QUERY Meaning of Stock/Level

This forum contains all archives from the SD Mailing list (go to http://www.systemdynamics.org/forum/ for more information). This is here as a read-only resource, please post any SD related questions to the SD Discussion forum.
""Alan McLucas"" <a.mclucas@a
Junior Member
Posts: 10
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Alan McLucas"" <a.mclucas@a »

Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>

There appears to be an increasing trend for some authors to proffer
models having ""stocks"" containing soft variables or intangibles as being
legitimate for their research. I have seen two examples of this in the
past week alone (one in a review of a journal article, and one in the
examination of a thesis).

Stocks in system dynamics models represent quantities of material. The
stock-and-flow representation has precise and unambiguous meaning:
stocks accumulate or integrate their flows; the net flow into the stock
is the rate of change of the stock (Sterman, 2000: 195).

If this is true, then the veracity of any system dynamics model we build
will be dependent upon what we mean by the term ""material"". Ambiguity
can occur when we build models where a stock contains material that is
not strictly physical in nature. Sterman (2000: 199 -200) gives an
example of a manager having an expectation about how customer ordering
might change over time. The stock ""Expected Customer Orders""
<<widgets/week>> is affected by the inflow ""Change in Expected Order
Rate"" <<widgets/week/week>>. In this example the <<widgets>> do not
physically exist, that is, there is no physical material flowing.
Unfortunately, the possibility that we can have stocks of materials
that are not physical in nature is sometimes interpreted as meaning that
it is quite legitimate to build models where the stocks are soft
variables or intangibles.

In the preceding example, if we were to substitute an intangible such as
""Trust"" for ""Customer Orders"" or ""Expected Trust"" for ""Expected Customer
Orders"" then the SD model fails. In the case of intangibles, inflows
and outflows CANNOT produce accumulations that are calculated by
numerical integration: a model having a stock ""Trust"" with an inflow
""Change in Trust"" will fail the essential mass-balance test.

The only way that such a model could be correct is if initial stock of
""Trust"", measured as a*1<<trust units>> had a further amount of trust
added at the rate of x*1<<trust units / units of time>> and an amount of
trust deducted at a rate of y*1<<trust units / unit of time>>, with the
current stock being calculated as Trust = INTEGRAL (Inflows - Outflows,
Initial Trust) (see Sterman, 2000: 195). ""Trust"" is an intangible which
cannot be represented by this integral equation. Therefore, we cannot
build system dynamics models that contain stocks of intangibles such as
""trust"". I appreciate the desire to build models which incorporate soft
variables and intangibles, because ignoring such variables leads to
erroneous models (Forrester), but building models that purport to
represent soft variables and intangibles as stocks is not the answer.

Am I missing something?

Regards,

Alan

Dr Alan McLucas
School of Information Technology and Electrical Engineering,
Australian Defence Force Academy,
Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>
posting date Wed, 9 Apr 2008 15:18:08 +1000
_______________________________________________
<jayforrest@jayforrest.com>
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by <jayforrest@jayforrest.com> »

Posted by <jayforrest@jayforrest.com>

Alan highlights an issue that is, in part, why I moved more to qualitative
system dynamics (or perhaps more precisely, qualitative method involving
logic from system dynamics) for I, as a futurist, routinely deal with
situations where intangibles are significant and fuzziness clouds
quantification. For me, quantification of intangibles raises substantial red
flags and usually demotes the output to speculative status.

While I work primarily with causal loop or influence diagrams, I would
suggest there is value in modeling intangibles in conventional SD format for
the explicit implications of the SD language provides increased clarity and
basis for inference and analysis over purely causal diagrams. In Alan's
example, he lists Trust. In a pure causal loop diagram there may be some
ambiguity as to whether it behaves as a stock or not. When represented as a
stock it implies a sense of memory and stabilizing influence - but a
precarious one that can disappear rapidly in the event of a catastrophe
(like a lake when a dam breaks). It also implies (which should be in the
model) a flow of trust in and trust out with some coflow with events. While
I find that causal loop diagrams facilitate discussion of such issues, the
rigourous logic of SD contributes to stronger logic when contemplating such
issues. On the other hand, I don't feel I need to quantify the model to
learn from it. Rather it is for internal checks for logic and consistency
and for communication with others who understand at least to some degree the
SD logic.

In summary, I agree with Alan that quantifying intangibles invites serious
problems both in the nature of the quantification, the boundaries of the
model, and the interpretation of the output. In my opinion such models
deserve careful consideration and potentially qualification before accepting
the output or implications. But I also think that representing intangibles
in stock/flow form offers insights in both contemplation and communication
that make qualitative modeling in this stock/flow form useful.

Regards!
Jay Forrest
Posted by <jayforrest@jayforrest.com>
posting date Wed, 9 Apr 2008 15:11:54 -0500
_______________________________________________
""Sheldon Friedman"" <sheldon
Junior Member
Posts: 3
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Sheldon Friedman"" <sheldon »

Posted by ""Sheldon Friedman"" <sheldon.friedman@comcast.net>

The idea of trust as an intangible was raised by Alan McLucas. How about
thinking of trust as the following: -What creates a trust between
people? For me, it might be the number of times I am told something by
someone and then how many of these transfers of information were lies or
truths. Trust builds based on the number of truths, trust is diminished
based on lies. How about a ratio of lies to all transfers of
information? Or a ratio of lies to truths? Certainly, humans use some
form or internal meaure to determine if they can trust-the challenge, it
seems, is to determine what the measure is.

Posted by ""Sheldon Friedman"" <sheldon.friedman@comcast.net>
posting date Wed, 9 Apr 2008 08:46:51 -0400
_______________________________________________
<wakeland@pdx.edu>
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by <wakeland@pdx.edu> »

Posted by <wakeland@pdx.edu>

Regarding the post by Alan McLucas regarding qualitative stock variables
that begins ""There appears to be an increasing trend for some authors to
proffer
models having ""stocks"" containing soft variables or intangibles as being
legitimate for their research...""

I think I appreciate Alan's concern, but nevertheless I believe strongly
that it is appropriate to model qualitative variables as stocks in some
cases.

Stocks are those variables we model by specifying the variable's rate of
change (inflows minus outflows) rather than specifying a formula for the
variable itself.

There are many examples. Consider a variable that we might name
""Current Degree of Awareness."" This is intangible, and yet it has
persistence or memory. Its value today is similar to its value
yesterday, plus an amount related to the rate at which new sensory
stimuli are tending to increase our awareness minus the rate at which
awareness tends to atrophy or dim over time (probably specified via a
time constant called something like ""awareness half life""). Performance
""level"" for a given sport or game is often modeled this way.
Performance improves with practice and atrophies without practice, but
at any given time, current performance level is strongly related prior
performance, so it is natural to model such a variable as a stock.

Wayne Wakeland,
Associate Professor
Systems Science Graduate Program
Portland State University
Posted by wakeland@pdx.edu
posting date Wed, 09 Apr 2008 22:35:22 -0700
_______________________________________________
Tom Fiddaman <tom@ventanasyst
Member
Posts: 25
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Tom Fiddaman <tom@ventanasyst »

Posted by Tom Fiddaman <tom@ventanasystems.com>


At 05:59 AM 4/9/2008, SDMAIL Alan McLucas wrote:
> Stocks in system dynamics models represent quantities of material. The
> stock-and-flow representation has precise and unambiguous meaning:
> stocks accumulate or integrate their flows; the net flow into the stock
> is the rate of change of the stock (Sterman, 2000: 195).
>
> If this is true, ...

I think the problem here is that the premise is false, i.e. that the
restriction to material is overly stringent (except as I'll argue below
that soft variables are material). Another synonym for stock or level is
""state"". The state of a system often implies a quantity of material with
conserved flows, but not exclusively.

> Unfortunately, the possibility that we can have stocks of materials
> that are not physical in nature is sometimes interpreted as meaning that
> it is quite legitimate to build models where the stocks are soft
> variables or intangibles.

If this were true, how would we include soft variables in models? Would
the world even have soft variables?

> ... In the case of intangibles, inflows
> and outflows CANNOT produce accumulations that are calculated by
> numerical integration: a model having a stock ""Trust"" with an inflow
> ""Change in Trust"" will fail the essential mass-balance test.

I agree that inflows and outflows of intangibles couldn't lead to an
accumulation of a physical quantity, but I don't see why that precludes
calculating an intangible state as the integral of the rate of change of
that state.

> The only way that such a model could be correct is if initial stock of
> ""Trust"", measured as a*1<<trust units> > had a further amount of trust
> added at the rate of x*1<<trust units / units of time> > and an amount of
> trust deducted at a rate of y*1<<trust units / unit of time> >, with the
> current stock being calculated as Trust = INTEGRAL (Inflows - Outflows,
> Initial Trust)

This seems reasonable, except for the dimensionless constant 1 buried in
the equation, for which I do not understand the purpose. There's nothing
about a stock/level/state that requires a separate inflow and outflow.

> ""Trust"" is an intangible which
> cannot be represented by this integral equation. Therefore, we cannot
> build system dynamics models that contain stocks of intangibles such as
> ""trust"".

Again, this seems to hinge on the premise that stocks must be material,
which strikes me as overly limiting. Certainly there are intangible
variables in Industrial Dynamics.

> I appreciate the desire to build models which incorporate soft
> variables and intangibles, because ignoring such variables leads to
> erroneous models (Forrester), but building models that purport to
> represent soft variables and intangibles as stocks is not the answer.

This begs the question, what is the proper representation of soft
variables?

Let me suggest a counterargument:

Suppose that beneath every soft variable is a system of material states
that obey conservation laws. Thus for ""trust"" one could imagine a
conserved accounting of neurons in a particular electrochemical state.
Neurons would arrive in or depart from that state, comprising an inflow
and outflow or net rate of change. Then one could in principle construct
a mapping between the neuron accounting and the state and rate of the
soft variable, and be confident that the soft representation did in fact
imply material conservation. In fact, I don't see how it could be
otherwise, unless the universe permits persistent states to exist
without a material substrate.

Also, a thought experiment, borrowing from an old observation about
averages:

Suppose that I'm sitting in a bar. Looking around, I notice that most of
the clientele are grad students, so I mentally calculate that the
average income is about 14,000 $/person/year. So, I have an intangible
state representing my expectation, denominated in $/person/year. Later,
I look up and notice that Bill Gates has walked in. It takes me a minute
to update my expectation of the average income to 120,000,000
$/person/year, at a rate of 119,986,000 $/person/year/minute. Now, would
it matter where that 119,986,000 came from? I don't think so. Going back
to the neural analogy, the representation of my expectation might be a
bit like a binary string, where a change of large or small magnitude
could be accomplished with equal material implication by flipping more
or less significant bits.

You could argue that this is a poor example, because it's too discrete.
But that problem applies to material stock/flow systems as well. If the
expectations of 100s of barflies were averaged, the aggregate movement
of the system would be well-represented by continuous integration.

( http://edstrong.blog-city.com/paul_krug ... _usa_1.htm )

Tom
Posted by Tom Fiddaman <tom@ventanasystems.com>
posting date Wed, 09 Apr 2008 20:42:39 -0600
_______________________________________________
Bill Braun <bbraun@hlthsys.co
Member
Posts: 43
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Bill Braun <bbraun@hlthsys.co »

Posted by Bill Braun <bbraun@hlthsys.com>

Alan McLucas asks about soft variables and stocks.

Interesting question. Imagine a group of people who work together and
there are two cups on each person's desk labeled ""Trust"" and ""Doubt"". We
start with ten chits in ""Trust"" and zero in ""Doubt"" (i.e., the
assumption of complete trust). Based on our interactions any of us can
take a chit from the Trust cup and drop it in the Doubt cup. Likewise we
can take a chit from Doubt and drop it in Trust.

As a stock/flow diagram it looks like [Trust] > Losing Trust > [Doubt] >
Gaining Trust > [Trust].

The material chits are a proxy for information/perceptions. At any point
in time we could bring the system to rest and assess the relative trust
levels in the group on a continuum ranging from 1.0 (complete trust) to
0.0 (complete doubt). We could expand the model to include ""Expected
Trust"" or ""Expected Doubt"" based on the flows.

Does that speak to your question, Alan, and is it a valid SD
representation (albeit metaphorical) of modeling an intangible?

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Wed, 09 Apr 2008 09:02:16 -0400
_______________________________________________
Stephen Wehrenberg <stephen.w
Junior Member
Posts: 8
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Stephen Wehrenberg <stephen.w »

Posted by Stephen Wehrenberg <stephen.wehrenberg@verizon.net>

Alan,

Wouldn't your proscription depend on the purpose of the model? If the
purpose is to attempt to predict, soft variables would create the
problems you assert. But if the purpose of the model is otherwise,
perhaps not.

I have used SD models purely to gain a better understanding of the
relationships of many of the soft variables you mention (I'm a
psychologist). I have also used models to promote a group's common
understanding of the nature of a problem. In neither case was the
purpose of the model not achieved because of the introduction of
variables that have no physical property.

Steve

Stephen B. Wehrenberg, Ph.D.
Posted by Stephen Wehrenberg <stephen.wehrenberg@verizon.net>
posting date Wed, 09 Apr 2008 18:40:36 -0400
_______________________________________________
""Galan, Ricardo"" <RICARDO.G
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Galan, Ricardo"" <RICARDO.G »

Posted by ""Galan, Ricardo"" <RICARDO.GALAN@saic.com>

Alan,

I define stocks as a representation of the accumulation of any amount.
If you could stop time, you will only be able to measure the quantities
in stocks. In the classic bathtub model, the stock is the level of
water. If you stop time, you will be able to measure the amount of water
in the bathtub, but not the flow of water from the faucet or out trough
the drain. Similarly, when looking at soft variables like ""customer
satisfaction"" or ""brand loyalty"", you can stop time and know what the
levels for these variables are. Many companies keep track of these
levels, and know what they are at any particular point in time. These
need to be accumulated separately, since they are independent of many
other variables that should be used for calculating expected sales.

Also, it might take some time to change the actual level of soft
variables like ""customer satisfaction"" or ""brand loyalty"". A change in
product quality might change brand loyalty, but just by a small amount
if the change in quality is sustained for a small amount of time.
Therefore, if the company makes fixes problems with product quality
quickly, they will not see much of a difference in the brand loyalty
that the have accumulated.


_______________________________________________

Ricardo Galan
Science Applications International Corporation
1710 SAIC Drive | MS 2-6-9 | McLean, VA 22102
Posted by ""Galan, Ricardo"" <RICARDO.GALAN@saic.com>
posting date Wed, 9 Apr 2008 13:40:56 -0400
_______________________________________________
""Alan McLucas"" <a.mclucas@a
Junior Member
Posts: 10
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Alan McLucas"" <a.mclucas@a »

Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>

Ricardo's comment reinforces the dilemma.
The concern I still have is that whilst we might legitimately use
stock-and-flow diagrams to enable conceptualization, the strict
mathematical relationships that control those stocks and flows simply do
NOT apply to soft variables such as ""brand loyalty"". This means that
the conceptualisation cannot be translated through to building a
quantified stock-and-flow representation of ""brand loyalty"", or other
soft variables or intangibles that do not adhere to the mathematical
rules for integration. This is the issue raised by Geoff Coyle in 1999
in Wellington, NZ. Clearly we still struggle to establish the limits to
qualitative SD.

I thank Tom for pointing out that we need to create proper
representations soft variables.

Certainly, we need to include soft variables and intangibles in our
models. Jay Forrester tells us that not to include them can only
guarantee one thing, our model will be wrong. He also tells us that any
model we build must produce dynamic behaviours that appear like those we
encounter in the real-world problem. Further, and most importantly, the
algebraic relationships in our models must be derived from real-world
cause and effect.

This last point creates explicit demands for validating models. This is
my main concern: we must be able to validate the behaviour of the model
and demonstrate how it relates to the real world, and this must be so
for all aspects of our model, including our representations of soft
variables.

I am yet to be convinced that representations of soft variables that
often appear in SD models are routinely and comprehensively validated,
rather than just appearing to be plausible. I would dearly love to be
proven wrong. If I am wrong, then this creates an opportunity to
compile a compendium of models that have been validated and, hence, we
can take as being proper representations of soft variables.

Dr Alan McLucas
Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>
posting date Thu, 10 Apr 2008 08:14:23 +1000
_______________________________________________
Bob Eberlein <bob@vensim.com&
Member
Posts: 26
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Bob Eberlein <bob@vensim.com& »

Posted by Bob Eberlein <bob@vensim.com>

Hi Everyone,

This discussion is interesting, but seems to have a lot of force in a
direction I consider extremely misleading, if not simply wrong.

First, the whole concept of a stock, level or state variable is an
abstraction, so there is no absolute test of appropriateness. Just as
Newtonian physics works very well on slow moving macroscopic objects so
are stocks and flows a good way to come at a variety of problems. Any
concept that can be quantified in some sensible way is a candidate for a
model variable, whether or not it is practically measurable. And any
variable that can change only over time, needs to be treated as a stock.
In fact, even in everyday language, we speak of things such as building
anger, rising emotions, falling into despair. Stocks every one.

The direction that is troubling is the belief that models containing
these types of variables should be treated as qualitative. Much of the
power of System Dynamics comes directly from the ability to combine
concrete measured concepts with intangibles. For example, many very good
project models include worker morale right along side the nuts and bolts
of getting work done.

That duality is a strength. Including it is vital to capturing the true
dynamics of problems being addressed. Combining tangible and intangibles
also provides a mechanism to measure the values of those intangibles
indirectly through the model. The contribution of that indirect measurement
can be tremendous.

It is not appropriate to think that every element of a model needs to be
validated against existing observations. Many important things are not
observed, or at least not observed well. It is useful to validate every
element of a model against some understanding of reality based on what
is observed and everything else that seems important. The inclusion of
soft variables does not take away from our ability to compare the hard
variables to measurements. The appropriate inclusion of soft variables
in a nuts and bolts model makes the model that much more valuable.

Sorry for the venting - but I had to get that out before I exploded.

Bob Eberlein
Posted by Bob Eberlein <bob@vensim.com>
posting date Thu, 10 Apr 2008 09:20:36 -0400
_______________________________________________
""Chip Hines"" <hines.chip@gm
Junior Member
Posts: 5
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Chip Hines"" <hines.chip@gm »

Posted by ""Chip Hines"" <hines.chip@gmail.com>

It would seem to me that in many situations soft variables have a
significant role, and that if they are not accounted for there is no
hope of understanding the entire system dynamics through a model. The
points about introducing more uncertainty into the model are indeed
valid, but if you ignore them the model isnt truely representative of
the situation. Creating the model with soft variables does carry with
it the obligation to be as thorough as possible and to be sure that
their inclusion is known. Creating one without them requires that
analysis and documentation of the effects of exclusion are well known.

I also believe that while the person building the model may not always
understand the dynamics of soft variables, it is likely that the end
user (assuming they arent the same person) will want to be able to use
the model to explore how these variables effect the system.

chip hines
Posted by ""Chip Hines"" <hines.chip@gmail.com>
posting date Thu, 10 Apr 2008 08:32:41 -0400
_______________________________________________
George Richardson <gpr@albany
Member
Posts: 20
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by George Richardson <gpr@albany »

Posted by George Richardson <gpr@albany.edu>

On Apr 9, 2008, at 7:59 AM, SDMAIL Alan McLucas wrote:
> ""Trust"" is an intangible which
> cannot be represented by this integral equation. Therefore, we cannot
> build system dynamics models that contain stocks of intangibles such as
> ""trust"". I appreciate the desire to build models which incorporate soft
> variables and intangibles, because ignoring such variables leads to
> erroneous models (Forrester), but building models that purport to
> represent soft variables and intangibles as stocks is not the answer.

On the contrary, soft concepts can indeed be represented as stocks and
sometimes must be. The representation has nothing to do with numerical
details about integration, but rather has everything to do with the way
such a soft variable persists through time and changes over time within
the time frame of the model.

Way back in 1961 Forrester addressed the kind of thinking that surrounds
the selection of stocks (levels) in a system dynamics model. He said,
in part:

> ""Levels exist in the information network as well as in the physical
> network of material, etc. 'Awareness levels' exist in the mental
> attitudes that influence decisions. Levels of satisfaction, of
> optimism, and of recollection of a past disastrous depression, all
> influence economic behavior. ... All memory and continuity from the
> past to the future exist in the levels of the system."" (Industrial
> Dynamics p. 68).

That memory would include things like trust and similar soft variables.
Forrester talked about the ""snapshot"" test for stocks, and that bit of
wisdom has been repeated in most texts since, e.g., Richardson and Pugh
pp. 176-177. They conclude

> ""Thus potential levels in a system include such obvious accumulations
> as poeple, inventory, production capacity, serum cholesterol, and bank
> balances. Included as well, however, are such things as cultural
> traditions, average sales rates, habits, and perceptions, for these
> too would not disappear if time were stopped.

There are numerous excellent examples in our literature of studies
involving models that capture soft variables as stocks. A recent
example from our work at the University at Albany (Luna-Reyes et al.,
Anatomy of a Group Model Building Intervention: Building Theory from
Case Study Research, SDR 22,4) is of interest because it reveals ""trust""
as a crucial variable in the case and the emergent theory, and the
authors and the client group chose to model that as an accumulation.

That study and others have built upon a theory of the way soft variables
like trust or commitment grow as artifacts (costocks and coflows) of
working together. ""Producing"" can do two things in a system: it can
create objects that accumulate in an inventory, and it can increase (or
maybe decrease) the accumulation of trust or commitment, as people work
together. Examples of the theory and its application in serious group
modeling interventions can be found in a couple of conference papers by
Luna-Reyes et al. in the proceedings of the 2004 System Dynamics
Conference.

In these interventions, and a host of others, it has been essential
that soft variables such as trust are treated as accumulations.

..George

George P. Richardson
Chair of public administration and policy
Rockefeller College of Public Affairs and Policy
University at Albany - SUNY, Albany, NY 12222
Posted by George Richardson <gpr@albany.edu>
posting date Thu, 10 Apr 2008 20:37:48 -0400
_______________________________________________
""John Gunkler"" <jgunkler@sp
Member
Posts: 20
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""John Gunkler"" <jgunkler@sp »

Posted by ""John Gunkler"" <jgunkler@sprintmail.com>

I think Tom Fiddaman has helped clarify one of Alan's points -- namely, that
what we often call ""stock and flow"" representations are also called ""state
and rate"" representations, and that ""states"" need not be material.

But is there an underlying, and quite legitimate, issue here of
""measurement"" (which I think Fiddaman is referring to when he mentions
""representation of a soft variable"")?

In SD, and in most of the other fields I've practiced in, there is a serious
problem with allowing people to include concepts in their thinking that have
not been ""operationalized."" In the Six Sigma process improvement
methodology, for example, there is a requirement that everything we include
in the analysis have ""operational definitions."" By this is meant, simply,
that we must specify a procedure (""operation"") for measuring the concept --
and it is this ""measurement"" result that will always enter into our
equations.

When SD modelers include ""soft variables"" without specifying how they are to
be measured (or operationalized), I agree with Alan that the resulting model
is problematical. This does not necessarily mean, as Jay Forrest and others
point out, that the models are useless. They may help people clarify their
thinking to a degree.

But I am reminded of a cartoon which showed three people on a basketball
court -- one of whom was a short man with seven legs. The second person was
saying to the third, ""Coach, I found that seven-footer you wanted.""

Misinterpretations caused by failing to operationalize the definitions of
soft variables could lead to results equally absurd, don't you think?

John Gunkler
Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
posting date Thu, 10 Apr 2008 11:53:06 -0400
_______________________________________________
""Douglas Franco"" <dfranco@c
Junior Member
Posts: 8
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Douglas Franco"" <dfranco@c »

Posted by ""Douglas Franco"" <dfranco@cantv.net>

Alan,

Numbers are not necessarily, 1, 2, 3.1 or I, IX, XX like in Romans. The
math essential requirement to make a set numerical is to have a relation
of order among the elements of the set. If you can say that a person or
an institution do have more trust than other, then trust is a quantity
rather than a quality. There are many qualities which in a deeper sense
are really quantities, because you can distinguish between more of such
variable from less of such variable. People usually tend to forget
qualities; therefore human perception of a quality is usually under a
vanishing pressure. Stock representation of such process imitates the
building and loosing trust. The error involved in neglecting these soft
variables is usually greater than the approximations from stock and flow
representation.

There are improvements needed in SD to deal properly with information
flows, some of them you already mention in your email. It is very
difficult to come up with an undisputable solution to these problems,
but SD is quite useful if you can settle for success when looking for
significant improvements.

Douglas Franco
Posted by ""Douglas Franco"" <dfranco@cantv.net>
posting date Thu, 10 Apr 2008 15:20:55 -0400
_______________________________________________
Richard Dudley <rgd6@cornell.
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Richard Dudley <rgd6@cornell. »

Posted by Richard Dudley <rgd6@cornell.edu>

Alan states that stocks should represent only ""quantities of materials"" and
should not be used to represent soft variables.

But I believe that the use of stocks to represent soft variables is
widespread in SD modeling and, in fact, is encouraged. I believe that the
use of stocks in this manner is perfectly legitimate.

Something like ""trust"" is a real thing in ones mind. It ebbs and flows in
response to many influences. At any given point in time a person will have
a given amount of trust in a policy or a politician (for example), and this
trust will affect other model components such as ""compliance with policy"" or
""support for a politician"".

I don't see any reason why soft variables cannot be represented as stocks.
Perhaps I misunderstood a the question?
Posted by Richard Dudley <rgd6@cornell.edu>
ÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌÌPosted by Richard Dudley <rgd6@cornell.edu>
_______________________________________________
""Malli"" <malli@madisonindia
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Malli"" <malli@madisonindia »

Posted by ""Malli"" <malli@madisonindia.com>

I have been studying system dynamics for the last 2 years, and have been
following this discussion on qualitative variables with great interest.
The fact that system dynamics as a modeling worldview allows the modeler
to incorporate both measured variables as well as those that are not
measured is what attracted me to this field. However, while quantifying
a soft variable such as morale, should one also think about the scale?
Essentially, is one referring to the variable using an interval scale or
a ratio scale? As one knows, on the interval scale, the zero is
arbitrary - If the temperature today is 100 degrees Fahrenheit, and
yesterday it was 50 degrees Fahrenheit, then one cannot say that today
is twice as hot as yesterday.

On the other hand, if one is using a ratio scale, then such statements
can be asserted. A market share of 20% is twice that of 10%

While quantifying soft variables, such as morale, if the morale dips
from 100 to 98, then what is the scale that one is referring to? Should
I be alarmed or not? If the body temperature increases from 98.5 Degrees
F to 102 Degrees F, I definitely am! however, seen from a ratio scale
this increment would be a small blip!

This is one aspect that has always puzzled me about the quantification
of soft variables

Malli
Posted by ""Malli"" <malli@madisonindia.com>
posting date Fri, 11 Apr 2008 17:46:45 -0700
_______________________________________________
""David W. Peterson"" <david@
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""David W. Peterson"" <david@ »

Posted by ""David W. Peterson"" <david@ventanasystems.com>

Here are four observations on state variables / stocks / levels:

1) Contrary to what some have implied, it is actually difficult and subtle
to decide what is a stock and what isn't. Outside the context of a specific
model or purpose, one cannot sort the real world into ""things which are
stocks"" and ""things which are not stocks"". What is a state variable depends
on time scale, perspective, and level of detail. For example, in a model
concerned with logistics stability, an inventory would be an obvious state
variable. But for a long-term growth model, inventory would be interesting
only as its average size, and could be represented as the flow through the
inventory, multiplied by a suitable dwell time (how long the average item
stays in the inventory).

2) The choice of state variables is crucial to value of the model. Once the
state variables are chosen and defined, the model is practically locked in
to an implied level of detail, time resolution and perspective. One
approach to making models (which I endorse) is to choose the state variables
first (instead of, for example, feedback loops). The next step is to
identify the flows in and out of the state variables, and then writing the
equations for the flows may take a while, but is straightforward. Feedback,
because it is important and ubiquitous, will emerge automatically in the
resulting equations.

3) For most models, there is a flexibility or freedom of choice in state
variables that has no effect on the results -- the same dynamics and
conclusions can arise from multiple choices of (dynamically equivalent)
state variables. As a trivial example, a model might have state variables
of (a) workforce and (b) total people-years of experience in the workforce.
But the exact same behavior could be based on (A) work force and (B) average
years experience per person in the workforce. In the former case, an
auxiliary variable is required to compute the average experience per worker;
in the latter case, the state variable directly gives the desired average
per person, but its flows are more complicated.

4) Within a model, the very same variable/concept can be simultaneously a
state variable and a flow. In the first example in the preceding point (3),
the workforce is a state variable but it is also the flow into the ""total
people-years of experience in the workforce"" accumulation of people-years.

David W. Peterson
Posted by ""David W. Peterson"" <david@ventanasystems.com>
posting date Fri, 11 Apr 2008 14:17:04 -0400
_______________________________________________
""Compton, Dan S"" <dan.s.com
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Compton, Dan S"" <dan.s.com »

Posted by ""Compton, Dan S"" <dan.s.compton@boeing.com>

Bob,

Good put. However your post is sad on two counts. That it had to
exist. And that you felt compelled to apologize for being passionate
about some foundational truths of System Dynamics modeling and
simulation.

Clients who allow explicit modeling of soft stocks seem pleased that
these stocks allow for additional relevant relationships. The whole
relevant system is modeled; not just the part that is directly
measurable.

--Dan
Posted by ""Compton, Dan S"" <dan.s.compton@boeing.com>
posting date Fri, 11 Apr 2008 05:46:17 -0700
_______________________________________________
Jim Thompson <james.thompson@
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Jim Thompson <james.thompson@ »

Posted by Jim Thompson <james.thompson@strath.ac.uk>

My reaction to the original query and some responses is similar
to Bob Eberlein's response of 11 April, and prompts me to add a
couple of thoughts.

When Bob and others refer to ""models containing these types of
variables"", we mean that there are feedbacks from soft variables
that affect system behaviour. To include a variable means that
the modeller thinks it represents part of his experience.

Variables such as 'trust' and 'morale' often reflect measurable
experiences. For example, 'trust' can reflect 'promises kept',
and one can measure promises made, kept and broken...at least in
theory. Morale may be partly a function of actual performance
versus expected performance, and these too can be measured.

As Bob pointed out, not everything that affects a system can be
measured directly and not everything that can be measured
directly affects the system. One contribution of system dynamics
methodology is finding out what matters.
Jim Thompson
Posted by Jim Thompson <james.thompson@strath.ac.uk>
posting date Fri, 11 Apr 2008 12:03:14 -0400
_______________________________________________
Tom Fiddaman <tom@ventanasyst
Member
Posts: 25
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Tom Fiddaman <tom@ventanasyst »

Posted by Tom Fiddaman <tom@ventanasystems.com>


At 10:40 PM 4/9/2008, Alan McLucas wrote:
> I am yet to be convinced that representations of soft variables that
> often appear in SD models are routinely and comprehensively validated,
> rather than just appearing to be plausible. I would dearly love to be
> proven wrong. If I am wrong, then this creates an opportunity to
> compile a compendium of models that have been validated and, hence, we
> can take as being proper representations of soft variables.

Ahh ... that is a different beast. I'd hesitate to assert that much of
anything gets ""routinely and comprehensively"" validated (not that that's
OK). But that doesn't mean that soft variables are unvalidatable. All
are at least subject to tests of robustness in extreme conditions and
dimensional consistency. Some, like expectations that get written down
(official forecasts) can be directly checked against data. In most other
cases, soft variables can be at least be indirectly verified. That is,
one can say something about the plausibility of a structure including
soft variables by observing it in context; if the aggregate model yields
implausible behavior then the soft variables may be at fault. One can
use the same strategy to indirectly measure soft variables (e.g.,
estimating the parameters of a perception process), which is not
fundamentally different than drawing conclusions about a physical
process by observing its inputs and outputs.

The idea of creating a compendium of validated soft variable structures
is attractive, as long as the compendium includes information about the
circumstances in which each was derived and its possible limitations. I
suspect that this would be rather straightforward for some of the
typical expectation-formation models, and perhaps also for more
existential concepts like hunger. It would also be quite useful. I think
at least one common formulation, the TREND function typical in SD
software, is flawed and ought to be replaced in most usage. I think the
idea of a compendium gets much stickier when we start talking about
love, fear and loathing, conservatism, etc. There we have not only
ambiguity about structure but also about definition, which may not be
widely agreed upon. Modeling love might actually be a good way to refine
and communicate that definition, though that may just be my nerd
perspective, certainly less satisfying than the corresponding
experience. Some of the ambiguity around soft variables may eventually
be resolved as technology for directly measuring the state of the brain
improves.

I think the caution in Alan's note is worth taking to heart. It's easy
to fill a model with SMOOTHs and other soft variable structures without
paying much attention to whether they are meaningful. At the very least,
one ought to explore extremes and try alternative and parameters to see
whether the soft variables are dominating the behavior. For example, in
some economic models, shortening perception delays and eliminating
biases leads to an approximation of general equilibrium, and certainly
the difference between behavioral and equilibrium perspectives is
important.

Regarding measurement, quantification, representation, operationalization:

I agree with Bob that it's not necessary to directly measure every
variable in order to have a useful and valid model. What I had in mind
when I said representation was really ""operationalization"" and
""quantification."" Those deserve careful attention when constructing soft
variables, particularly for things other than expectations of measurable
quantities. Operationalization and quantification mean more than just
assigning a scale. One must also consider the underlying process and
information sources that would cause the variable to change, and
possible limits to the scale in extremes. Jay Forrest provides a nice
example of the thought process:

> ... I think there may be a better way of modeling
> trust. My thought arrives from a perception that trust has a limit - a
> cap -
> where additional acts of trust building don't create further trust. So
> perhaps there is a stock of trust building acts that has some form of
> time
> depletion and a stock of trust destroying acts with a different time
> depletion factor ... Trust is I think directionally a (graphically S
> shaped) fxn of the balance
> of the two where acts of either kind CAN affect the outflow rate of the
> other.

This has at least four nice features:
- it provides a description of the mechanics by which trust goes up or
down on the basis of events that could potentially be measured, for
example by asking individuals to recall them
- by distinguishing trust building and destroying acts, it recognizes
that there may be asymmetry in time constants for response to each
- it generates saturation naturally as a consequence of the behavior of
truths/(truths+lies) rather than via an arbitrary function
- it captures an important internal dynamic, that the stocks of building
and destroying acts might influence one anothers' outflows (I'd add that
they might also influence inflows, i.e. when one has a large stock of
trust building acts, one might at first neglect to notice trust
destroying acts, just as scientists with strong faith in a model might
at first reject apparent anomalies in its performance)
These features could easily be absent in a more simplistic and less
operational representation, like a single stock of trust adjusting to an
indicated level, with saturation captured by a lookup table.

While measurement is not necessary, I think the Six Sigma attitude is
useful. I'm reminded of a company I worked with a few years ago. People
there actively talked about the sense of loyalty instilled by the
company's stable employment and internal promotion policies. Then, out
of the blue and within a few days of Christmas, they suddenly hacked off
a huge chunk of staff, destroying morale and any semblance of a social
contract. We all ""know"" that morale is important, but quarterly results
are important and measurable. If someone had built a model of the
company with morale in it, and searched for a way to operationalize and
measure morale, it would be possible to get a read on (a) whether morale
actually matters to performance and (b) whether that act actually
destroyed morale. The mere prospect of measurement might have influenced
the decision.

Tom
Posted by Tom Fiddaman <tom@ventanasystems.com>
posting date Thu, 10 Apr 2008 20:41:16 -0600
_______________________________________________
""Ralph Levine"" <leviner@msu
Junior Member
Posts: 2
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Ralph Levine"" <leviner@msu »

Posted by ""Ralph Levine"" <leviner@msu.edu>

We have been looking at this discussion about the use of soft variables
in system dynamic models with intense interest. As community
psychologists, we see many places where the integration of soft,
psychological variables with conserved variables that conform to the
bathtub metaphor makes sense. The inclusion of soft variables can
contribute to better understanding the dynamics underlying the problem
at hand for both modelers and stakeholders. System dynamics is
primarily noted for its emphasis on feedback mechanisms that determine
behavior. Most system dynamic models deal with feedback associated with
human actors, who perceive, make judgments, and take actions, at various
parts of the system to control the flow and ebb of the material
processes of interest, the levels or stocks, such as the amount of water
in a reservoir, the size of the inventory on hand, and the size of the
population in an urban area. All of those examples, are what we would
call “energy/material” variables that are conserved. On the other hand,
feedback mechanisms, which are a very much part of system dynamics, are
informational in nature. As noted by Jay Forrester, in a number of
places, e.g. Principles of a basic feedback loop is a circular path
composed of a decision, action, a state variable, and information about
the state variable. We have found in our modeling efforts that soft
variables play a valuable role in elaborating the informational aspects
of loop processes.

A number of points dealing with whether or not one should represent soft
variables as levels have already been given. We are going to address
other issues in this controversy, such as measuring soft variables, and
the costs of not including soft variables in a model when relevant.

Let’s address measuring soft variables. Some of the posts described soft
variables as “intangibles.” They may be intangible, unlike holding
dollars in your hand, or putting your hand in a bathtub filled with
water. However, the notion of being intangible does not imply that soft
variables cannot be quantified or measured. Douglas Franco addressed
the quantification issue in his example of quantifying trust. We think
he is right on target. In our work, we frequently quantify soft
variables by anchoring them alternatively between 0.0 and 1.0, or 0.0
and 100. So, if we want to quantify the level of depression, 0.0 would
be absolutely no degree of depression and 100 would be the situation
maximum level of depression one can feel, which may or may not be
attained.

Although related, there is a difference between quantifying a soft
variable and developing a operational measure of the variable. We think
that John Gunkler has a point in saying that we must specify how the
variables are measured (operationalized). Actually, the technology of
measuring and scaling social and emotional variables is widely used by
research and applied psychologists. It has been around since the
1920’s, if not before, and it is not a big deal, if you have the
training. The mathematical theory underlying psychological measurement
is well known, and from a practical statistical point of view, there are
a number of tools, like confirmatory factor analysis that can test for
internal consistency and reliability of one’s scale. Concerning the
validity of one’s scale, it is a matter of using correlational analysis
to correlate one’s scals with measures of other concepts that in theory
should correlate with ones scale, if the scale actually measures what is
was designed to measure.

In general, psychologist scales, based on Likert scaling technology,
fall somewhere between ordinal and interval scales, not ratio scales.
This is fine for most correlational analysis, but a bit of a problem for
using such scales as data to be fit by the model. If the soft variables
in the model, like trust, are quantified as ratio scales, a perfectly
good model could be rejected, because the data, based on interval
scales, may not conform to trajectories that were generated by the
model, assuming the soft variables would be measured on ratio scales.
We have written about this problem before. Stevens, who developed
methods for measuring psychological and social variable on ratio scales,
also found a way to transform data that is measured on an interval scale
to a ratio scale.

Now let’s discuss some of the implications (and unintended consequences
) of leaving out soft variables out of “quantitative” SD models. Let’s
take the suggestion that one should not include soft variables in one’s
quantitative SD , simulation model. One unintended consequence is that
it may limit or eliminate the use of a coflow process that may be the
core of one’s model. The problem being modeled may actually deal with
the lowered quality of new hires, for example. The coflow process is a
very important, helpful tool in SD modeling. However, unfortunately
most of the attributes, such as “Total Effective Experience,” are soft
variables. The example of Total Effective Experience was taken from
John Sterman’s text, pages 497-212. Would it be better for the field of
system dynamics to encourage Dr. Sterman to cut out a good portion of
his twelfth chapter, or at perhaps keep it as an abbreviated section on
coflows as an example of what not to include in SD models.

Actually, in the classic coflow model, the attribute, as noted by Jim
Hines, can be represented b y a first order information delay, the
smooth. The smooth is not conserved, yet we believe that most system
dynamicists would not want the smooth to be eliminated in one’s SD
model. Indeed, it is an important tool representing the dynamics of the
stakeholder’s problems. Perhaps to get around Alan’s objections to
including soft, nonconserved variables in SD models, one might continue
to use practice of representing psychological processes as smooth.
Along the same line, one could also continue a very traditional way of
representing soft variables by defining them as converters (auxiliaries)
first, and then sending the value of the converter to an information
delay. An example of this can be found in Jay Forrester’s urban dynamic
model. First, he created a soft variable, AMM, the Attractiveness for
migration multiplier as a converter. However, it takes time to perceive
attractiveness, so he fed AMM into a first order information delay,
which he denoted as AMMP, the Attractiveness for migration multiplier
perceived. We also should note that AMMP had a significant dynamic
impact on the stock of underemployed people.

There is certainly more to say about going beyond the use of first order
information delays to model soft variables. There is the decision
whether to only include them in simulation models or only in CLD’s and
stock and flow diagrams. Stock and flow diagrams convey more information
than CLD’s, but operational SD models put you in a much better position.
One can then gain a lot of information concerning strength of various
levers of change by performing a sensitivity analysis on the
quantitative model. Such a model undoubtedly contains a number of
nonlinearities, the effects of which may not be known without the
ability of running the model. Also, one can learn a lot about the
functioning of individual loop processes, and in some cases, one can
assess the dominance of particular loops with some software now being
developed.

Moreover, a model that includes the soft variables can be validated by
performing numerous tests developed by Forrester and Senge, Barlas,
etc.Errors in specifying the equations containing soft variables should
be picked up this rigorous way of validating models. If one has time
series data, one can also perform Theil inequality statistics to
ascertain the source of errors the model makes in fitting the data. If
those who feel that including soft variables in one’s model can lead to
false conclusions, certainly these should be picked up by performing
these tests. We also should mention the potential problems in using the
qualitative model as a policy tool, even with the inclusion of soft
variables. In our opinion, insights gained by a qualitative SD model
are more likely to be in error, because of the inability of validating
the model, or CLD in a rigorous way and not seeing the implications of
nonlinear effects of delays, and loop structure over time.

Finally, a couple of years ago we started a Psychology Chapter for those
who are interested in including psychological and social processes in
their models. Undoubtedly, the issues raised by Allen McLucas should be
brought up at our annual meeting in Athens this year. It is a great
opportunity to discuss and debate the role of soft variables in system
dynamic models.

Respectfully,

Ralph Levine, Ph.D.
Posted by ""Ralph Levine"" <leviner@msu.edu>
posting date Fri, 11 Apr 2008 17:51:45 -0400
_______________________________________________
""Alan McLucas"" <a.mclucas@a
Junior Member
Posts: 10
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""Alan McLucas"" <a.mclucas@a »

Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>

George,

Thank you. We can always rely on you for your sage comments.
Thank you for your examples.
I have been deliberately playing ""devil's advocate"" with the way I have
sought to elicit comments about this topic.
>From a conceptualization viewpoint I agree that representing soft
variables as stocks is a legitimate approach. However, representing the
growth and decay in a soft variable by using stock-and-flow diagrams to
conceptualize is not the same as claiming that such a model is strictly
valid. If the purpose of the model is to hypothesize about how changes
in the soft variable occur, so that some form of intervention might be
designed, then this is a legitimate approach.
However, I am still concerned that models appear from time-to-time that
clearly rely on the stock-and-flow representation of the behaviour of
soft variables satisfying some basic ""truth criterion"", at least in the
minds of the modeller. Sorry, I am not at all convinced that this is
either legitimate or appropriate.
Regards,
Alan

Dr Alan McLucas
Posted by ""Alan McLucas"" <A.McLucas@adfa.edu.au>
posting date Mon, 14 Apr 2008 08:20:16 +1000
_______________________________________________
""John Barton"" <bartcons@big
Newbie
Posts: 1
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by ""John Barton"" <bartcons@big »

Posted by ""John Barton"" <bartcons@bigpond.net.au>

Alan

For those like yourself who get concerned about the credibility of
models that include some categories of soft variables as stocks, you
might find the following approach useful. Specifically, I have used this
approach to model intangibles as drivers of shareholder value in
business models; a context in which clients become very dubious about
measures that appear unreal.

The first thing to note is that an intangible such as ""reputation"" only
takes on meaning when it translates into action (either actual or
contemplated). Furthermore, concepts such as ""reputation"" result from
the cognitive integration of factors which contribute to reputation.
While these factors may initially be unconscious, interview techniques
can help identify these factors are and what relative importance they
have to decision makers. Such factors can be eventually related back to
a number of easily recognised measurables such as years of experience,
performance, and size of client lists etc. In most cases if not all,
these factors can be modelled as stocks. Given the weighting factors, a
""reputation"" index can be formed using these stocks. This index can then
be related to a further action using a graph function that describes the
usually non-linear behavioural assumptions (defined by the graph's
convexity). In the limiting case, this index may have a single component
which is the case when a surrogate measure is used (for example,
absenteeism for morale).

This technique is made rigorous in terms of pragmatist philosophy by
reference to Peirce's pragmatic maxim (meaning is provided by
contemplated or actual action), and his semiotic framework in which a
symbol (the index) produces a certain reaction (the action driving a
rate) determined by an interpretant (the empirically determined weighted
average used in the index).

Apart from avoiding any issues about modelling intangibles as stocks,
modelling the basic factors provides a completely transparent approach
to simulation strategies aimed at improving intangibles and their outcomes.

Best wishes

John

John Barton, PhD
Principal
John Barton Consulting
Business Strategy Dynamics
Office: 2 Arthur Street
Sandringham. 3191
Victoria, Australia
Posted by ""John Barton"" <bartcons@bigpond.net.au>
posting date Tue, 15 Apr 2008 16:30:58 +1000
_______________________________________________
Jean-Jacques Laublé <jean-jac
Senior Member
Posts: 61
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Jean-Jacques Laublé <jean-jac »

Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>

Hi every body.

I use a soft variable if I feel that it is a useful addition to
the model: it means that it increases its usefulness.
This is the first consideration, whatever the technical
problems that it may generate.

It is unfortunately very easy and tempting to push a model
outside its interval of usefulness, which is limited and knowing the limit
is essential. SD can solve a part of many problems but rarely the totality.

A second condition of using soft variable I follow, is that the soft variable
must depend in a large part from other hard ones. It is often expressed as
a ratio or a probability. Example, if I consider the trust of a client into the
quality of a product, I will eventually consider the willingness of the client
to fulfil all his needs with the considered product. This willingness will
eventually depend on the percentage of the past defect of the product.
This willingness is necessarily expressed as a ratio from 0 to 1.

So I will express the actual willingness depending from the percentage of defect
by a lookup and the willingness by a smooth of the actual and past willingness or
as a weighted moving average of the past willingness.
Using a smooth is mandatory if one wants to keep the willingness into
the 0,1 interval. One could use a simple accumulation with a
limiting capacity, but it does not represent what happens in the
reality.

I can make the willingness depend from other factors too and it will be
necessary to weight them. It is always possible to see the influence of
the weighting system on the objectives that summarize the efficiency
of the policies and use sensitivity analysis.

Regards.

Jean-Jacques Laublé Eurli Allocar
Strasbourg France
Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>
posting date Tue, 15 Apr 2008 11:47:59 +0200
_______________________________________________
Jean-Jacques Laublé <jean-jac
Senior Member
Posts: 61
Joined: Fri Mar 29, 2002 3:39 am

QUERY Meaning of Stock/Level

Post by Jean-Jacques Laublé <jean-jac »

Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>

High Jay

You write
> Jean-Jacques' logic
>of striving to make the soft variable dependent on hard variables strikes
>me as an appropriate ambition.

I try to make 'soft variable' dependant on hard one's. But it is not always
possible.
And there is not only the quantification of the soft variable that is
difficult, but how much and how it influences the hard one's.

This is why I insist on the consideration of the utility of adding soft
variable or any other variable even hard.

I do not like theories unless grounded on practical examples.

I have just finished a small model that explores the following problem.
I rent vans and trucks primarily. I have the possibility that the supplier
of new cars buys back at a predefined price the vehicle he has sold. I can give back the
vehicle after a time ranging from 18 months to 36 months. I do not pretend to solve the
problem taking into consideration all the factors of the reality, but I build generally
exploratory models much simplified that enlightens my understanding of the problem and may help
define a policy on a improved grounding than the one I use presently.

In the model I built in about a week time, I took into consideration most of
the main hard factors, the basic demand with the seasonality, the average
revenue by day of renting, the relation of the number of vehicles and the percentage of utilization
possible, maintenance costs etc. But I did not include an important soft variable. Clients prefer
to rent a new car than an old one. If I renew the vehicles in 36 months instead of 18 months,
they will be two times older. It will have an effect on the demand. But how do I evaluate
that effect?
I can use a variable factor and test the sensitivity of the model behaviour
relative to that factor.
But if I add that soft factor, my original model that has 8 loops, will
increase to 13 loops.
Going from 8 loops to 13 loops, will make my model at least twice more
difficult to manage.

I am sure that I can find some more factors and I will finish with a model
that has 100 loops.
I will be very pleased as a modeller, but as a user?

My first model with the 8 loops shows clearly that it is better to renew in
18 months.
If I add the soft variable, increased interest of the client by newer cars,
it will reinforce the result, as it will be even more interesting to renew in 18 months.
So adding that important soft variable does not add anything relative to the
objective assigned to the model.
I wanted to illustrate the fact that the inclusion of soft or hard variable
is always relative to the original objective. Exploratory simplified models help clarify the
problems.
It is too often necessary to reduce the pretension of the objective so as to
reduce the complexity and have a model that helps you a bit but with certitude than
one that might help you a lot, but will fail to do it.
Regards.
Jean-Jacques Laublé Eurli Allocar
Strasbourg France.
Posted by Jean-Jacques Laublé <jean-jacques.lauble@wanadoo.fr>
posting date Tue, 15 Apr 2008 17:44:14 +0200
_______________________________________________
Locked