Plenary Program
Strategic Evaluation of Flexible Assembly Systems
- Combining Hard and Soft Decision Criteria -
Erich Zahn
Jurgen Greschner
Betriebswirtschaftliches Institut
Universitat Stuttgart
Keplerstr.17, 70174 Stuttgart, Germany
Abstract
The evaluation of investments in flexible assembly systems lacks of an appropriate methodology.
First a brief review of decision making processes regarding complex investments is given. Such
decisions have to be made in the tension of hard and soft decision criteria which often produce a
dilemma for the decision maker: Considering only short-term effects in terms of hard criteria will
usually kill investment proposals. In contrast managers feel the need for the investment but have
difficulties to justify their intuitive insights. As a possible solution a Systems Dynamics based
approach is proposed to bridge the gap between rational and intuitive judgement. The approach
combines qualitative and quantitative criteria by using a computer-aided step-by-step modelling
concept.
Introduction
The evaluation of complex investment objects is not an easy task. We are dealing with
interventions into complex socio-economic systems. Here decision makers have to consider:
- hard and soft decision criteria,
- interdependent aspects,
- high capital investments,
- time-consuming implementation,
- long time horizons,
- often significant organizational change, and
- important impacts on the competitiveness.
All in all such decisions can be considered as strategic. They must be submitted to a
comprehensive economic evaluation. However, up to now there are no methods available which
permit this evaluation in a satisfactory way.
Available Approaches for Evaluating Complex Investments
Conventional methods of investment evaluation differ primarily in whether the time aspect of
expenses and earnings of an investment is taken into account or not. Static calculation methods
ignore differences in the time structure of the series of payments. Dynamic methods take into
consideration the values of the payments which are due at different times. Both types of
procedures, however, presuppose clearly defined expectations with regard to the development of
the payments concerning a particular investment. These procedures evaluate individual
302
System Dynamics '95 — Volume I
investment projects that are not interconnected according to one goal criterion and are limited to
quantifiable aspects. Interdependencies between different investment projects are neglected.
Besides, additional premises have to be fulfilled if the respective methods are to be applied
successfully. This results in further limitations.
Conventional methods can be used at best to estimate the expenses of an investment. The
consideration of the benefits is possible only in a limited way. These methods, although widely
used in practice, thus often result in a "killing calculation" of an altogether advantageous
investment. One also speaks of a so-called "brake effect" which appears when these methods are
applied exclusively (Liider 1993).
In the case of soft decision criteria the gap between the existing procedures and the desirable
ones is even greater. Figure 1 shows three examples. The respective effects that an investment
produces are presented by means of lists of impacts (Horvath e.a. 1987). Balance sheets of
arguments are also used to structure qualitative criteria - in this case pro and con arguments are
listed side by side in a structured balance (Wildemann 1986). The most widely used method still
is the scoring model (e.g. Blohm and Lider 1991).
« Lists of Impacts
« Balance Sheets of Arguments
« Scoring Model
Purpose: Structured evaluation of decision alternatives
based upon the preferences of the decision makers
(concerning multi-dimensional objectives)
Procedure: (1) Formulation of decision criteria concerning the alternatives
(2) (subjective) estimation of the importance of the criteria
(3) (subjective) evaluation of reaching the objectives of each alternative
Example: Choosing between alternative plant location
Availvabitty
Criteria | Time Saving | Traffic Security | of Personal Scores
Resources
Importance of Criteria (9) Aggregation Rule
Alternatives N=nit)ta(t) + n(i2)%
at)=03| o@=04 | g@=o3 IO"?
Location C n(11)=8 n(12)=7 n(13)=9 N=79
Location D n(@1)=6 n(22) =8 n@@3)=8 | N=7,4
Figure 1: Scoring Model
The advantage of the scoring model lies in the fact that investment alternatives can be organized
with regard to a multidimensional goal system. However, interdependent effects are neglected
here as well. Besides, this model leads to a certain pretense of objectivity which contains the
danger of deceiving the decision maker from the real subjectivity of the estimates and
evaluations. Apart from this, there exist a number of different procedures and methods. All in all,
however, they have the following disadvantages:
- The integration of so-called hard and soft decision criteria in a comprehensive model remains to
be solved.
303
Plenary Program
- The processing of information about the conditions of the market and the competition is mostly
not satisfactory.
- The long-term effects of an investment are only considered in isolation.
- There is no support provided when information of long-term effects about an investment is
needed.
Intuition as an Element of Real Decision Processes
Since there do not exist any suitable methods for the evaluation of investments, decisions are
often made on the basis of intuition and of previous experiences. Although the weaknesses of this
kind of evaluation of investments are evident, real life often requires decisions on this basis
because there are no better methods available.
Intuitive knowledge about complex issues is stored in various ways in so-called mental
models by the decision makers. Senge (1990) notes three categories of knowledge as shown in
Figure 2:
- Knowledge of facts is noted first when dealing with complex systems. It comprises knowledge
about the state of the system, perhaps about skills for mastering problems in a reactive way.
This encompasses e.g. knowledge about financial aspects such as profits, costs, revenues, etc.
- Behavioral knowledge about systems is more far-reaching. It deals with statements about trends
and developments in the past which may possibly continue into the future in terms of trend
extrapolations, e.g. the development of demands, business cycles, etc. The consideration of
such knowledge permits a certain future-oriented or proactive decision-making.
- The development can only be influenced generatively or organized if structural knowledge is
available. This is knowledge about feedback loops, policies, delays, and gains.
ae
a Fact:
(reactive)
Figure 2: Knowledge about Complex Systems
The problem is that the decision maker has stored these types of knowledge mentally in a
relatively unstructured way. Furthermore, the difficulties of a purely intuition-based approach
increase the more in detail this pyramid is regarded. In real life intuition-based decisions are
seldom founded on correct and verified structural knowledge.
Especially in the case of investments in flexible assembly systems one deals with complex
socio-economic systems. And complex systems possess two basic characteristics which make
them extremely difficult to handle, and which present themselves as a basic evil for the decision
maker: Complex systems contain feedbacks and delays (Sterman 1989).
304
System Dynamics '95 — Volume I
This severely inhibits the formation of structural knowledge through intuition. By now there
exist a number of scientific studies which prove that human decision makers tend to commit the
same errors over and over again when dealing with complex systems without the support of any
instruments (e.g. Dérner 1983, 1989; Vester 1985; Sterman 1989, 1994). E.g. delays in socio-
economic systems are often underestimated. One simple example are delays in the process of
training people. New employees become productive only after a certain period of time. Until then
they may even inhibit the productivity of the present team of employees. The employment of
new people can thus - contrary to the intended effect - reduce the output of the company for a
certain time. Only in the long run do the intended positive effects become manifest. The time gap
between cause and intended effect often makes it difficult to bring them together and connect
cause an effect correctly. An exemplary collection of such basic errors in human's reasoning with
regard, to complex systems is shown in Figure 3.
o* Underestimation o Missing e Isolated
of Delays Assignment of Targets Problem Analysis
* Ignorance of “
Limits of Growth Tendency to
Dictator Behavoiur
e Erosion iv)
of Objectives
@* irreversible
se sa i Formulation of
e Shifting Focus of Interests
the Problem
© Attraction of — © Wrong Estimation
Short-Term Success t { of Exponential Growth
ik ay
@* Tendency to © Generalization Ignorance of
Oversteer of Single Events Side Effects
Figure 3: The Manager as Prisoner of his/her Own Errors of Reasoning.
Nevertheless, the mental models resp. the experience of experts and decision makers contain a
large amount of valuable knowledge that should be considered when a decision is made. After
all, this it is already done in real life. However, there it takes place mostly in an unstructured way
and is purely based on intuition, under the hazard of committing the errors described. Up to now,
unfortunately there aren't any methods or procedures available, or not even ideas concerned with
these, which could provide substantial support for the solution of this problem.
Figure 4 counterposes once more the two diametrically opposed methods which are used in
real life. The evaluation of investments on the basis of hard decision criteria, such as e.g. costs,
rentability, etc., and the application of conventional investment procedures are not sufficient if
one desires an overall evaluation. These procedures can be useful in evaluating the cost- and
payment-aspects, but the benefits and profits of an investment can hardly be analyzed properly.
The inconsistency in the consideration of flexibility costs and flexibility benefits easily leads to a
"killing calculation" of necessary investments.
305
Plenary Program
Evaluation of Investments
va in Flexible Assembly Systems. =]
Hard Decision Criteria: Soft Decision Criteria:
- Costs
~ Profit
- Cash Flow
- Break Even Time
- Knowledge from Experience
- Competitor Behaviour
(Pricing, Marketing, etc.)
- Qualitative Aspects
(Flexibility, Quality, etc.)
Danger: i Danger: UY
Killing Necessary Investments Wrong Evaluation of Benefit Aspects
Possible Solutio!
Structured Modelling Process to
- Quantify Soft Decision Criteria in a Step-by-Step Process,
~- Integrate Hard and Soft Decision Criteria,
~ Simulate Alternative Investment Decisions and
~ Initiate and Support Organizational Learning Processes.
Figure 4: Evaluation of Investments in Flexible Assembly Systems
This is contrasted with the sole use of soft decision criteria. Here a suitable evaluation of an
investment is not possible either. The estimation based on intuition of the effects of qualitative
benefit factors, as e.g. improvements in flexibility and quality, easily results in their
overestimation. Because of the dynamic long-term effects only seldom decision makers can build
up knowledge from experience that can immediately be transferred to future projects.
Because of the differing foci of analysis the simultaneous application of hard and soft
decision criteria in separate studies may lead to conflicting results. Often the investment does not
seem desirable when considering the hard criteria. The application of the soft criteria, however,
leads to the insight that the investment is desirable, perhaps even absolutely essential. Therefore
it appears useful and important to attempt a better integration of the two types of decision criteria
in a more comprehensive procedure. The intention is to achieve the following:
- to quantify soft decision criteria step by step,
- to integrate hard and soft decision criteria,
- to achieve a dynamic simulation of the interconnected effects of investments in flexible
assembly systems, and
- to initiate and support collective learning processes with regard to understanding the effects of
such investment programs.
306
System Dynamics '95 — Volume I
Integrating Hard and Soft Decision Criteria Step by Step
Figure 5 shows a step by step modelling concept for evaluating investment decisions (Zahn and
Greschner 1993). It aims to make a simulation model available in which the investment project is
shown in its economic network, This should make the evaluation of the effects of an investment
possible in relation to time. This simulation model can eventually be converted into a learning
model that is designed to distribute the knowledge generated in the modelling process among all
the employees of a company. The basic idea behind this is that it is not the scenario of effects
that is produced by means of such prognostic models but it is rather the process of modelling
itself which permits essential insights (Vennix and Scheper 1990).
Modelling Front End
Focus Modelling Stages Tools
Creativity Creativity Tools Brainstorming,
Fishbone Diagram, etc.
System Thinking Structuring of Generic Models, Causal Diagrams,
Problem Area Object Orientated Diagrams,
Hypertext Systems
Knowledge of Qualitative Model Qualitative Presimulation
System Structure
Knowledge of Quantitative Model ics Model
System Behaviour System Dynamics Made
Rees! | tearing Model Interactive Computer Game
Gaming and Simulation Front End
Figure 5: Step by Step Modelling Concept
The generation of the simulation model is realized in computer-aided sessions. A previously
formed group of decision makers is asked to participate actively in the process of modelling. This
so-called modelling group gradually uncovers the relationships of cause and effect that underlie
the investment project in a structured way. In this process hard, quantifiable decision criteria are
applied as well as soft criteria that are difficult to quantify. The aim is to uncover and describe
the problem while taking into account the intuitive experiences of the experts and decision
makers.
In a first step the modelling group can fall back on a general or generic model. This generic
model is based on a large number of empirical studies and theoretical insights. In the generic
model the basic structures of the field under investigation are shown - in our case investments in
flexible assembly systems. The generic model presents the basic cause - effect relationships in
the form of diagrams and of a hypertext system in which knowledge about the model is stored.
This knowledge can then be made available to all users.
307
Plenary Program
When structuring and delimiting the problem the concrete model is deduced from this generic
level with the help of creativity tools. This step provides a graphic presentation of the specific
cause-effect relationships in the form of causal diagrams and of an expanded hypertext system.
The following step encompasses the qualitative specification of the model - that is the description
of the cause-effect relationships as shown in Figure 6.
influences
Investment Capacity
Direction: positive
Force: equal
Delay: 4 - 8 weeks
Constraint: If sufficient capacity available
Figure 6: Qualitative Specification.
This qualitative specification permits a preliminary qualitative simulation which contributes to a
primary examination of the model structure and eventually leads to the later dynamic simulation.
On the basis of this qualitative view the behavior of the model of the later dynamic simulation
should become more easily imaginable and understandable for the users.
It has not been possible so far to analyse the dynamic phenomena of the investment because
of the lack of a simulation that takes into account the time factor. Nevertheless, this structured
discussion of a problem results in qualitatively better decisions than an analysis that is purely
based on intuition.
All qualitative causal chains between two elements, i.e. all possibilities for inter-
connectedness in the interrelated system between these two elements, can be generated and
examined. Problems can be discussed, as e.g.:
“What effects do investments have on market share ?"
“Which causal chains exist between these two elements?"
The next step on the way from soft to hard decision criteria comprises a quantitative specification
of the model in the form of mathematical equations. Thus a System Dynamics simulation model
is created that can then be employed to examine the dynamic phenomena of the investment
project. On the basis of the information generated through simulation the available investment
alternatives can be evaluated. Finally, this simulation model can be transferred into a learning
model with an appropriate user interface. With the help of this learning model the knowledge that
the modelling group has acquired in the process of modelling can be distributed in the company
as a whole. The aim is to initiate and support processes of organizational learning.
A Learning Model to Initiate Organizational Learning
By means of System Dynamics simulation studies a large amount of important and relevant
knowledge is created. However, frequently only the modelling group has this knowledge at its
disposal, and therefore only this group among all managers of a firm can make use of this
308
System Dynamics '95 — Volume I
knowledge. In the case of complex investment projects other decision makers need to be
convinced that organizational measures based on this knowledge are necessary. Decisive
premises for parallel and uniform action toward the same goal by all managers involved in an
investment project are an understanding of the necessity of the changes which this investment
will cause, and of the formation of common expectations with regard to the economic causalities
affected.
In workshops the learning model is used as an instrument to facilitate the development of
such common ideas and insights in a collective process for the people that are involved in the
decision about an investment - that is to construct a common mental model. For those involved
the learning model provides the opportunity to work out essential insights on their own (e.g. Kim
and Senge 1994).
Figure 7 shows the user interface of a learning model. The users have the option to make
various decisions. They can
- realize investments,
- fix prices,
- give orders for overtime work,
- order material, and
~ hire or fire people.
Datei Modelle Spiel Info
aE
=
|Bestellungen
Beschaffung
Personal Finanzen
Preis 7 Quartal Kennzahien
Uberstunden Kosten |Nettokapazitaten
Markt Personal
se
= = =
Liquide Mittel
DM
00000.00- a
& Posteingang
Zuwenig Personal in Woche 25
Zuwenig Personal in Woche 26 a
zuwenig Personal in Woche 27 wey
Zuwenig Personal in Woche 28 ie 4
Zuwenig Personal in Woche 29 Le
Station 1 ist uberlastet in Woche 34 a
to
oN
ae an
~190000.00—+ : : :
— selbstfinanziertes Investitionsbudget
Wochen
20.00
— — Investitionsbudget
40.00
Figure 7: User
See
Interface of a Learning Model
309
Plenary Program
Once the users have made their decision they start the simulation of 12 weeks = 1 quarter by a
mouse-click on the button "continue". After the program has run through the users can read the
information about the consequences of their decisions in reports, diagrams and charts. These
include data about business aspects such as revenue, profit, cash flow, etc., but also information
about capacity, flows of material, personal resources, etc. The diagram as shown in Figure 7
shows the curve of liquidity during the first quarter simulated. In real life the liquidity is a
decisive aspect with regard to the success of an operation. In the model this is the case as well.
Does the curve move below zero the company is insolvent. This virtual bankrupcy can also be
caused if something goes wrong with the second existential aspect, that is the equity of the
company. If there do not exist any more equity this will also cause a premature end of the game.
Furthermore, we have installed a "mailbox" which notifies the state of critical aspects of the
assembly system. In this case the lack of manpower and the overload of station 1 is pointed out.
As has been said before, working with the learning model aims at users who intend to acquire
the knowledge of the modelling group for themselves in an interactive way. For this reason the
program provides them with access to all the tools that have briefly been presented here, that is
the causal diagrams as well as the hypertext system, and the qualitative methods of analysis.
References
Blohm, H./ Liider, K. 1991. Investition, 7. iiberarb. u. erw. Aufl., Miinchen.
Dorner, D. 1983. Lohhausen, Reinbek.
Dorner, D. 1989. Die Logik des MiBlingens, Reinbek.
Horvath, P./ Kleiner, F./ Mayer, R. 1987. Dynamische Investitionsrechnung fir flexibel
automatisierte Werkzeugmaschinen. In: Die Betriebswirtschaft, 47(1):69ff.
Kim, D./ Senge, P. 1994. Putting Systems Thinking into Practice. In: System Dynamics Review,
10(Summer - Fall):277-290.
Liider, K. 1993. Investitionsplanung und -kontrolle. In: Wittmann, W.; Kern, W.; Kohler, R. u.a.
(Eds.), Handworterbuch der Betriebswirtschaft, 5. Aufl., 3 Bd., Stuttgart: 1982-1999.
Senge, P.M. 1990. The Fifth Discipline - The Art and Practice of the Learning Organization,
New York.
Sterman, J.D. 1989. Modeling Managerial Behavior: Misperceptions of Feedback in a Dynamic
Decision Making Experiment. In: Management Science, 35 (3):321-339.
Sterman, J.D. 1994. Learning in and about Complex Systems. In: System Dynamics Review.
10(Summer - Fall):291-330.
Vennix, J.A.M./ Scheeper, W.J. 1990. Modeling as Organizational Learning: An Empirical
Perspective. In: Proceedings of the 1990 International Conference of the System Dynamics
Society, Boston:1199 - 1210.
Vester, F. 1985. Neuland des Denkens, Stuttgart.
Wildemann, H. 1986. Strategische Investitionsplanung fiir neue Technologien in der Produktion.
In: Albach, H.; Wildemann, H. (Eds.). Strategische Investitionsplanung fiir neue Techno-
logien, Z{B-Erganzungsheft 1:1ff.
Zahn, E./Greschner, J. 1993. Model-Based Planning for Strategic Management: An Integrated
Simulation and Learning Toolkit. In: Zepeda, E./Machuca, J.A.D. (Eds.): The Role of
Strategic Modelling in’ International Competitivness. Proceedings of the International
System Dynamics Society, Cancun, Mexico:621-629.
310
