Document [original]

[111]

omission

in

the

simulation

literature.

This

refers

to

the

neglect

of

potentialities

of

simulation

as

a

scientific

research

technique

enabling

a

sophisticated

and

rigorous

approach

to

analyzing

&dquo;simulation

models.&dquo;

I

wished

to

point

out

that

relatively

too

much

effort

has

at

times

been

invested

in

the

modeling

stage,

while

neglecting

simulation

as

a

model

analysis

technique.

Thus,

Schultz

and

Sullivan

may

disagree

with

my

definition

and

propose

an

alternative

which

may

be

more

useful

for

their

purpose.

(Although

it

seems

that

their

five-point

definition

is

not

strictly

in

line

with

Aristotle’s

advice

to

definition

makers.)

It

seems

that

part

of

their

misinterpreting

my

argument

stems

from

our

different

disciplinary

backgrounds.

I

used

the

operations

researcher’s

approach

(see,

for

example,

Hillier

and

Lieberman’s

definition

in

their

&dquo;Introduction

to

Operations

Research&dquo;

[Holden

Day,

1967] :

&dquo;Simulation

typically

is

nothing

more

or

less

than

the

technique

of

performing

sampling

experiments

on

the

model

of

the

system&dquo;

[p.

4~0] )

which,

I

believe,

should

be

that

of

the

social

scientist

as

well.

Hence,

my

implicit

assumption

concerning

experimentation

with

quantitative

models,

for

whose

solution

computer

simulation

is

an

appropriate

technique.

(Pencil

and

paper

or

a

desk

calculator

are

other

possibilities,

but

one

would

not

recommend

them

too

often

in

this

day

and

age.)

Surely,

if

one

deals

with

a

conceptual

model,

expressed

in

verbal

terms

or

general

symbols,

some

of

my

remarks

are

presently

irrelevant.

In

the

latter

case,

the

word

&dquo;simulation&dquo;

assumes

a

different

connotation,

closer

to

Schultz

and

Sullivan’s

definition,

in

the

same

sense

that

a

word

such

as

&dquo;significance&dquo;

has

a

different

meaning

in

statistical

theory

as

compared

to

everyday

usage.

-Mordechai

Shechter

Technion

Haifa,

Israel

Comments

on

"Simulation

and

the

City"

0

In

the

December

1970

issue

of

this

journal

(pp.

411-428),

Berger,

Boulay,

and

Zisk

have

outlined

some

general

difficulties

which

in

their

opinion

have

slowed

the

development

of

simulation

as a

method

for the

scientific

study

of

urban

society

(p.

412).

Although

I

accept,

confimmed

by

my

personal

experiences

gained

in

the

United

States,

the

authors’

[112]

conclusion

of

basic

intellectual

difficulties

being

responsible

for

this

situation

(p.

413),

I

disagree

with

several

of

their

interpretations.

In

my

view,

some

of

their

statements

lead

to

an

increase

in

the

general

confusion

and

disagreement

in

this

field

instead

of

facilitating

a

reduction

of

it,

and

I

would

like

to

reply

to

these

authors

with

the

following

arguments:

(1)

Presupposition

for

any

simulation

of

an

urban

phenomenon

or

process

is

its

definition

as

a

system,

i.e.

as

a

theoretical

structure

consisting

of

elements

(subsystems),

relationships,

and

orders

of

action

(rules).

(2)

Depending

on

the

kind

of

descriptiveness

and

determinativeness,

different

approaches

for

the

simulation

of

one-and-the-same

system

have

to

be

provided-e.g.,

an

approach

to

investigate

the

unknown

relationships

or

orders

of

action

between

elements

of

an

urban

system

(mainly

those

between

the

social,

cultural,

or

political

elements

and

their

ecotechnical

environment)

will

differ

significantly

from

an

approach

that

tries

to

derive

optimal

solutions

from

a

system

completely

determined.

(3)

An

excellent

and,

in

my

view,

still

valid

description

and

determi-

nation

of

such

different

approaches

is

presented

in

an

early

RAND

publication

by

Geisler,

Haythorn,

and

Steger

(1962:

5-8;

see

also

Steger,

1965).

These

authors

define

a

continuum

of

systems

analysis

techniques

ranging

from

least

to

most

abstract:

Increasing

Degree

of

...

Abstraction

and

Formalization

(4)

The

three

simulation

techniques

defined

by

the

terms

One-to-One

Simulation,

Game

Simulation,

and

All-Computer

Simulation

(p.

413.

Ray

and

Duke

use

the

terms

Gaming,

Gaming

Simulation,

and

Simulation;

the

of

this

journal

has

employed

the

terms

Man

Simulation,

Man-Machine

Simulation,

and

All-Machine

Simulation

[Inbar,

1970:

4] )

differ

significantly

in

relation

to

the

type

of

initial

empirical

bases,

of

research

approach

and

procedures,

and

of

results

obtainable.

The

often

cited

statement

of

Abt

that

all

games

are

simulations,

but

not

all

simulations

are

games

(p.

416)

does

not

provide

any

further

understanding

since

it

neglects

completely

the

contribution

of

the

formal

&dquo;Theory

of

Gwnes&dquo;

to

the

techniques

oaf

all-computer

simulation

and

mathematical

analysis.

[113]

(5)

Thus,

turning

to

the

argumentation

of

Berger

et

al.

(pp.

414-415),

I

would

conclude:

There

is

an

essential

difference

between

the

techniques

described

by

Ray

and

Duke,

and

the

failure

to

recognize

their

identity

and

efficiency

has

greatly

contributed

to

the

failure

of

simulation

to

develop

generally

adoptable

models

of

urban

social

forceS,2

and

the

choice

between

these

different

simulation-types

should

only

be

made

on

the

basis

of

the

relation

between

the

objectives

of

an

analysis

or

design

and

the

kind

of

initial

empirical

material

available.’

(6)

These

conclusions

lead

to

an

assumption

of

a

principal

interde-

pendence

between

the

objectives,

empirical

bases,

and

techniques

of =an

urban

simulation

approach,

and

thus

to

the

request

for

the

investigation

of

efficient,

if

not

optimal

combinations

of

these

three

components.

Such

kind

of

theoretical

concept

of

a

&dquo;general

appropriateness&dquo;

of

urban

simulations

could

provide

an

important

contribution

for

the

still

outstand-

ing

comparison

and

evaluation

of

past

and

present

activities,

which

in

turn

prove

to

be

in

my

view

the

inevitable

presuppositions

for

any

valid

contribution

to

this

field

in

the

future.

-Henning

Schran

Technische

Universität

Berlin,

Germany

NOTES

(1)

Depending

on

the

occasional

ability

to

abstract

and

formalize

the

process

investigated,

war

games

(p.

416)

have

always

been

based

on

these

three

different

simulation

techniques:

the

gaming

type

("free"

game),

the

all-computer

type

("rigid"

game),

and

the

hybrid

type

of

gaming

simulation.

(2)

An

instructive

example

of

this

failure

in

the

recognition

of

identity

and

resulting

confusion

is

presented

in

the

same

issue

of

this

journal:

the

"Genealogy

of

Urban

Simulation

Models"

(p.

481)

shows

an

assignment

of

several

projects

to

the

sector

"Games,"

which

proves

to

be

not

valid

in

relation

to

the

a.m.

typology.

(3)

This

conclusion

in

no

way

neglects

the

influence

that

costs

may

have

on

this

decision.

From

a

general

theoretical

view,

however,

this

influence

has

to

be

seen

as

an

exogenous

variable

to

the

simulation

approach,

not

as

an

endogenous

one.

[114]

REFERENCES

GEISLER,

M.,

W.

HAYTHORN,

and

W.

STEGER

(1962)

"Simulation

and

the

logistics

systems

laboratory."

RAND

Memorandum

3281,

Santa

Monica,

Calif.

INBAR,

M.

(1970)

"Editorial

introduction."

Simulation

and

Games

1

(March).

STEGER,

W.

(1965)

"Review

of

analytic

techniques

for

the

CRP."

J.

of

Amer.

Institute

of

Planners

2

(May).

A

Reply

to

Dr.

Schran:

. If

we

understand

Mr.

Schran

correctly,

he

diverges

from

us

on

only

one

important

point:

the

usefulness

of

commonly

accepted

schemas

for

classifying

simulations.

In

our

assessment

of

developments

in

urban

simulation

(and,

by

implication,

simulation

in

social

science

generally)

we

took

the

position

that

classificatory