a given
supplier.
More efficient way to arrange the data is to break down ORDER into
Once data have been captured and organized in data warehouses and
data marts, they are available for further analysis using tools for
business intelligence, which we introduced briefly in Chapter 2.
Business intelligence tools enable users to analyze data to see new
patterns, relationships, and insights that are useful for guiding decision
making.
Principal tools for business intelligence include software for database
querying and reporting, tools for multidimensional data analysis (online
analytical processing), and tools for data mining.
Slika 6. Viedimenzionalni
model podataka
Data Mining
Traditional database queries answer such questions as, How many
units of product number 403 were shipped in February 2010?
OLAP, or multidimensional analysis, supports much more complex
requests for information, such as Compare sales of product 403
relative to plan by quarter and sales region for the past two years.
With OLAP and query-oriented data analysis, users need to have a good
idea about the information for which they are looking.
What does it mean to the business to make better decisions? What is the
monetary value of improved decision making? Table 12-1 attempts to measure
the monetary value of improved decision making for a small U.S. manufacturing
firm with $280 million in annual revenue and 140 employees. The firm has
identified a number of key decisions where new system investments might
improve the quality of decision making. The table provides selected estimates
of annual value (in the form of cost savings or increased revenue) from
improved decision making in selected areas of the business.
We can see from Table 12-1 that decisions are made at all levels of the firm
and that some of these decisions are common, routine, and numerous.
Although the value of improving any single decision may be small, improving
hundreds of thousands of small decisions adds up to a large annual value for
the business.
TYPES OF DECISIONS
Chapters 1 and 2 showed that there are different levels in an
organization. Each of these levels has different information
requirements for decision support and responsibility for different
types of decisions (see Figure 12-1).
Decisions are classified as structured, semistructured, and
unstructured.
PROCES ODLUIVANJA
Making a decision is a
multistep process.
Simon (1960) described
four different stages in
decision making:
intelligence, design,
choice, and
implementation (see
Figure 12-2).
TA JE POSLOVNA INTELIGENCIJA ?
When we think of humans as intelligent beings we often refer to their
ability to take in data from their environment, understand the
meaning and significance of the information, and then act
appropriately.
Can the same be said of business firms? The answer appears to be a
qualified yes. All organizations, including business firms, do indeed
take in information from their environments, attempt to understand
the meaning of the information, and then attempt to act on the
information.
Just like human beings, some business firms do this well, and others
poorly.
Data from the business environment: Businesses must deal with both
structured and unstructured data from many different sources, including
mobile devices and the Internet. The data need to be integrated and
organized so that they can be analyzed and used by human decision makers.
Business intelligence infrastructure: The underlying foundation of business
intelligence is a powerful database system that captures all the relevant data
to operate the business. The data may be stored in transactional databases
or combined and integrated into an enterprise-data warehouse or series of
interrelated data marts.
Business analytics toolset: A set of software tools are used to analyze data
and produce reports, respond to questions posed by managers, and track the
progress of the business using key indicators of performance.
Managerial users and methods: Business intelligence hardware and software are
only as intelligent as the human beings who use them.
Managers impose order on the analysis of data using a variety of managerial
methods that define strategic business goals and specify how progress will be
measured.
These include business performance management and balanced scorecard
approaches focusing on key performance indicators and industry strategic
analyses focusing on changes in the general business environment, with special
attention to competitors.
Without strong senior management oversight, business analytics can produce a
great deal of information, reports, and online screens that focus on the wrong
matters and divert attention from the real issues.
You need to remember that, so far, only humans can ask intelligent questions.
User interface: Business people are no longer tied to their desks and
desktops.
They often learn quicker from a visual representation of data than
from a dry report with columns and rows of information.
Todays business analytics software suites emphasize visual
techniques such as dashboards and scorecards.
They also are able to deliver reports on Blackberries, iPhones, and
other mobile handhelds as well as on the firms Web portal.
BA software is adding capabilities to post information on Twitter,
Facebook, or internal social media to support decision making in an
online group setting rather than in a face-to-face meeting.
Figure 13-2 illustrates the as-is process for purchasing a book from a
physical bookstore.
Consider what happens when a customer visits a physical bookstore
Figure 13-2
KLASIFIKACIJA IS PREMA
PODRCI KOJU PRUAJU
Prof. dr Lazo Rolji
Informacioni sistemi
Transakcioni
informacioni sistemi
OLTP
OAS
Menaderski
informacioni sistemi
PCS
MIS
DSS
GDSS
EIS
BI
odluka.
Sutina u funkcionisanju ovih sistema je da izvrnim rukovodiocima na
stratekim nivoima upravljanja obezbijede dovoljno relevantnih
informacija za postavljanje bitnih pitanja u vezi sa poslovanjem
organizacije.
SISTEMI ZA PODRKU
ODLUIVANJU
Prof. dr Lazo Rolji
Pojam DSS
DSS je na modelima zasnovan skup procedura za obradu i
tumaenje informacija koje daju podrku menaderu u
donoenju odluka.
DSS je interaktivni raunarski informacioni sistem za pomo i
podrku strukturisanog, polustrukturisanog i nestrukturisanog
odluivanja.
Karakteristike DSS
Za nestrukturisane i polustrukturisane odluke,
Za menadere razliitih nivoa,
Za grupe i individue,
Za nekoliko meusobno zavisnih i/ili sekvencijelnih odluka,
Za sve faze procesa odluivanja
istraivanje,
dizajn,
izbor i
Implementaciju.
Karakteristike DSS
Za razliite procese i stilove donoenja odluka,
Adaptibilni i fleksibilni,
Lako upotrebljivi,
Unaprijeuju efektivnost odluivanja,
Lak za konstruisanje.
Vrste DSS
Budui da postoje mnogi pristupi donoenju odluka i zbog irokog
Vrste DSS
Individualni DSS (za polustrukturisane i nestrukturisane probleme),
Grupni DSS (za dati domen vie alternativa, procedurama
konvergencija miljenja i izbor alternative za zadovoljavajue
rjeenje).
Veoma vane organizacione odluke donose grupe ljudi i to naroito na
stratekom nivou. Znai, kako organizacija postaje kompleksnija, tako
sve manji broj odluka donose pojedinci. Odgovornost se distribuie na
vie uesnika i odluivanje postaje posao grupe ljudi.
1. Sistem datoteka
2. Sistem analize podataka
3. Sistem analize informacija
4. Raunovodstveni model
5. Reprezentacioni (procjenjuje posljedice pojedninih akcija)
6. Optimizacioni modeli
7. Sugestivni modeli (vre proraune koji generiu sugerisane odluke).
1. Simulacioni paketi
Od softvera iz ove klase vrijedi spomenuti:
Crystal Ball Pro (Decisioneering, od 2007. ORACLE),
Denitive Scenario 2.0 Select Edition (Denitive Software) i
Ithink Analyst 5.0 (High Performance Systems).
U zagradama su navedene softverske rme, proizvoai simulacionih
paketa.
DATA-1
(http: //www.treeage.com).
DATA-2
(http: //www.treeage.com).
Decision Grid-1
(www.decisiongrid.com)
Decision Grid-2
(www.decisiongrid.com)
Spreadsheet-based decision
support systems
Excel Solver
MS Excel Solver je dodatni alat ukljuen u MS Excel tabelarn kalkulator.
Koristi se za optimizaciju linearnim ili nelinearnim programiranjem za
izraunavanje minimuma ili maksimuma nekih varijabli.
Excel Solver je:
Optimizacijski alat kreiran kao dodatni modul u MS Excel.
Osim standardnog modula ugraenog u MS Excel instalaciju, postoji i
samostalni proizvod Solver, koji se moe koristiti iz MS Excel-a, ali i drugih
alata (npr. Matlab-a, Visual Basic-a) - proizvoa Front Line Systems.
Metode koje koristi:
linearno programiranje i kvadratno programiranje
konino (kupasto) i konveksno nelinearno programiranje
cjelobrojno programiranje
zaglaivanje (smoothing) kod nelinearnog programiranja i dr.
Management
Management problems
Decision variables
Maximize profit
Provide earliest entry into market
Minimize employee discomfort/turnover
Determine what price to use
Determine length of time tests should be run on a new product/service
Determine the responsibilities to assign to each worker
Constraints
Types of Problems
Repetitive
Standard solution methods exist
Complete automation may be feasible
One-time
No standard solutions
Rely on judgment
Automation is usually infeasible
DSS in Summary
A MANAGEMENT LEVEL COMPUTER SYSTEM
Which:
COMBINES DATA,
MODELS,
USER - FRIENDLY SOFTWARE
A Hypothetical Decision
Making Example
Is the route optimum? Are all likely users connected? What are the
possible routes?
Growth of traffic: To what extent does development of railway
depends on development of port, new town, airport, industrial area and
agricultural area?
Competition: To what extent would development of an improved road
would eliminate the need for railway?
Engineering problems: How much electricity is needed for electrical
train?
Supply problem: Where will supply of equipment and constructors
sought from?
Operational problem: With inadequate supply of local skilled workers
where will operating team be obtained from? Will foreign operating
contactors be used?
Time Scale: When to start the project and when it will be finished?
Cost: What will the total cost of project be?
Infrastructure: Will services available include: telephone, fire, water,
radio communication, hospitals, hotels and housing?
Step 2
Step 3
Decision To Abandon
Step 5
Decision To Abandon
Step 4
Decision To Abandon
Decision To Abandon
Operation Starts
Decision To Proceed
Decision To Abandon
Step 1
Step 2
Step 3
Step 4 ,5
Step 4
Step 5
Arhitektura DSS
Architecture
A Detailed Architecture
Typical Architecture
TPS: transaction
processing system
MODEL:
representation of a
problem
OLAP: on-line
analytical
processing
USER INTERFACE:
how user enters
problem & receives
answers
DSS DATABASE:
current data from
applications or
groups
DATA MINING:
technology for
finding relationships
in large data bases
for prediction
TPS
USER
INTERFACE
DSS DATA
BASE
EXTERNAL
DATA
USER
Tabele odluivanja
Prof. dr Lazo Rolji
Uslovi
AKO
ONDA
O
D
L
U
K
A
A
K
C
I
J
E
Vrijednosti
Ostvareni
uslovi
da ne
Pokazatelji
aktivnosti
X --
Aktivnosti
Funkcije
Gornji lijevi kvadrant sadri u pojedinom redu verbalno formulisane uslove opis uslova. Redovi se popunjavaju tako to se odgovara na pitanje: "koji
su uslovi".
Primjer 16: Ako sa eljeznike stanice A elimo otii do stanice E moemo oekivati da
preemo dvije ili tri liste destinacija (A do B, A do C itd.) dok ne ugledamo A do E,
koja nam pokazuje na koji peron trebamo otii.
Izbor moguih odluka i potrebne aktivnosti mogu se prikazati na sljedeoj tabeli:
Primjer 17: Navodimo jedan primjer koji ilustruje odluivanje prilikom ulaska
u zgradu, a koji se moe predstaviti sljedeom tabelom odluivanja:
vrata otvorena
vrata zakljuana
ulazimo
otvaramo vrata
zvonimo
ponovimo tabelu ulaska u zgradu
da
x
-
ne
da
x
x
da
ne
x
-
ne
ne
x
x
Diskusija rjeenja...
Uoimo da u ovom primjeru (namjerno smo tako uradili) imamo redundatnost u
navedenim situacijama (uslovima) odluivanja, pa radi toga i u pravilima
odluivanju (prvo i tree pravilo), jer kada su vrata otvorena onda i nije bitan
drugi uslov za akciju ulazimo, i naravno ako su vrata otvorena onda ne mogu
istovremeno biti i vrata zakljuana.
Mi smo u prvom pravilu odluivanja (prva kolona, prvi red, u prvom kvadrantu
tabele) za sluaj da su vrata otvorena, u ostvarenim uslovima za drugi uslov
vrata zakljuana jednom stavili znak - (kao nije bitan taj uslov), a drugi put
smo (ponovili) stavili oznaku ne u treem pravilu odluivanja (trea kolona u
drugom i treem kvadrantu), to je dovelo do ponavljanja situacije odluivanja i,
logino, do istog pravila.
Primjer 18:
Izrada rauna transportnog preduzea zavisi od tereta koji se prevozi i rastojanja
na koje se on transportuje. Ako teret robe prelazi 10 tona i rastojanje je preko 50
kilometara, raun se pravi na slijedei nain:
koliina puta transportna tarifa minus 5% rabata.
Ako transportni teret ne prelazi 10 tona, ali je udaljenost preko 50 kilometara,
od vrijednosti transportne usluge odbije se 3% rabata.
Isto ovo vai u sluaju ako udaljenost ne prelazi 50 kilometara, ali je teret
preko 10 tona.
KIBERNETSKE METODE
ANALIZE I ODLUIVANJA
Prof. dr Lazo Rolji
Kibernetika
Kibernetika je, kao i opta teorija sistema, relativno nova nauna disciplina
koja na odreeni nain sintetizuje dostignua ostalih naunih disciplina i na
osnovu takve sinteze dolazi do novih naunih spoznaja. Nastala je nezavisno
od opte teorije sistema, a njen osniva je Norbert Viner (Norbert Wiener).
Kibernetika je nauka o upravljanju i vezi u mehanizmima, organizmima i
drutvima.
Kibernetika (po V.V. Kafarovu) moe se definisati kao nauka koja se bavi
prouavanjem sistema bilo kakve prirode koji mogu primati, prenositi i
preraivati informaciju u svrhu upravljanja (voenja).
Na taj nain kibernetika sadri pojmove: informacija, prenos i prerada
informacija te upravljanje sistema. Pri tome se kibernetika uveliko slui
automatici.
Ako su, meutim, informacije koje se vraaju iza obavljene aktivnosti
sposobne da izmjene opti nain i stil djelovanja upravljanog sistema, onda
U zavisnosti od toga kako ulazni uticaji djeluju na izlazna dejstva, a naroito kakav je povratni uticaj izlaznih
dejstava na ulazne uticaje (x), razlikujemo dva osnovna oblika povratne sprege:
pozitivnu povratnu spregu, a to je ona povratna sprega u kojoj izlazna dejstva xr djeluju pozitivno na
ulazne uticaje, tj. poveavaju ih i
negativnu povratnu spregu, a to je ona u kojoj izlazna dejstva xr djeluju povratno tako da smanjuju
ulazne uticaje xi
U prvom sluaju vre se kvalitativne promjene stanja sistema. Primjere ovakvih sistema nalazimo u
organski ivotnim procesima, u rastenju elija, organa i organizama, kao i mijenjanju drutvenih sistema.
Drugi oblik povratnog dejstva predstavlja konzervativni proces.
Uslovi
O
D
AKO L
IF
U
K
A
A
K
ONDA C
THEN I
J
E
Podruje
konstelacije uslova
Podruje slijeda
aktivnosti
Aktivnosti
Vrijednosti
Ostvareni
uslovi
da ne
Pokazatelji
aktivnosti
X --
Funkcije
Diskusija rjeenja...
Primjer 18:
Izrada rauna transportnog preduzea zavisi od tereta koji
se prevozi i rastojanja na koje se on transportuje. Ako
teret robe prelazi 10 tona i rastojanje je preko 50
kilometara, raun se pravi na slijedei nain:
koliina puta transportna tarifa minus 5% rabata.
Ako transportni teret ne prelazi 10 tona, ali je
udaljenost preko 50 kilometara, od vrijednosti
transportne usluge odbije se 3% rabata.
Isto ovo vai u sluaju ako udaljenost ne prelazi 50
kilometara, ali je teret preko 10 tona.
6. ta je dijagram toka?
7. emu slue dijagrami toka?
8. ta se smatra kompleksnom situacijom odluivanja?
9. Koji su osnovni dijelovi tabele odluivanja?
10. ta se u tabeli odluivanja naziva pravilo odluivanja?
11. Koja su osnovna pravila za konstruisanje tabele odluivanja?
12. ta je konani smisao izrade tabele odluivanja?
SPO -03/6
SISTEMI ZA PODRKU
ODLUIVANJU
Prof. dr Lazo Rolji
Uvod
Why DSS?
Increasing complexity of decisions
Technology
Information:
Data, data everywhere, and not the time to think!
Number and complexity of options
Pace of change
Perceived benefits
decision quality
improved communication
cost reduction
increased productivity
time savings
improved customer and employee satisfaction
A Multidiscipline Study
It is clear that DSS belong to an environment with multidisciplinary
foundations, including (but not exclusively):
Database research,
Artificial intelligence,
Human-computer interaction,
Simulation methods,
Software engineering, and
Telecommunications.
Klasifikacija - podjela
Using the mode of assistance as the criterion, Power (2002)
differentiates five types for DSS:
communication-driven DSS,
data-driven DSS,
document-driven DSS,
knowledge-driven DSS, and
model-driven DSS.
An international development agency recommended that the iron in the area should
be mined and refined locally and melt using industries which has to be established.
The refined iron is possibly exported to Germany and Japan for car industry.
For success of project it requires supply of skilled labor. To overcome this problem a
training center has to be established to train workers by the time plant gets ready.
The development agency also recommends the fertile land in the area should be
prepared for intensive farming to provide food for the consumption of the people
working in the industry.
The railway should link the industrial area, farm and port.
Growth of traffic: To what extent does development of railway depends on development of port, new town,
airport, industrial area and agricultural area?
Competition: To what extent would development of an improved road would eliminate the need for railway?
Engineering problems: How much electricity is needed for electrical train?
Supply problem: Where will supply of equipment and constructors sought from?
Operational problem: With inadequate supply of local skilled workers where will operating team be obtained
from? Will foreign operating contactors be used?
Time Scale: When to start the project and when it will be finished?
Cost: What will the total cost of project be?
Infrastructure: Will services available include: telephone, fire, water, radio communication, hospitals, hotels
and housing?
Step 2
Step 3
Decision To Abandon
Step 5
Decision To Abandon
Step 4
Decision To Abandon
Decision To Abandon
Operation Starts
Decision To Proceed
Decision To Abandon
Step 1
The conceptual need for a project arise mainly as a result of an basement
of future requirements.
It may be made by a team of experts.
Typically a conceptual study will identify the technical solution required,
the economic merits, and acceptability of project in socio political terms.
It may require discussion with financial institutions wither or not they will
provide necessary funds.
Step 2
Assuming the decision has been made to develop the project further then a detailed
assessment will have to be made of all technical, economic and socio-political factors.
In this step the degree of uncertainty associated with each factor will begin to emerge.
An understanding of uncertainty associated with any proposal is essential for a feasible
decision making.
Step 3
If the outcome of step 2 is to proceed the project, then a tender
specification has to be prepared.
It should define, exactly what work the tender is required to do. Ideally it
has to define every thing that has to be done.
The magnitude of uncertainty associated with this stage is a reason for
possible variations in cost and duration of projects.
Before a tender specification is issued it is prudent to confirm that the
project is acceptable to regulatory authorities and that the adequate
finance is available.
The financer need to be convinced that the project is viable, that the
proposer is sound and has the experience and capability to derive the
project to a successful conclusion.
Step 4 ,5
Step 4
The first action is to decide if one of the tender should be accepted.
The tenderer should have the appropriate experience, capability and
adequate financial resources.
Step 5
Assuming all steps completed satisfactorily, a decision has to be taken
to start the project.
Even if the project starts, it might have to be stopped if the
environment it operates is changed.
Management
Management is decision making
The manager is a decision maker
Organizations are filled with decision makers at different level.
Management is considered as art: a talent acquired over years by trial-and-error.
However decision making today is becoming more complicated:
Technology / Information/Computers : increasing More alternative to choose
Structural Complexity / Competition : increasing larger cost of error
International markets / Consumerism : increasing more uncertainty about
future
Changes, Fluctuations : increasing need for quick decision
Management problems
Most management problems for which decisions are sought can be represented
by three standard elements objectives, decision variables, and constraints.
Objective
Maximize profit
Provide earliest entry into market
Minimize employee discomfort/turnover
Decision variables
Determine what price to use
Determine length of time tests should be run on a new product/service
Determine the responsibilities to assign to each worker
Constraints
Cant charge below cost
Test enough to meet minimum safety regulations
Ensure responsibilities are at most shared by two workers
Types of Problems
Structured: situations where the procedures to follow when a decision is needed can be specified in
advance
Repetitive
Standard solution methods exist
Complete automation may be feasible
Unstructured: decision situations where it is not possible to specify in advance most of the decision
procedures to follow
One-time
No standard solutions
Rely on judgment
Automation is usually infeasible
Semi-structured: decision procedures that can be pre specified, but not enough to lead to a definite
recommended decision
Some elements and/or phases of decision making process have repetitive elements
DSS in Summary
COMBINES DATA,
MODELS,
Taxonomies
Using the mode of assistance as the criterion, Power (2002)
differentiates five types for DSS:
communication-driven DSS,
data-driven DSS,
document-driven DSS,
knowledge-driven DSS, and
model-driven DSS.
Model-driven DSS
A model-driven DSS emphasizes access to and manipulation of a statistical, financial,
optimization, or simulation model. Model-driven DSS use data and parameters provided by
users to assist decision makers in analyzing a situation; they are not necessarily data
intensive. Dicodess is an example of an open source model-driven DSS generator (Gachet
2004).
Other examples:
A spread-sheet with formulas in
A statistical forecasting model
An optimum routing model
Simple file systems accessed by query and retrieval tools provides the elementary level of
functionality. Data warehouses provide additional functionality. OLAP provides highest level of
functionality.
Examples:
Accessing AMMIS data base for all maintenance Jan89-Jul94 for CH124
Accessing INTERPOL database for crimes
by .
Communication-driven DSS
A communication-driven DSS use network and comminication
technologies to faciliate collaboartion on decision making. It supports
more than one person working on a shared task.
examples include integrated tools like Microsoft's NetMeeting or
Groove (Stanhope 2002), Vide conferencing.
It is related to group decision support systems.
Document-driven DSS
A document-driven DSS uses storage and processing technologies to
document retrieval and analysis. It manages, retrieves and
manipulates unstructured information in a variety of electronic
formats.
Document database may include: Scanned documents, hypertext
documents, images, sound and video.
A search engine is a primary tool associated with document drivel
DSS.
Knowledge-driven DSS
A knowledge-driven DSS provides specialized problem solving
expertise stored as facts, rules, procedures, or in similar structures. It
suggest or recommend actions to managers.
MYCIN: A rule based reasoning program which help physicians
diagnose blood disease.
Architecture
Three fundamental components of DSS:
the database management system (DBMS),
the model management system (MBMS), and
the dialog generation and management system (DGMS).
the Data Management Component stores information (which can be further subdivided into
that derived from an organization's traditional data repositories, from external sources such
as the Internet, or from the personal insights and experiences of individual users);
the Model Management Component handles representations of events, facts, or situations
(using various kinds of models, two examples being optimization models and goal-seeking
models); and
the User Interface Management Component is of course the component that allows a user to
interact with the system.
A Detailed Architecture
Even though different authors identify different components in a DSS,
academics and practitioners have come up with a generalized
architecture made of six distinct parts:
the data management system,
the model management system,
the user(s)
Typical Architecture
TPS: transaction
processing system
MODEL: representation
of a problem
TPS
DSS DATA
BASE
EXTERNAL
DATA
USER
INTERFACE
USER
Examples:
Linear programming models,
Multiple regression forecasting models
Capital budgeting present value models
Applications
There are theoretical possibilities of building such systems in any
knowledge domain.
Clinical decision support system for medical diagnosis.
a bank loan officer verifying the credit of a loan applicant
an engineering firm that has bids on several projects and wants to
know if they can be competitive with their costs.
DSS is extensively used in business and management. Executive
dashboards and other business performance software allow faster
decision making, identification of negative trends, and better allocation
of business resources.
A growing area of DSS application, concepts, principles, and techniques
is in agricultural production, marketing for sustainable development.
A specific example concerns the Canadian National Railway system,
which tests its equipment on a regular basis using a decision support
system.
A DSS can be designed to help make decisions on the stock market, or
deciding which area or segment to market a product toward.
DSS Characteristics
(DSS In Action 1.5: Houston Minerals Case)
Decision Support
Systems
Decision
support
provided
Information
form and
frequency
Periodic, exception,
demand, and push reports
and responses
Information
format
Information
processing
methodology
Information produced by
Information produced by
extraction and manipulation analytical modeling of
of business data
business data
Definitions
DBMS - System for storing and retrieving data and processing queries
Data warehouse - Consolidated database, usually gathered from multiple primary sources,
organized and optimized for reporting and analysis
MIS - System to provide managers with summaries of decision-relevant information
Expert system - computerized system that exhibits expert-like behavior in a given problem
domain
Decision aid - automated support to help users conform to some normative ideal of rational
decision making
DSS - provide automated support for any or all aspects of the decision making process
EIS (Executive information system) - A kind of DSS specialized to the needs of top executives
Exception Reports
Reports about exceptional conditions
May be produced regularly or when exception occurs
Push Reporting
Information pushed to manager
Drill-down
Display detail data that comprise consolidated data
Data Mining
Main purpose is to provide decision support to
managers and business professionals through
knowledge discovery
Analyzes vast store of historical business data
Tries to discover patterns, trends, and correlations
hidden in the data that can help a company improve its
business performance
Use regression, decision tree, neural network, cluster
analysis, or market basket analysis
Features of an EIS
Information presented in forms tailored to the
preferences of the executives using the system
Knowledge Management
Systems
The use of information technology to help gather,
organize, and share business knowledge within an
organization
Enterprise Knowledge Portals
EIPs that are the entry to corporate intranets that serve as
knowledge management systems
Podjele SPO
Mogue je izvriti sljedeu podjelu (grupisanje) SPO:
(1) prema nainu korienja i
(2) prema funkcionalnim oblastima
(3) prema HW/SW elementima
(4) (GSPO) prema broju (i udaljenosti) uesnika u donoenju odluka
(5) SPO prema tehnikama odluivanja.
SPO:
1) SPO sredstva-alati - elementi hardvera i softvera koji
GDSS
GSPO
U analizi GSPO, pored dijaloga, modela i podataka, sree se nova
komponenta u obliku komunikacionih procedura.
Ako je SPO interaktivni raunarski bazirani sistem, koji pomae u
rjeavanju slabo strukturisanih problema, analogno tome GSPO
GSPO
lanovi grupe zajedno sa tzv. grupnim posrednikom, odnosno
medijatorom, predstavljaju grupnu komponentu GSPO.
sobe za odluivanje,
povezane sobe odluivanja (tipa telekonferencije) i
mree odluivanja.
Raunarske mree mogu biti lokalne tzv. LAN (Local Area Network) ili
iroko rasprostranjene (WAN), kad se povezuju donosioci odluka koji su
smjeteni na razliitim mjestima (daljinsko odluivanje).
funkcionalnost.
Sugestivni SPO
(Knowledge Driven DSS)
Sugestivni SPO moe preporuiti ili sugerisati akcije menaderu.
Modeli odluivanja
(Model Driven DSS)
Model - Driven SPO stavljaju akcenat na izgradnju i prouavanje modela
(statistikih, finansijskih, simulacionih i sl.). Jednostavni statistiki i
analitiki alati obezbjeuju elementarni nivo funkcionalnosti.
Donoenje odluke je jednostavno ako su razmotreni svi relevantni
faktori koji se tiu odreivanja posljedica, to je uglavnom nemogue,
pa tako ni odgovarajui model ne moe biti sveobuhvatan. U svakom
sluaju, model koji predstavlja realni sistem moe obezbijediti rezultate
koji e biti osnova za donoenje odluke.
arhitekture.
interorganizacionih SPO.
Funkcionalni SPO
(Function Specific DSS)
Funkcionalno-specifini SPO predstavljaju sisteme koji su specijalno
dizajnirani da podre odreene poslovne funkcije ili tipove poslovanja. To
mogu biti sistemi za podrku neke funkcionalne oblasti u poslovnom
sistemu, kao to su marketing ili finansije.
Ovi sistemi se obino dizajniraju za jedan specifian zadatak, kao to je,
raspored letenja u avio kompaniji i sl.
Bolje komuniciranje;
Vea produktivnost kreativnog rada strunjaka;
Prednosti SPO
Prednosti SPO zavise od prirode donosioca odluka i situacije po
izvri,
Mogunosti SPO su ograniene mogunostima raunara (hardvera i
Slika 5. SPO
Ekspertni sistemi
Expert Systems
ES
A knowledge-based information system (KBIS) that uses
its knowledge about a specific, complex application to
act as an expert consultant to end users
KBIS is a system that adds a knowledge base to the
other components on an IS
Software Resources
Inference engine processes the knowledge and makes inferences
to make recommend course of action
User interface programs to communicate with end user
Explanation programs to explain the reasoning process to end user
Using DSS
What-if Analysis
End user makes changes to variables, or relationships among
variables, and observes the resulting changes in the values of
other variables
Sensitivity Analysis
Value of only one variable is changed repeatedly and the
resulting changes in other variables are observed
Using DSS
Goal-Seeking
Set a target value for a variable and then repeatedly change
other variables until the target value is achieved
Optimization
Goal is to find the optimum value for one or more target
variables given certain constraints
One or more other variables are changed repeatedly until the
best values for the target variables are discovered
Note on DSS
Decision support systems quite literally refer to applications that are
designed to support, not replace, decision making.
Unfortunately, this is too often forgotten by decision support system
users, or these users simply equate the notion of intelligent support
of human decision making with automated decision making.
ljudskog znanja.
(Stoiljkovi, Milosavljevi. 1995).
(Stojiljkovi, 1995):
pravi ekspert iznio svoje miljenje (stvara se atmosfera poput formiranja tima eksperata).
-Savjetnike - mogu koristiti i eksperti ali i oni ostali korisnici kojima je potreban savjet u
odgovarajuim situacijama.
-Sistemi za ispitivanje tipa ta bi bilo ako...- omoguava razmatranje odreenih
situacijama u kojima je potrebno predvidjeti efekte primjene alternativnih akcija.
(Stoiljkovi, Milosavljevi, 1995).
neaktivnih pravila za rjeavanje problema i pruaju razumijevanje zakljuka koji je dao ekspertni sistem.
- Interfejs (jedinica za dijalog) - omoguuje komunikaciju korisnika sa ekspertnim sistemom putem tastature i ekrana,
ali postoje i druge varijante: putem slike, tona, mjernih signala.