ML20211G567

Forwards Review of Methodology for Aiding Repository Siting Decisions. Site Selection Method Presented in Clear Manner & Appears to Be Good Attempt to Deal W/Site Selection Using Subjective Data

ML20211G567
Person / Time
Issue date:	10/24/1985
From:	Knapp M NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS)
To:	Miller H NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS)
References
REF-WM-1 NUDOCS 8606200009
Download: ML20211G567 (23)

Text

r D

ISTRIBUTION:

WM sf

/by,y OCT 2 41985-WMGT rf NMSS rf RC/85/10/22 RBrowning 'MBell MKnapp MFliegel RCodell & rf

-MEMORANDUM FOR:

Hubert J. Miller,~ Chief SCoplan, RP WMRP JSmith, RM FROM:-

'Malcolm R. Knapp, Chief WMGT

SUBJECT:

REVIEW 0F DOE LETTER ON SITE. SELECTION Richard Codell of the Hydrology Section has reviewed the Report on the DOE Site-Selection Methodology. He finds that the method is presented in a clear manner, and appears to be a good attempt to deal with site selection where there are mainly subjective data available. Further efforts should be made to review the referenced literature in decision theory. Dick is in the process of obtaining decision support software.

Several reprints of software reviews and promotional literature are attached.

Ist Malcolm R. Knapp, Chief WMGT

Enclosures:

-1. Review of " Methodology for Aiding Repository Siting Decisions" 2.. Reviews and promotional literature for decision

support software I

,yiuRet I

'fl',]ct.

l Distribution:

(

ddutnWilihik

^ ~ ~ ~

,j KC if

'(

..... g _.WNET._..... :...W.MG1

--' ' 'b IFC -:

2 y._ _:_ _ f_

E RCodell:dl MKnapp

..... :.___........ :. __ _ __ __... : _ _ _.. e6062 00009 e 51024

-- - - - - - - - --- ru ' -

LATE :85/10/ 2 f : 85/10AL4 00"--i PDR v

.S a

Review of " Methodology for Aiding Repository Siting Decisions" Performed By Richard Codell I have reviewed the subject report, which was attached to a letter from Ben Rusche, DOE, to Peter Myers, National Academy of Science dated September 16, 1985. The report addresses the formal methods used by DOE in its site selection process for High Level Waste repositories.

The DOE methods are based on Multi-Attribute Utility Analysis, which consists of six steps:

1.

Identifying and organizing objectives 2.

Establishing performance measures," Single-attribute Utility Functions", and scales to measure the extent to which the site meets the objective 3.

Verifying that the objectives are independent of cach other 4.

Assessing the single-attribute utility functions 5.

Assigning scaling factors and weights 6.

Performing numerical and sensitivity analyses t

The methodology attempts to formalize subjective judgments of site suitability where there is insufficient information on which to base objective conclusions.

It attempts to differentiate between the responsibility of technical experts and policy makers.

Technical information, understanding of physical systems and expert opinion are responsibility of the experts, while evaluation of the

e.

4

+

2 site in terms of values and preferences are the responsibilities of the policy makers.

The theory's most basic tenet is that performance can be decomposed into independent sub-measures or " utilities" of the performance. The total performance U is therefore a summation of the individual utilities:

4 U = W)U)(x))+W2U I*2)+**** + N U I*n

• 2 nn i

i where U is the nth utility, W is its weight or importance, and x is the n

n n

independent variable on which u depends. Several techniques are presented to n

estimate the weight and establish the independence of utilities. An important point about utility functions is that they are not necessarily linear functions of the independent variable. For example, if utility U is defined as cumulative dose to the population in 100,000 years, and x is canister life, an j

increase from x = 1000 year to 2000 years gives a greater benefit than an increase from 10,000 yr to 11,000 years.

The other major tenet of the method is probability theory.

A measure of the utility U depends on the probabilities of the individual utilities, which are l

functions of the probabilities of the independent variables.

The measure (say, the expected value) of the utility is the utility weighted by the importance assigned each sub-utility and the probability of the independent variables.

i

3 For example, if U) is cumulative dose due to travel times x) and U2 is cumulative dose due to waste package lifetime x '

2 E(U)=[(W) U) (X)(S))) P) (S ) dS) +

W U I*2 Ib )) P IS ) dS

(

j 2

2 2

2 2

2 S

S j

2 are all scenarios likely to be encountered which would affect where 5) and 52 travel time and canister lifetime and P is the associated probability. This would be more complicated however, if S) and S2 were interrelated (i.e., P) and P were joint probabilities). Sensitivities of the measure to assumptions and 2

weights can also be evaluated at this step.

The method described above appears to be well thought-out and a useful way to formally compare sites.

I could find nothing significantly wrong with the approach based on my knowledge of probabilistic risk assessment, but I cannot claim to be an expert on either PRA or decision theory.

I am aware that the Office of Resource Management has an effort to study and use software for decision theory, I think it would be useful to determine the capabilities of NRC in this area.

It also would be useful, should you decide to pursue this matter further, to assign someone to review the public literature on decision theory.

I have not done this, but I intend to obtain and review Decision Evaluation software packageffor the IBM Personal Computer.

I include several reviews and promotions on decision support software.

bu Richard Codell

s msom.comuns Software That Helps You Understand.Why software? To recap, it's software that lets WeU, don't hold your breath... but it You can compare candidates on the B.y Michael Schrage one list, rank and analyze the criteria one does have some very good things going for individual criteria or clusters of criteria or

.u,,

ro sue,.

uses to make a decision.

it.

is tote. You can create sublevels for each of C collect-What can I tell you? I have all me faolish;can me a sucker; call Why is that important? Because it gives First and foremost, the price borders on your main criteria (for example," street me unreliable; call me anythm, g but people a chance to come to grips with how reasonable. At $145,it'sless than smarts " grades and verbal skills could be and why they make the decisions that they one-third the price of Lightyear. I subfactors in the intelligence category).

an obsessive interest in do. It offers a corr.puter-aided lens for personally would like to see a price closer DecisionMap makes it easy to skip back decision-support / decision-modeling i:itrospection. It lets you dissect ymr to $99, and forth between factors and candidates to software.

priorities. It enables you to twiddle and Secondly-and this is vital-it twiddle with ywr rankings. This isn't a Maybe I like analyzing things too much; manipulate your decision assumptions to intelligently exploits everything that is sterile Msion tree with branches of maybe I believe intellectual serendipity and see how decision outcomes might change if decent and easy to use about the Macintosh choices to make. This is very free-form.

mtuttion can be pinned to a piece of you modified your criteria just a little bit.

nterface. In other words* this is a very and the mouse and menus makes software hke a butterfly; maybe I just hke Why is that important? Too often, people assembling the DecisionMaps a breeze.

flogging a dead horse. It doesn't matter. So lay out spreadsheets, model problems and suah deM pway h ps ym see-

. Is this usefuP I suppose that's up to the how you rate ymr decision pnonties.

here I arn writing about yet another generate scenarios without really M

M I M6*t be surprised to intriguing program that offers new ways of appreciating to what extent their internal For example,let s say ya re seeking see marketing groups and sales people use exploring how people make decisions.

biases and priorities taint what should be

. decision support,in hinng an employe.

DecisionMap as a sales tool to guide Regular readers of this column have an objective analysis and decision.

With DecisionMap, you would establish the prospects through a buying suffered through my infatuation with Applications of this type of software range relative importance of five key factors say, decision-ranking the buyer's priorities, Lightyear-a decision-modeling program from hiring employes to assessing references, intelligence, ability to grow, for example, for the IBM-PC-and a subsequent mea investment criteria to computer dating.

skiU with people, and energy.

Will t sell? Don't ask me that, but this culpa column after the company failed In essence, we're talking about Unlike other decis'on programs that program is worth a look.

tbysmally in selling its product and had to nonquantitative ways of examining vital make you assign numerical values to those Softstyle, based in Honolulu, has a be rescued by Thoughtware (yet another information. Lightyear's approach wasn't choices in order of priority, DecisaonMaP toll-free number:(800) 367-5600.

software maker).

bad but I think that Softstyle's new has you use the Mac's mouse to create Another decision-support program I've That second column performed an DecisionMap (for the Macintosh) ready little bar graphs to chart how important seen is Expert Choice, for IBM-PCs. The autopsy on Lightyear's faihire (priced too could be something special, those differing criteria are.

price is stiff at $495 and it's a little too high, tefocused marketing plan, etc.) while "DecisionMap will do for decision Of course, you can go back and change decision-treeish for my taste, but some bravely suggesting that there was still a modeling what Visicalc did for financial the height of the bars-and DecisionMap users swear by it. It's from Decision future for decision-support software.

modeling " boasts Softstyle's president, automatically calculates how the move Support Software at 1300 Vincent Place What is decision-support /modehng Craig Slayter.

changes your overall choice.

McIzan, Va. 22101.

m l

m

.l

v. c

. ~..,

d.. L s s. : ".;..7. f v.. b '_f ~e. y y.5 q

_,v... s,

M w

~,- v.y '.

.,: y.c m

~

r;. i>.

.,ay

- I

,n. 3.~....m r< <

..e. :

u>..

3..; ;. - ' '

v!.

.., ^

a,{*

...c

. 3._.T, sn e p.V ; :

, =.y L..

-. ';:.. L, ?,,,a n..

.y n

J

.Y

.g-

'. %. y

- : -. I

.. cm

+.~. _.. T.

. gg*: ;

- 4..

. [

... ", g :..'

t,.q,. a,.,. S ',p

.;t,,. s 4,..

0-c--

+py.e

-4 g.

., p gl+,,J+;n,r-

1. A.-

y

. 3 s. 3.!,,e, f-:. 7,

) <;.

i

.._-'a

.. w-3 a %

e

.g uy r-

.c

,,,. 3_ _ :-

. n r.

.2:,.i; F

. ; g _ i

.s.

,.u ; _,. + _..

qc

~.-.n

- ;. -......p e, - f. ~., '

L.. _

..,.,' s A..

.: e w,,:

.~ - -..... -, --..-

. r.

y _ 4.;! :..... - : s* L & ",c,Q

~ -

i~w

^ ^ ?ky ;;f g?

f 1.w e.

.e S

W,..<'.'.s

<t' A -.

,,,.j,,,._-,

%, 7 4,

f. ;. 3,

s

,g C e.

.. - n, g,,,,.,

..,v.

,\\l,'

Y

,.,'.' f.,.. :

, I

),A-

.h _ _j *,', ] '

l

,ll,,.. '. ;- [,.?

.' ',l

• -Q

^

~ ~ I' j

.g v.

.. ~ ~, Is,, ;..9 ' w,::.. N; -.,

..,~, 6 '

(....

r s.

a,

7. c..

g.y_

r. y

.V v.

, ";..,e. ? '. (

..s t

g.

~.Q., ;.7...

'..i -,,. ' -,

6,g

.g s

s.

~.

,s..

,, 4.,

• W '.,' -

V.

, u (..

3,

a",y,[.!-

.q-l.

.7

+; '.. s.. -'.

, - ' '., ~.; y -:

• ', ' ;. ' ~

h :~

.-p.l. ' ' ' _

"I

- 't

,. - V.;

<>;? _

r J

[i V

,..c 8

c

- 4;,.; _. ~ _. s.

s, p h,_... ;.-

w-,..

,a w :s,,. - - <

.; 9.i '.. 4 2 ; ;

,d, I ",. '.. _j. " [V ' ~R.

- ~.

T.,7 W. H8f4

,s.

"J ' '

a s'j

. 8 ^

f * -n 3,'," { %,;::t.c

' D.,.

.e 1-i.;

s-i..'.*'~.,

=

. 'L:

y k, s.[E :.

l, *.. "...%,f * -

4 &_. ;>p

-h

.t..

f,,: " '

t r, -s w ' ',. s.

a s

i-n,..y -.,

.3.m... _,

u e

.l ~ ;.' N, f.

• ... ; \\ Y. f..-

f 'f 1..-

N._;

~]

,f f

!+

.,;=,. 4 - '

- * ( f,._?w, -",.;?./ --. f;

,,u,i

,.(,.., ?. :;..

{

• y? y,;. "

.. ?

..:..y

>: :,,$,e n _;... % ;,,

. y

3. '

n

..'}

'e"_

g.

V h,

. ' ye.,.. ' < 'b:,. - ~., y"; pl, >; y:.

. 'e. a.. &

- 4 y.. -

c m.

4. :

i

- - ni g*

.,k,g.. $w _.. 3= e. --

.y : j..

,y>'/

-l y.:._.;.,.

,} ",,, ;>

v l..

r y..

..q,y -

,;y--. g.,.

..g..

.,7.-,,, #

.g..g

. *, < g.g1., y,.

- #;3.,...f,

, q

. =..

,..o.

e.

.3 e

, f,s.,.:

1,f

.g

,+, -.

y g...

....- y y

,;..,,..e,v,

. E

,m A

.p

p..;

.,e i

q

.x

,g

.I o n h..

s

- r;.6

,e...,g. ' y. ;.<.,. p. e v

..,. '.., Q

.1 ',

g-'

' f 4 t,,-

.T.

ii n

s.:

n,

y

,.,1

c. 4

. - A',

Y-

,...s. g's<, J.D *. q

,a, m;

-.. g_ _.

A.-

. g.,.,

...'y D

._Q,ys.... - :- y.,1..

J y+

-p 9.,.s. g.. -

v&u.,,.,,,ph.y?

r'.

'* e

%.. V.

.4,,..

1; a, -

.*.e

~ -

r y.

3+

r.,....- +. ;

,,, 1'.. t.

-s,

., a.. g <...

,6 4

.a

,.. g.:.

,. 5- +.-

?,. 4<,.j: y,...'

c.

s

.g.. ' v...,. s

r. g..

ep

-..- q 1 s,,..

.. &.,,. q..x, d,.

q,

e.

.c s

4....' 5.,

Y

t *

,.. ?.,G y,.st.*, - -

v

~ --

~., w,

.- m..,.

s.

., =.

y, y

,4... Kj.,,.. ?..[

.c :, y y.' ;a 4.

., []. g q, w 9 e % y

-H. c.

.4g-A,.,..,..;Q -

u 9

.; a N <1, #..

u.

x,. c. f - - ;;

". (,4,3 pj. i.:.'g,y; f g; ' 4,.,,,, g '_ %,i.._,g M_ y. " '. 7f. y ;,., ;..e.

P...-

z

_ q,...1,,.,___,.

g__-.

r.

4.f.

3.t,f.. J.

.i. <

. g....

• . r.;

1 v. fg}m...; -

y

, =,.. -,

i,-

n.

.g r

., ;,.f. 3-

n... s,,

,, ~, s 4 4l ;;, &, t. ?,

+ 4 g: _3.g(.0 p e.N sg ;p j.. 6 -C',',, L ?.cd.a ' s s '

O, s.9.,.,,iy.n, c

y+,;,

'.._ c,.

sv.

M 4.

..s,

-s

.m

,.2 m

.s,.&,.-

,y L f,4,. 3._g,.. 4 ', '; W<. ~l:l,,.':.r%.v.

. uw

,3 y,. ~.;..., g :q.,

<o 4; 7. g.1,. 6.f.e

.. O,

". ) ;

.h.,

. :.u ri.

..y-... ' y..

m.

a

g.,: g...g

.c,

1

. ' 3~.l ~.) $. '. - T :. 5l,

.?. h n N - f.y U i fy

.y

$T.$ h y pl ?.kG V, & &

~

^

-f

}.

- h[,$h d' p-h..,,y'%.

.g.

. }@ '.,,,,,ae

, -Q,),j.;

4

[a e

la

,, apg 3,.

l

~

' rQN. N ',,",i

' g. ; MJ _f L. Y y-f-

.=,fs3 ' (.,N p_ O * '. g h,T';Y

(,,,' ' h I,}a - k.

-. D.,4 ??@j *Dr. 4$,.. A-p._

., 4', f 1, ],j.'j'it,;,,

} J..,

'h]

,. fp i q,?

.M[$,y.' M

['4,'.

'r i

g,[4 o_

'. 3,

, g,j.,

, u'b7J l,jg i' >_

4,7

.y

'- p'y..,,

i -b p3 :. Q,,.;,, / g, -

3 *.-'s e, y _" r %..,

_f. l -

q. '.

l '

gi q.

c.,1

,,,t c

p-p,,,,'

, j,

w.

7 c

,. c 4,

. 3

-A,,

. g 1%,

a

-r 6

e-

'E l

. i c,

j.

)

1 )*j'h.}m.. >. f

.

• l.i $- l F-$,Qh}...

. F?h n c, y,,

• y.

5

-,,..t-,.+.,.'& 4 4..,.... k.y$,(-

r

. ~.

=

?'

h-

-i q_

/ O, j ' '. ",'M i

. 3 t

b

• k.,

' I;

W.,

1. S,,g =7 m, " ' 1 ;8 i,, M....f.,,g dN,,,~

N.i s.a

{hf_. y;&:,.. i$.y,. O i i,,w.:,"..f, N, 9

• V,',' %.,s

.. N. c.. :,

~

%. Q..a gs. q f %,

g, <y#l

i,

QgQ h3g,Q-

. p, y,

v.-

m

3..,

' n

.:g.9,8,, g;. 'y;,,y;+

7,

jf$

h w

~;m w n f,

~[ { m 4,f ;

}a f w. vl l w' y,,,. { f f }

s[,ggff.

j mm g r,u

e sm,a

. m-. xys -.:

e, L

. pt, f%. ~

.Qp;OJ m y' f, g>'-

m hk ef W,: yp'.,. Q. g, p..

.,,w

. v-a

.%{ 9r,4,.$,%,,,

jn. t:

g,.9.tz

.. M.9 4.' r "Q, ye. _.,lt R

g 7 :.

y.

cs.b,

..t..,.. :

w n.

p$W Q&.

,s

..,.t_t

' i s

y a

,. 7 a

&w l &,.

k,, Y,

>:{ Q '

Ei '&'G k)WhW 'm e w. h+ 4 m?

k.

n w&a ?, :z$m# byh$y:- t

=

k Q

f*a ya, s

yma f.

.II Q % q p% b d $ y$ g f % d d @w $ @ [ m g% g $ M f ' M M M Q,.c, a.

en

j...

n x.

gz 24p.

m d

AdQ M

W

• 3

Managing RNiay3 D8Cisl083. The critical role for managers is to make decisions and communicate them. Until recently, most business people have had only spread-sheets and data bases to help them. But decisions involve more than just data. Now, Lightyear introduces a new dimension in management software, and for the first time, managers can combine quantitative and qualitative factors into an action-oriented tool-the decision model.

With UghtyBar-Ybu're the Expert! Your own expertise guides the decision-modeling process. Lightyear provides you with a framework to organize informa-tion, analyze data, and evaluate the options. Need to evaluate which computers to l

buy? Wondering which new venture your corporation should pursue? Need tojudge which loan applications are good risks? Lightyear enhances your ability to decide l

among your alternatives and provides results that you can use to communicate your decisions effectively Getting Ah8ad with UghtyBat The things you normally do to make a decision, you'll do better and faster with Lightyear. Lightyear takes your information in the same way you would express it-in words, numbers, and gut feelings-and incorpo-rates it into a decision model. Your results provide you with a visual way to explain your decision and how you arrived at it.

l USl#9 the UghtyBar SySf85. Four easy-to-learn steps is all it takes to learn how to l

build a decision model.

1

.l

. an.l$)n:. m.' w '.m'.' p'$ :..,

... ~. m;. 1 3m.h,.

c lY

.... a'

]

1

-. m.. < :.. t

..t 4.

u.

c w.-

s; w g:.

n y

A

.e.g g mh+.y?:.~., >, :,.s:s.

. h.

d:f.,y y$a,ar. a w.

._... A - : g y. s 'a -.

4. :. : ',. j.o :

s-S+ jQf YN0?STN:lM.U.Y n ? j

.. y.. ;.+. p

.,f. :Q

.. m-s<

4y

- x; 7.g' ge-v, q _ ;;. ' '- > -

?

3"p}:.,.,.%,' [ ;R':' '.. ( Q' ;

r l;fl

.]7L,

. " ~ '^Q

.gll-.

, Os Ri

.WL ;. @%

h MA MM, ml:

m.

M#e

• i

,..l

. :,C-s??;o;y..:y,

u_QVy...y[ ~_ g

- m

.. [.

. ;.y_.

'qg,,?

j.7

~

>;,3. - : M. r u.,; y.

.V '. ' e,

.cQm. E.g. s.

,,1

.,9 yee'. 'J f.?(ggppp 0, j'.

. '.a.- a.

'.. p : :...' : ;.. '. F e7*

%e.s -; n

.,. N,ay.,,.e; s.,y d,

o s-.

~

a, *4 1

5. a g.(pigs 16 *i, W.l, M.,,a.o

?

..4 pa.

c,

',nM'.,. ?

. %~

.a r.

.e

,1

..j M...,,%

. b s., g S;;.wspry,'

! ( a.a,y,.

s

%4

<J s,

/f 9>

.. - 1.;y a

3.

..g.+.

79.

..y'..

.+

.!.N.,

h l' *. ',' ' 'tg.,b,',.

,y I.

' 9 l

- d';,.. -

e'-

<,, t t WW...i ', d; (. t: 0

,,m.'.

" S

-[

'N- *i ' ^',-

j

~,, g ' '[ * ' * '[ y c :.

, 'l f (,

i. t @,0. -

'l

[. *

,I1 1

+

A. j#

(,

y i-lj, i'i

,..: E

.4.,, G *y.: [ h8 %

J' + [.. y.

,g, [

-l ) * ; _,e

('

(,g 1'; ( (,. [j/{ l'(f '

. g k:

.,. (.

.)- [ j

.i h

q - 7' j(-[

y ? %}g[.,. h1.hh

.f-Yq/-

2,.

.,. W p < p.e ;.x Q ;. :.

3 w:y o.ym

~

y;,

+-

u., 7 mA y:

y %;.,

ew

5 e.

_.;m we. -

.f :. p..;;

l A s.:, y [y w,. w.a nm' gaum y w, n ; g~. n t q ym. 9pe 1., 7.;,

j wyf wy; c.

.e<

.n. c

'~ n n*

m.

. fgg y;..g.,

g, r.

t

[,[:-

i?..h 5f[

h-id

, C[Yl.}

E-E

~~

5 [? }/ } f. [. R f'). j Q f i.Dl_':[ g.Q l3 d

.y:.M W f_[,[ [ V

[

h*,,.k}gg I.f

• c'.

?

s...."' &

G..*

' * ' ' " ' ' ' ^

• ,E

'*y'

• }.

I

. ::. -} f. 9 '.l.

a,

.l. lj.;,

ll ll,;;_ *;. ' ' ; ', 'L k

.,.g

' % f}*,

, h,, ~ l.

' *( h g

ji..

'} ftk- '. '

E

.;. 'kll.- y L.,".;l. y LA '

.,..,,,.,.g

-,ye.4.. o.

.-.; 7

., j t. 3,., -

p>

2 c...,.1

.., E 4

.s z ~

t.

g y>

..9,:..

3<

,y., 5 _.. l-Q ',_ wt l egl.;,,f y _:.L.. c.... - +

.+-

& ;, k;. '.'.

s

.c F

.~.',.4

.h

.. * \\,~.:,. :E. -

.'l' 4

s p.c 4,

L' <

....s w

.l.* : f Y, l l

' l

e..-.

,t.

u,

....x.

~

^

c

.1 * *. s.

• 3.

t.

.. y.c ~. -

~ i'l. #

}-. Y,J '., f Q.;.. ~ ' # '

..?-

N*.*'Jg

l :. L. = V,

..n

..c,

. N jw w

7 i.'

' -i :., ;

f.'. -

..#,'.., ' ' M T..l0%.,.-}f,,l. ' i

.j.

U..,

,..- V. b }t.:, 6 ~

=.

s:

g s

m u

g

- t y.... r

.% w

t...

en pa.

s'

{

s

~s.,.

t ;i.

e.v.,.,..

1

+

. 4 y

n

,?

,t

..py-ta' 3

.+.

1,

,+ _; -

.. J c<

n q

.~,..:.

, W :. -.., *

.,.y.

n ' ::.*.

...r.

3

'r

~. q-.,.,. ' ~,.

5 c..,

e y

, ~-

.i i p>,

.. w.

,i_

• [~

.[

.f.,

J _ '..',-

-4 g *. {,-. ;

w

.2 a.

q :_ g. #,.9.~.

b,.

7*

', - s.

p.,-

,., ' '(

s s ',

I',.

3'

,.(, * '. ~'., -,,... -

f.,.

+

.,g f'

o

.a gJ f

+

I

'.~.t.

, ' ',-{

T i.

.g lc,,' '.,.

w f*

i.

• g

,t g i ;.

m

.,b.

, [; '

..'d.-

M...

4

,f ! *;-(

7... '.y. ?

',4,

}3 e

p ' " ;..',e, - # 4 ',

m_

+

t..n.. $... !

1,

1. ~,

,f

,$, ' - ', f (

,,.. ' ",.,,..V.

.<g.

f.-

f_

" J.,

l

~

mpests)

.,[

m.w;w

[ p. !

+.y. p 7 y a w;r. c. im, v.

..e,,e. [g i, g:l(o

,3

'.'.4*y / m...(7 y-k.

' f y

ub.

f,,

d p9' ~ g,.

..C

{. ', j k,. l.,'

v

\\ ' [.i.

, 4 4.,,.,;

,} { [ f.., '.?.

]'(3g 7,f.

- - s..y 4

i.

,{

... {,,,.. ;. %,, ",..-:'Wf)

}') / '

,...d

'*'.I'

' ' e h..'" ~

.h.{r.

,Y

[Y

&~

.5

=-

A$f i*f'T Y, % V Y.b i

' > ; L: : (

n M? ' lI

'L'~,,,.:.0+k,. n 2

, ' 3(a. % V l k. I r

.. - ' ' N(.,. }. y,s.. g..',,.s..',

Q

'. F.,

C.,. ;. m.., ;,. ? 1 g '.,%, '.,

-u y, S.*, fM q.m.:. 4.4 gl...,.

>9,, w

= r-4

~.

,i

,. a a

.J.

._ ?-

.. )

<>qm 3'

'..N ;',, 3

'p.

a. -.,

4~

.a3

,.

• w. '-

,,,,.0,

\\...h '.. -'

. _.,., [.... i. s

.g...-.

.J t. 3.a,.

v f

. ~..

?

y.

s.:

.. - m g,.. /, -

• ,R.'<

p-4 g

• a.

g e

4

.s v

Qi

' ym. #,y.

,qgy3yg 3g g

??

' by :

}l

• %w

-]p A_,J% _4 $

.4 q

.' '., P.,, 4. - u

.- ~

3. A g

(

';f.) % W2,',

c... QM, p l,p. ? 3 4:g$ hw p(" m

?L:

47 1!.h.

.f > d i '

w

.o

,~r/h,';j.

- y

- y ;c -

l

' f%}9,gg%., y.

cr s

.k f q$%W% M.;;.

. q i U.s ' i

/.

e K

L-u Q.[

s L y/g yy.yp QC JM&,gi~phd.p&7 j -s#Dgigg fj.

~ '

. y 9 :d. u.; ;

p m Q %.

y.n

.v..

~ p s.v.r ~'.-

j. p%q-:lc 5

.e *.

.s.n :. FA{

- gg'1." W% N'm;,--.*,' _ h. g$g> q y 1

f y

4 M n &.'.

~

V,M..:-.i.

W

,;. ; y. ;q.,:

y '.- %r

1 '.-

' l.,, l ?,. '. t,

..,,;,f g ;r; ; '.nm

., gQ@,.. c h.

5 v 1 m, gJc

'4

_a

.4

'V w.,.

9 : p.-

,,,,,,,,,,,3 c

,f:w, rN_.,.,.

p

_..?

- g

[

1 e

p:y c..l

- 2 g,:.(,' y; 3,,,,, g- -

9 y*'.+, mu $.fi, gk, s.-[d.

m 1

v.

1 - 9.,j.%u y.g-

• C

.s

. Jfy' ', & -

t,,,..

,;. ~.$g %,.,,.[A-,,9

~...

a.. a.

y;;:-

sn Q L W #. j').

.,,,,,, Q e

..c. >,,

.., t - m. ).,

9

,.t. a,.,..,,

. c v

h * );A.

'M.

... ~.

...", y.,.4.

,,p _,

)

,/;~

.c_

?

, y 4

.k T,

.. &.y y

y g.g; y,

y, g;.

,. g s,,v

4 3

'w h E

6

'M Af y-

.v Lightyear gives you four main evaluation screens. Each evaluation displays the results with horizontal bar graphs and numerical scores. With Lightyear, you can see how you are making your decisions-which criteria and rules most strongly influence the ranking of your alternatives. You can see how your alternatives rank with the criteria you defined, or with the rules, or with both.

Uglity9ar is Fast and Fl8Xlble. Lightyear's flexibility allows you to change any part of the model. You can add an alternative, drop a criterion, change a rule, increase a value, give more weight to a criterion, or create a new verbal category. Once you make the change, press two or three keys on your computer, and Lightyear instantly displays the revised rankings. Lightyear lets you adapt your model to new informa-tion so you keep yourself well informed about which options remain viable.

" 4.;

a. -

y

. g N..,

R

,Y:,-

'e.

-l*

,. i Nl".N

g 9.p@~.' '

le M,4l.x sf

.p -

?

o s ; 'g Asy &y t,6 Mow -

W:

2",4,,$,

't

- k ).

~

I.

.c.

w

a. n-y

,7,

).- -[

D' f.Q!.,5 g,f.y; 4.

y "-

y s.we

.;gy 3

..,. g +7.,.. 1 w

a.

.y %,...y... q : gg

, - [.. ;-

T, y _. 'ejg

' ',:.) ** -

.,1. -.m.

3 c

l3 ' [ y 4 :' l.J. ij, ;

,. }-l.,

7,,-.qg-

~..

1. q.: ;.a.

. w. ' r

• y.,

p

~w.

c.

q.

. +:%.. * ' ' a. ', " -

- ' < f. f.~'.e

? _ _, ',,

s..-

. 'a.

.. r.

-a

- !s.

v.. > x,..u ;2:,

m.. c ;,. ~ -

u

"(

i "f

.j

%thescoresor4smaug... M f

IM

,Use this method to understand the dif-7 0?'

A S

' %.. 2 firesices between two possibilities. Om.a s.QfdWA%ceseehowcertirMMjof Deixmit evig to'Deasury Notes.gl,[j g C

N"$%%u coul(use itjust as easily to comJ M#

Mpere any#khertwo optiora pTE.

@$, NAN

'95$id 1.

., 7 -

. :.. - y x.

y s-Q m;y a.

.9 sy! >,

%y.

+

S.

Ibfri on the Power of Ughtysar. Choosing between two products or 200-no job is too small or too large for Lightyear. Use Lightyear to get results-and to communi-cate them etTectively to others. You can even build models with others. incorporat-ing their viewpoints to reach a quick consensus. And when you're done. you can print out any part of a Lightyear model for a presentation or you can use it with your word processor for a report. Lightyear provides you with the power you need to get your job done more effectively.

)

O O

=

W 4

=.

a L

N

~ '

y,gg gm MMWMtWWW4W ~'

L'&

w

'~ ~

w-----

q g

1, 8'

.,, ~.

, y fR

'if 1 4

h' cr,i;jjzy ;.

Q OG w

1

- + -

^

7 Jj v

y 1

S v

s.<

• I

,f b

,u

. ? ?,

D

w a-Q 4

_1

~

=

,'(

?

=

+%_

' 1 t

g

, ys

. c"?

x w:7 ;,

g m

e

.k h

(a~

a

[

g 4

e g

g

. c i

3-t a

e 2-e

- u a

h 3 f,.

, E3

h. l

.q

-}

l

c. __.

2

,{

'p'(.

k i

4.

x k[

1 a

Eik

.y, a

Er s

J V

~

=-

/)

,t;

.I 1

y

~ ;.:

)

F

-9+

w i

p l'

'l T

1 e

'li ff-A-Ki*~tly ykhC' P['.,.

ryg b

4,

~

'~

e L

r n

5 l_

h l

F 5

m 5

K' i

?

e-

Lightycor requires an IBN1 PC.XT, AT, Ibrtable.

CON 1PAQ,or 100% IBN1 corrpatible computer,192i; iW.1 minimum, at least one double-sided d,ive, and a color er monochrome monitor.

i.., gm I, ! :u,? ' ' 'i $ * '},} ? '* 1

.,.3 u,. u MJ c',,; -.i 1333 Lawrence Expressway Building 2 inu. re. ered i,.de

,n..d PC, XT, AT..nd Portable Santa Clara,CA 95051

.re ti.de..rks of the Imern.isonal Business M.cIhnes Corporatme courm...de ra or courw co.,.= corpor

u 3

EXPERT CHOICE:

q l

A DECISION SUPPORT l

I4-SYSTEM

"~

wi ooo f

es D

\\

by Bobbi Drexler Expert Choice is one of the new soft-The next step is to judge the merits of ware tools that supports the decision each of the alternatives and then to rank making process.

It provides a framework th e with respect to each of the criteria.

for structuring a couplex proble into To indicate the nature of the comparison manageable components. Problem elements the user chooses one of three types: (1) are clustered according to their relation-importance, (2) preference, or (3) likeli-ships to one another. A proble solution hood. mese pairwise comparisons can be is the result of judgments based on pref-Performed in either the verbal or ntaneri-erence and subject knowledge, cal modes.

Expert Choice alerts the user to inconsistent comparisons by calculating Expert Choice is designed to ruti on an an inconsistency ratio on each set of 28 IBM PC, IBM XT, or compatible cmputer.

judgments. A high " inconsistency ratio" could be attributed to a clerical error or It requires 128K m mory, and one double-sided disk drive. We software will sup-maybe a loss in concentration. The user port a hard disk and a color graphics or can check the ratio at any time. The monochrome adapter. A printer is option-final step is to re-examine Judcynents.

al, but re.w...eaded.

I used a COMPAQ to Judgments can be changed if desired.

If review the nMkage. Expert Choice costs satisfied, the result is the problem solu-tion.

S495 and was developed by:

I created a model to determine the best Decision Support Software, Inc.

1300 Vincent Place site for the Navy's fourth annual micro-McLean, VA 22101 ccrnputer conference. See Figure 1.

Ex-(703) 442-7900 Pert Choice processed the information I entered into the various nodes that repre-Expert Choice is based on an analytical sent goal, criteria, and alternatives.

heirarchy process. This process organizes The screen that provides onscreen compari-the factors of a problem into a visual sons is shown in Figure 2.

top-down tree structure.

Expert Choice scurra -m r.au cmrrmcr stre refers to this structure as a "model."

The first step in building a model is to ma' determine the Goal, the overall objective i

of the model. The structure branches I

I downward from the Goal.

Intermediate levels represent the Critieria, which are

"* l j"

l

'^ " 9 5" 9 j" ^9-the basis for making judgments. At the

.woec

-woe 6

-zuet

- roet

-scoet bottom of the tree are the Alternatives,

,C,

-C.

{:M3

@3

EC which are the options of choice.

Chips Ahoy

~

ace "cr7:uenE A en nice cowzance s:Tr terms. A three page technical reference section is provided for additional reading

.7,,,,,.

on the subject. The docmentation guides cost you through several example applications to allow the user to became familiar with Expert Choice. The onscreen dmon::tra-tions are presented at two levels: (1) i simic short, for those anxious to move on, and nacenare (2) detailed, which provides more explana-og r a n. 2.

tion.

After building the initial model, I real-The designers of Expert Choice consis-ized I needed to make same changes.

I tently sphasize that the results reflect entered the edit mode and changed some of the opinions and judgments of the user.

the criteria by combining, adding, and The docmentation states that "it does not delcting nodes.

I thought of some addi-tell you how to think, but it does make tional alternatives, so I used the repli-your own thoughts clearer, easier to or-czt2 comand to reproduce the alternative ganize, and more presentable to others."

brrnch nodes under each node in the crite-I found these claims to be true.

Expert rie level. My inconsistency ratio was low, Choice did not make me came up with new so I used the synthesize comand to obtain thoughts or ideas.

It did provide me with the overall decision. Using the synthe-a format to organize the elements of a prob-cize comand, you can print the results Im, which is the strong point of the rnd opt to sort alternatives by priority.

package. Although it forced me to organ-See Figure 3. for the resuIting printout ize my thoughts, I arrived at my decision that indicates, given the criteria, the before I ever applied Expert Choice.

It "best site for the Navy micro conference" is most effective applied to a ccatplex is the Pavilion Conference Center, Vir-proble that involves a large neber of 29 factors that, if mitted, could impact the ginia Beach, Virginia.

final decision. Expert Choice is useful stuenz uAn meno courtancz stre to present a decision to upper levels of otAr as sonrm er mesm wta^u. rumustsrocr recx. e.s3 management.

It provides a substantiating case in an organized rational form.

PAYlLloN S.M3 -

smee e.2st -

Expert Choice could enhance group deci-sion making.

It would require a small

,,m,,,tu xxxxxxx - nnnnn,

group that is well informed and highly l

g motivated. The group has to be willing to run. 3.

i participate without letting differences of f

Expert Choice provides comands to edit, opinion get in the way.

If that given insert, delete, rename, mark, move to, and situation is possible, Expert Choice would replicate nodes. To accenplish this, pr mote brainstorming and the sharing of ideas.

Expert Choice makes use of the < Alt > key and special function keys. 'the onscreen help menu provides a complete description Expert Choice is not for the average f

of each. To move around an Expert Choice business application user who needs word l

tree, use the arrow keys on the neeric key processing or a data base managment pack-age. A decision support syste is intend-l pad.

ed for mid to upper level management.

In 1

l Tne e M is well organized.

It begins the past, there have been very few tools with "How to Use This Manual" and goes on available for the managment segment of to detailed chapters on how to apply Ex-the user comunity.

If applied, Expert pert Choice. The documentation is clearly Choice bec mes a tool to effectively plan written and concise.

It includes an in-for costly, long-term projects that re-dex, table of contents and a glossary of quire intelligent decisions.

copied wiin permission October /1984 Naval Data Automation command l

r

'l i

l Expert Choice Decision Support Software, Inc.

l 1300 Vincent Place McLean, VA 22101 l

903)442-7900 Compatibihty: lilM PC, XT, compatibles l

Released: 5/84 Price: $495 Application: decision support l

Expert Choice is an exciting new concept in the business A

software market. It is an easy to use decision support soft-k ware package designed with the executive,ind the consul-tant in mind. Expert Choice helps you reach complex deci-I sions by synthesizing data and facts.

l To make a decision with Expert Choice, you incorporate your personal goals, criterii, categories and choices within f

a generie format resembling an imerted tree. Setting up this i

decision structure is a simple process that requires very little typing. First, you enter your goal, which then becomes the top of the tree. The tree then branches out under the goal l

as you enter more lesels of criteria. Additions and deletions t

can be made at any time.

Once the criteria are established, Expert Choice leads t

you through comparisons of the elements. Decisions can be i

made in either of two modes: verbal or numeric. Each mode has three types of comparisons: "importance,"

"prefereree" and "hkelihood." When making a verbal comparimn, you rate each element on a scale ranging from

" extreme" to " equal." a verbal comparison based on im-portance might read " interest rate is equal to moderately

v..

more important than down payment." A numenc compari-i son is made by assigning a value of from 1.0 to 9.0 to each item, v ith 1.0 being equal.

liased on these weights, Expert Choice offers conclu-sions. The output, which consists of bar graphs and the global weights assigned to each factor, can be displayed or sent to the printer. Calculations can be made at any point in the comparison process. This allows you to change the im-l portance of criteria as more information is added.

Espert Choice is a unique tool. The program's flexibihty makes its applications virtually unhmited. In the few hours I spent with the product, I found a few shortcomings in its editing capabihties and error messages; however, these were Jetinitely mershadowed by the informative manual and the insight I gained from structunng my decision on Expert Choice. Expert Choice is a pioneer in a new area of soft-ware and detinitely worth your perusal.

-EA l

i BUslNE ss Soti4 Ant 69 NoVE MBER 198 4

[

k

among several alternatnes$

Zwick knks the results from Expert Choice to 6tes from Lotus Develop-

=

ment's Lotus 1-2-3 spreadsheet for a D

composite on tendmg conditions. The mformation can be used by ditierent decision makers involved in a loan project, he says, allowmg for the great-est possible utility of the data.

"Our clients can integrate the Ex-LOW MSk IO8nS pert Choice data with their own in-formation and then have a basis for Jack Zuick uses a decision deciding on whether or not they support program to evaluate should make a loan."

risks associated with loans.

Prior to implementing Expert Choice, Zwick says his 6rm provided Jack Zwick, president of a McLean, 6nancial information concerning in-VA firm that specializes in executive

ernational loans in report form. Al-training and consulting, aims to take though all the necessary information away some of the risk in making inter-was contained in these reports, he says national loans by providing his clients it was difficult for his clients to use the with timely information and decision data with other collected information.

making expertise.

With Expert Choice, the data can be Jack Zwick Associates has been integrated with other programs for in business seven years providing com-more effective decision making. "It puter training, consulting and custom-puts the decision makers right in the ized application software for large middle of the process," Zwick says.

banking interests such as American Zwick originally discovered Expert Express and CitiBank and for smaller Choice while reviewing business soft-regional banks.

ware for his clients. The review pro-Recently, however, Zwick's con-cess is a major part of creating expert sulting operations received a boost systems that include training, custom-when he purchased a $495 program ized application software, tutorials called Expert Choice from Decision and manuals.

Support Software. He is using the pro-Consultants from the 6rm discuss gram with several maior banking cli-specific financial applications with cli-ents to help make strategic decisions ents, Zwick explains and then take in placing foreign loans.

popular software packages such as "With the program, we are able to Ashton-Tate's dBASE 11 and Irame-help them (clients) determine how Work, or Lotus 1-2-3, and design risky a situation is," Zwick says. Ex-complete executive user program pert Choice is a decision making pro-around them.

gram that enables a problem to be vi-As for the hardware used at the sually stractured in the form of a 6rm, Zwick says, "We use whatever hierarchy. To determine the risks asso-our client does," which runs the ciated with a given loan, Zwick says gamut from the IBM PC to Apple's the 6rst step is to establish a set of Macintosh to the Compaq portable.

rules concerning a lending decision.

The computer most widely used at the Then speci6c data such as the finan-6rm, however, is Otrona's Attache cial status of the lender, the economic

-a portable computer Zwick says condition of the country and the ti-

"has a lot of zap."

nancial status of the borrower are then Having used and reviewed a fair entered into the program.

number of computer systems, Zwick it organizes this information, rank-concludes," computers are just a deci-ing it accordmg to priority. It can also sion making aid."

conduct cost / bene 6t analysis in cases where costs and benents should be Sharon Patin is a freelance writer Copyright 1984 M&T Publishmg considered separately, Zwick explains, based in Palo Alto, copied with permission BUslNEss $oFTWARE 91 NOVEMBER 1984

7 79-

• <.i a,

.\\.

..- ~4 7 3.; ;;,.x. p'

,a, z. v, t.,f,;, f, T. y,

s., -', '..' : '.

t.,

j.,-

%C

. g l

.'.. l" .,. /.. Q.,...y.4,

. ;. y'. 4 e.'

...r

.,.;. i N,

.,,...g

>. ' :~.ya ? '.., _1;. - y.,. -.:

1

• h.-.*

N.:. ). '*

. g*.e -[ J,.. -. -

s

,. r '

p---

. u

- _*f

/

y.

i

,,<..;. > : ;= -

1,

sc pr.-

4

" s -

,.,i r c, -.*'[.

..g

., _ - +-,- -

.a e

.t' m '.,;.4L4,,

q j.,

3 : Y.,' -" g-Q}' ; r :

q L

3."

g*

.( - ; ; y..'y

.. ?

-J N,.

+ -

,?

v 4. , '...

+%

.j, - +,

.?

l-

~

4.

. ~.

6 3

s

..g e

. l y,.

. ~ -

m 3,

,y*f f '.: '

A.'. f~~',

_ '; y 'n,_,

.f f '. & - l3

.f. '.,,

y 3,-

".- +

f t _;;, - : [

. 3

- " '..- l

'.4 ;.:

i Y

0 0., - y.** ' ', %.:

4

~

l.' i f, s[. c(.

I4 ;,'.*

y I

,..)

' ~

~

. k-Q...,

.p*-

. }...'

,/ 'y,

E ':

'j ;

',,',.i...

, r p,

,e

[r h

,t'

[,,

.4 7.

'r g..','

j

,9' -

gr

~

,f.

.'>,.g..'.,'

.i.

I 1.,

-,i y

g..>j A.g., 3.

- g - (.., $.-

r,,,,

,f

4..

c.

g.

. 3

+p o.

q

s. -.

=

s 1,

. ". +,., f.

• y*,*

w

- g, ' c;, a..-

.s,3..

Di,,.,l 8,% [ c l' ' ',.,, e I..}

'., 3

l,'

}

-" L -v

^

' p q..

,,3 -

,;,L,

. ' ' :n.

?,

O

%. ~.

~~l ? ' ',,.",1,;

.[

I*

~. (

fy

'f '

F

-1

}

l-

}. / ] r h

k-

,"Y *. '

"... l;d g,

Y%

a'

.-.... 6 +.;, J.

,g,'j.,

[.-

',4

?.

q

?,.,

- m w

u4....,

,.J.

Q3 y

...t,. ;,.

~.

,. \\,.

(. p e.'-

-[ [.

..I

.., 4 -

. [h 2

• .s

. n w

a ~. ~

E3

..] d] )

.O

-%., g -

f.gg...,

v

- [y ; ' *-

'.I,-r y,-

k,. t

..t,...

,..-[".

.[ _

6.

4

.- w

., i, g..,4 fd. D D ~, %

x 4=

d

, (.

.,, 45 b,w,_

g- -s--s 4

",' ^ y ' <

- "..,>4-,

v'

^ i [.I

.}'

u.-

4.

.~

n,,.-

,,, +.

"+

i 3,

,c

-,) ;.,

. g 4 <.

., s 1

+

.y g z.:.

s

.. + - '

6,,.

3

.Q , '.,. g -

j ',b. - -- y

..;,.a.,,,

4

. n a.

.= :

g,_

. g '.. k;,. e ;.

t -

-;. -, '. ' i..

, + k.

.. "'* j e,4*f.+4.g f;

2, ;j.,

.g.

s.,, b,,-

i

. I

,g,,

, s

..w*

.f

-r.

(,

,N ';' _N M

,,';.1

( g.;

.l,.

b. '.= ;,,

~

39 '*" [,

-k 7

'e

[(.{

-]

l-Y

'w l=e m

& pd' ~4

[.

'l.. L '.

We

=

+

-f,% ' ' -

.)

.--.t,,-

.x p y r,,

p=.=w m. ~. + w - ~ m.a m r + y..w p w L

rw a

'p'-.t

..q gy..

=

e q g w.m -- - w ye en sum an

/. ;

wmemar w me.

gun. ? seen gn:..p p p,

,,,Y

' Y._ ',.,.;. :

I 1, F

'I.I

'y g +; $

3"..,'..'

N Y9 9 9 9 9 " *O ?.N..T*g1 i.N. MP

.M

,.~.. %,,i,- ' ',

g- -

4'l'.'

..,.r_v

. l..,

\\ y ' '

W V..> ;.

' )

e-c

- 3 e,..

i. :. '

~ ~ ~. -

.e

,.p-f,;P,',.fr, y. t i

L s.

s g

(

..d ';O

- e 1 --.

. 3.l-k.

., f p3 Y,

"..,f.d?

e

. 5,.

7..

_., u

,Y h.hth' h.

p,

(,

.,~ % f,

,,D;I',' !r,/. r..,.,, 4 g_...

< h e,.%,,,:

-n.

,'A...>..,.4,.., ;.~

\\'

c,,

.,g

" Y,,

~*q

--1

, ty,4,,e, P*..y *ser - -.. ~ ;y0 M.- _

' dak e' o. ' " '.

U

,.3.,

.y ; :

.'%'". ' [. -(-t ;- - g ll l p, h [ H,i.

(.m

.} 3,

.-, 2, /,['

d j

. _7 " *.. 4,. j.f. $, f.

Q,.

,j.^v c-

.[

,i[., '

q. J i~.

(k 4- [~

N.h'

'Y }*. k[-' -- f ^ k, kh ? !,

f, }-

.. - ;.= -.

,6,,"+

15 g ). -

,.,]N (' n.*

i h,,

• J i.. /

,5

• Y.6, % h..,W;.

[., f

,,; 7 A :. ;. ~

- y -

f

.. j, 4.,

.y,[.,

2 N

5 c'%':Q

i y Q' '. v '$...-_f V'.' +'.-~

yl, _. = y

-.s,.,.

' wh i p,.

-= g ;

-f,

' Q'i

_ _., -, Q} ' ' i...j- ; $

d'

. g

,g'}.,'.,-#-

, ' %... ?

_ ~f s

% i,

'-,o.

v=

$e v

$ g.. x },Q' - ( ':. ' '..

4., ;[ lf.,.,,,3,J. i [T ; e,A %,,., "A m 3.g

'.m.

.(l,

x. (.,,t n,

y; ',o 7*,

g[$ :,%

% of

, x,*'

..,g,.

,g '..

m.

4 4. >,

'}

,1,-

.'..,,, ~ "

i Q '.'_.

[.

(

h I

^*'

., ;._,. :, ; 41.

,, 55 4

' $$ [

i _.. s y

.7

[',s-p.

.j d, t'

% -l ?--

. -" 4.: -

_4, tv

.Se

. '.. *;- gA 1,

.:;. ; -..q..

+

r

.5,

>g s

-t

_,c,

/

-g r: :

j x

(. y' q.. - 4.,,.,

[.,

... - yn.a

.e,#-,,..

. s_-

N ' t Q'[:.

,@~

p 3 m_.~s.w x,,.:...,

, (

y.

[.[

w

. m..

. +..,

s;' g g - :.. -.

7....

__ g _

g,.

. m y

4. s. <

-l n'

,- e

.-v

.9 u r y'

N,-

g'

.s.. a,. - ;'-,, -

. e

g... -

.s.

• e

' wf, -

y-

'L

~-

t.

s

.g,5s

'.Y g.

1. s ;..

,.?

}:. ', ' [ R,....,.,. s,,... [ _

l l_

+

v.

I.

^

..s

.y

..g.....

f.

p..

v., ? _.

r,.

.... ~ _, :, 7, _

.'L

^.:

w.

.- ' _3y_.

. w

- 3. _'. l ;

c b

g.,

e

_s,e

+.

T m

2.c ;...

.T..

e

-. t e

. - c -

,. c.7y-s.

,. ;-. 4 -..

w..

y.

s

- um -

- j,

?

v a i

.1 9

-g

..1 m.

7.,,..,.,.,,.

- ~

e e

1. -.

... ~ '

.b A

..4 8

3 4

, g i

a 4 -.,

. E 7

y i..

(

4; L'

s 4.*

~

-.. g, s.

m V-..

y r,

.A n

~r

~

3..

f

.~

, ',.;,- 6

., g-

1. f i '

i *

{.,'

4 i

,, i.

. '- ^ 4.

. -. w -

1 t. '.

.; y<

r"

. e.

p '

~.

.a f., " (..

s,..

~.'

,[ -

1.' t..'a

r. 1

..v x'y s.3 t

- 9

- 1 s,

-t

. '..+ -

~

~

.m

., i

.y

.g

- g t- - (

_I..

D

-'s

,., 93 4

L_-

~,

9 f.

a.

-s r

t.:

.y

', f,

~ * '.

y

_.4'

~

+

e -.

s - -

.e 4

a.

v

-t j

+ y

(

f' c

• ..' *. g

,,, 4,

d i :

~ *..

~3 a -

g - %. +.

'.,,, ',., ~ -

g m

.u w

'h

=

ps

. +.

_,g 4

O I*

b,

' 7.f

[.[

d

[*

j 4

s

.,.. r:..

. e s

~..

+

..; n

<o

-z g

yJ. -'"

g

', ~.... P

~

s-y -

s}' -

3 '. ',

x 1

.,. ~

a._"

W.

. f..

,e

,-?

w s -

t a.

~

. x g.

s -

e

^^

j=.. :.,

n

~f.

y-

'. e'

' l t

.1-.

t-

-[

.g'

,c, '~

' p

~

?

".., -,. _n * ' '

y.

[',

. g 3

~j$ 3' J J.. '"

0*

' h(.

T a

,f v

" 4 I

-2 8

e.

h

' +.

- '. od ' '.' 4

.h., -....,*..

., e s,,

^ - "

g

't-. '.

7

.c

'f

~ -,

W

.g e:

4, i,. _ '. -

-- 4.

s i.

_. '. i'. in,.

p.

• - -:.=._

.. f-

, j

+, '

s

.9

)

+

r t :.

^

- 4 Q 4.,

a

• -' s.

p.,

,.,,... 4 *

-., -. f,y, 2

• • ~ ',,

. _,..,p;,,,*. _.T y

, f a

c,.

r; h.

=j ",

p,-

f.

b, e_

l*

-. +

..r x.

. h.. -;.

3....<

i -

-, ~

~.

} '_

- -. s.

3..'

1

.t a s

+-

,}

4'.

t*

Q.r g

,..g it e

. I,

-..~ -

s

,)%,, +. .. *

,.[

^ >..

.,,(,'.

, '~

'J.,. =-..- -

.s

• ~

~

y.

m..;.. -

L. 9 l

g.i - -

..:..,"w.

• [

'..' pV .

[ - ,..

l~

f O~,,

' ~~

t

..1 j

n.

r.

a.

c

-d

- M

,p-r.

e.

f _.-.

.v

• t f.=-

': ri 5

g.,

,.s

.u a_-

3-

['

3,,. - c.-,

f,

.c Y,

~..

a..-

4; -.

e-4 4

.4

~j

. S.*, p -- - ;

.w d.-

- " -. ~ - i..

t.

,zu,,

t

,.f a g

-54

.. ~.

...m g,.

4_

4

' /

g.

1. n.,... -. -.

' +

• w.-

,s

=

^.-

d % b'*

).

e.

y y"

- y

-' L 1,. *,,

rf s

.e s'

.i

.t g,..'

-. p' * -*

.g

,j,,

T.

7

.-f_

., s g...

e.d,_

.4,

. _7

v. i,. ' Q '

"g.

t

\\'.

_. ~.. -.'

-... y ;

. 1,. -....,

- I

, g.

f.,

w 1-e.~

4

~.*

..,.. +

e'%+ -.a.. x,.

w.=

~.,.,#,

g

..rr

.r.

. - +

=4

-<gg..{. ','..' j.? '

- -.- b).' -,

s

.-y-h

".L. \\ ! "

.1 4..,t,q.

E,

• L

-,'4.

.,-p.,'

t.'.

s, v l 5,.,

- (

's

}.

o -. ' '.f *$

s

~ '

M-

% 'h.

1

. ~'

c 9,., *

', - :.]

- :. -. - ;~. ' -

e.$

/.,<

r..

,. te

- y-4 u

p.-

,y

-w+

p

-,g..

~~ ;

y

,k..

g-

, yy

. _,., M. ' ' ',

.3..... >. <.. K f,. :

~ ' -.. r :'.,' ~..

~

f *

• -' i

~

1

- r

-. 6 r.,

e,

' ~.,... E..., [.... s,. [..

.. Y, '....

.,,. -.; r.

,.l.g ', - l

,,-r, p>

g) -

a g

4,

?

a

., y/....-

-4

+-

2

% r. -' ' y y

p s

x

. h,(

. I '4

.T

,. A.,..i

',.b

- i

, s

~

g. g i. ;.' ::,... Q N_ p* g* 'y

.n~

y b

-; A

- : ~. ; :.

4, y

- ; ip

,..,,t<-

.. w.

g ~.,.,

t-

- f. ' U,.]f..,. f

. f.',,( : q 4,' :.,

si:.;,

+

n

. y4 -

J.s"

' -l j i (

., [. J. ~, '. K, :., [,4.-ly g;

p y v

.s V

2

,?

. +.;.

x..

_ t

.u,

. 4s _

s.

., y.. <,.

w

?.

. s.

,..^-

x...

..,,g';'.

-,- - u -.. -

y,-.A-

- -,,..p 3

5

-... y.4 p

g_.

.s

) -.v f

- +.*

,. a -.- -e.,, f. -.

9'

.,f s t*

n o

),s 1

s...,-

eq r-

. A. ;, =

5,. s -... p p.

- -. ' _ ',. ~.

tq.

_,g ~.,

4-.

g_

_... '_ ? a. <

. _ '.,.y' 4,

g g

_,; c....

,.fk1+...-. -f.'.., 3

- s-

. '.x. -..

g,

,,..,, 3

s.... * *..:....

y 'f. p.,.f,, w; bV c.

.s.

...,m

. s-

'.y

?

,.. j.:..'._w

+.x.. < 'R..

-T

?;

w.

-. % l l..

. :.,. =

n....,,.

.y s.

.4s r

..,, u ;. -; Myny.. p.* > W v'. (. ' :, ' %j.Q.. V

/ L

%. a '. -

m

.E <. S cl ' P ; ' '

.. g -

g,,,; ;,. i :, t-l ([

.},

9 %}

l

. g m, g gep g m yg,gg

Arborist puts the power of a sophisticated Operations Research technique at your fingertips.

What is the tulue of a good Screen A adapt the structure of your tree to decision? How can you be sure changing business conditions that you have considered all the without having to go back to the attematives and the value or risk beginning.

associated with each component At each step along the way, in the decision chain?

you assign values to the units of When you analy:e complex the problem (such as monetary situations, and run short on time, gain) and a probability for each there is a tendency to limit the chance event.

scope of each consideration, hop-With these values and proba-ing that the cumulative result will bilities, Arborist determines the be acceptable. A better, more expected value of each point in the comprehensive approach to deci-tree, the best attemative at each sion making exists-an approach decision point in the tree, and the M -

expected value and distribution of that alk>ws you to consider the entire treeof attematives and E

E potential outcomes of the entire risks.

decision.

There has to be a better way.

Developed as an Operations There is hardly a business l

Research discipline, this "scien.

Arborist is that way. Now the activity that couldn't benefit from ti6c" approach to decision making decision-making power, the com-better, faster, more effective deci-is an important part of many prehension of uncertainty and the sions. Arborist can be valuable in MBA programs.

ability to consider the uMe pmh-evaluating decisions ranging from Unfortunately, once out lem before making a decision,are personal 6 nances ta long range of school, MBA's rarely use this available with Arborist and your corporate planning.

valuable toolin their day-to-day personal computer.

Attomeys will 6nd that the evaluations. Constructing deci.

You are free to concentrate complicated question oflitigate i

sion trees with paper and pencil on the content of the problem vs. settle can be much easier to is time-consuming and frustrating, while Arborist captures the infor-resohe when the choices have especially when the dynamics of mation and represents it as a been evaluated on a probability /

the problem mean that the alter, logical series ofchoices. You have cost basis.

natives are changing quickly. The the freedom to change, re6ne and Engineers can use Arborist to 6nal, and perhaps most valuable, evaluate risk, makedesign approach evaluation-calculating a value determinations, and include the for each branch of the tree-can many factors involved in a com-consume many hours when plex problem.

done manually.

w3 yl.

-,e--,y

--v y

m 7%

-.--.----.---m

(

d s

Saecn 8 Arborist makes decision tree Saeen ( :

{

[

analysis easy to use.

L 4;i 2 h..N. di m *. N.i ' D

, ;

j7.3[ u/o4

,. g f...,

is its case ot use. This screen

,.... s y-e.,..,'

The real power of Arbonst

-g i

.. v.

. ~. '

• Jl 5-A shows you how easy it is to con-

$:.f..-(. [, ~

.?

+

J

.. ; r. -,

?

-! N I[ !

-r.

J su ? 14. 0. - -P

• V.'

e.

,.e

. _~

t struct voor decision tree. hee

'./

e s.

. ~p gw d W '. E< '. g,4 p...:vgp.c. k M..-+i g..~l

• w ;

4 i..

Saeen A). Each juncture requir-c

..,,y

..y.s 4.

.. j 4

s ng a decision is indicated by a o Dg wp M ~ f. square, chance events are indi-y '-qN3

,. ;. ~. e,

.m C.,.y ((. s g

,g.

i ' ;

,.[..

Cated by a Circ!e.

~

}*M f ya. i ' I lk7

,- i

-f.~.a, (

.. " 1.-

s ':.. :... _. i f. c.., '-

wt..

.1. -

As 3ou construct the deci-

.. _.,, n o.

sion tree, you always have the

.-e.-

. x..

v__ 3[,,.j.3 b i. { :? f f s.i..f ; {. ;,

(.

, ability to view the larger context ;y,,,.y 4.y. p~. ;, -

7-i, i>f the pr(>blem in the upper

~'

' - ~ ~ -

m@

5

' naht-hand window of the display. l

-- _ m

- - - - - - = = =

%,hile you concentrate on the problem, Arbonst does the At any point you can ask Arbonst's value as a unan.

putations tor you and keeps

.zrbonst to evaluate the tree (cal-co cial planning tool encompasses track of the results. W, hen ques.

culate the expected value) and business plans, hnancial analysis-nons of probabihty occur. Arbonst Jisplav a distnbunon of the out-pncme analysis and long-rance presents two choices hee Naecn H1 wme values hee 's.aeen (,) And planmne.

It numencal input is desired, the t;le expected value of the best And, as a marketing tool, the det isiiin f ree apprtiach to pn > duct probability i>t the event can be soiunon is displased at all nmes.

entered <iirectlv. it numencal prob-intn ductions, corporate redirec, abihrv is dithcult to Jeternune, the non, resource and personnel P'"babihty w hee l can be adiosted Jerlovment and strategic plannmu to arnve at the hkehhoid or that could lv cnicial to the continuine l'UIC L IWe -

klea!Ik1 i)t a ci> Hip. Ins.

When the situanon changes ArkXirlst Can be lised by group 3 AIh'USI makes if extremcIv easy of decision makers to review prob-I " "dity vi>tir structure. Y< >u can lems and agree on decisitins.

copV subtrees and even store them tor later reuse. Tlus is a tremendolls time saver when patterns or branches repeat themselves or ahen similar sittlations arise.

s of A.

m

A powerful new problem-solving tool.

In our fast-paced world The combination of LISP individuals and organi:ations are (a symbolic processing language),

pushed to make many decisions in multiple window display, graphical increasingly shorter time frames.

representation of data, and plain Computers have the potential to English commands makes the make decision makers more Arborist decision-analysis program effective.

responsive to your thought process, But, up to now, there hasn't so you are free to concentrate on been software available that can the problem (or opportunity) you adJ discipline to the decision-are analy:ing.

making process without adding Previously, decision-analysis new problems and a lengthy learn-software generally required large, ing curve.

high-speed, mainframe computers Texas Instruments is a and Operations Research profes-pioneer in the emerging field of sionals. As a result, decision trees Artificial Intelligence-tech-were used primarily on large nologies that will make computers problems by large organi:ation.s.

more powerful and easier to use.

Arborist makes this vital planning Tl has employed some of these tool practical on personal com-techniques to reduce the barriers puters. For the first time, this that have prevented the wide-technique is available to everyone spread use of the decision-analysis who needs to make better and technique, commonly called deci-faster decisions.

sion trees.

I 1

J Arborist can help you make better decisions.

Arborist could become the Decision Tree Definition Features:

Arborist Recommended System most important decision-support Standard expected value com-Configuration tool that you add to your personal putation, with maximi:e/

TI Professionaland computer. For the name of your minimi:e option Ibrtable ProfessionalComputers:

nearest authori:ed sales location 10 Mbyte hard disk Pmbability distribution gen-call 1800 527 3500.

eration, cumulative and non-Diskette Drive cumul tive, at n poin n the Arborist Software Features Minimum 512K bytes of memory tree, wit resu sp a3 UserInterAce:

gr phically

3. Plane Graphics Option

• Graphical display of decision tree

" 9" 5t*C'" "

User-de6ned modeling formulas v

• Interactive " browsing" of the tree and variables Color monitor stmcmre

, Sensitivity analysis on modeling hts-DOS 1.25 or 2.1

• Interaction with the graphical variables de6ned in the tree, display for de6nition and edit, with results displayed ing of tree structures graphically Other System Configurations

• Graphical probability wheel aid

• Expected value computation can IBA18 ItrsonalComputer :

for probability assessment be performed on partially 10 Mbyte hard disk de6ned tree structures

• Cursorcontrolled, Point / Select Diskette Drive menu interface for commands Outpia Features:

. Minimum 512K bytes of memory

• Single line explanation displayed

. Printout of arbitrarily large tree for all commands, plus Graphics Controller (required) expanded HEl.P facilities avail-

. Output of probability dis-Color iMonimr l

able with single keystroke tribution and sensitivity data in disk 61es or printed output NIS-DOS 1.25 or 2.1

• Accommodates color or mono-chrome di> plays, an 8087 Tree and subtree strucrures saved Numeric Processor, hts"-DOS in disk 61es 1.25 and 2.1 1

• Thunrhaul d'r Jit es.mehe 4=uara' ha Nen.'emeJ.'n N l

res..mmend,J Tioceques.c t.

.e shce s.crisurar.c.m mun', m ord ed Ji rla dualds Oh me a venJem4ra,4 Tesa. In.eiumeen in.,.rmed

\\1.

te k4n rear. cal Bu.mc Ma hine, u.

TEXAS INSIKUMENTS Creating useful products and services for you.

nm - -

s