DECISION SUPPORT FOR INNOVATION MANAGEMENT:
APPLICATION TO THE LIGHTING
INDUSTRY
Heinz-Dieter Haustein and Mathias Weber
I n t e r n a t i o n a l I n s t i t u t e f o r A p p l i e d S y s t e m s A n a l y s i s , L a x e n b u r g , A u s t r i a
RR-83-29 December 1983
INTERNATIONAL INSTlTUTE FOR
APPLIED
SYSIXMS ANALYSIS Laxenburg, AustriaInternational Standard Book Number 3-7045-0050-5
Research Reports, which record research conducted a t IIASA, a r e independently reviewed before publication. However, t h e views and opinions they express a r e not necessarily those of t h e I n s t i t u t e o r t h e National Member Organizations t h a t support i t .
Copyright O 1983
International I n s t i t u t e for Applied Systems Analysis
All r i g h t s reserved. No p a r t of t h i s publication may be reproduced o r t r a n s m i t t e d in any form o r by a n y means, electronic o r mechanical, including photocopy, recording, o r any informa- tion s t o r a g e o r retrieval system, without permission in writing from t h e publisher.
Cover design by Anka James
P r i n t e d by Novographic, Vienna. Austria
In today's t u r b u l e n t economic environment, every decision affecting t h e development of i n d u s t r y necessarily carries a n increased risk t h a t t h e antici- pated economic and social goals will n o t be achieved. The description of deci- sion making does not always include t h e notion of risk. S o m e t i m e s t h e "vola- tility of c o s t factors" or changing economies of scale (innovation being t h e p r i m a r y reason for t h e change) are held responsible for u n c e r t a i n t y a b o u t f u t u r e development. These phenomena a r e also used t o explain t h e decline in capital f o r m a t i o n a n d i n decisions to invest t h a t we a r e c u r r e n t l y witnessing.
The economic a n d decision sciences a r e trying t o cope with t h i s s i t u a - tion by devising m o r e sophisticated m e t h o d s a n d procedures for supporting decision making. This Research Report reviews s o m e m e t h o d s t h a t a r e appli- cable t o t h e analysis of innovation p a t t e r n s , with t h e a i m of basing t h e neces- sary decisions on m o r e sound reasoning. The r e p o r t t h e n describes t h e appli- cation of s o m e of t h e s e methods t o innovation m a n a g e m e n t i n t h e lighting i n d u s t r y . I t is hoped t h a t t h e i r application will r e s u l t i n b e t t e r decisions being m a d e i n t h e allocation of resources for innovation.
Tibor Vasko
Lkputy Chairman
of t h e f o r m e r Management and Technology Area
CONTENTS
SUMMARY 1 INTRODUCTION
2 A REVIEW OF DECISION SUPPORT MODELS AND THEIR RELEVANCE TO INNOVATION MANAGEMENT
2.1 Models for Evaluation a n d Selection of Innovation Projects 2.2 The Decision Tree as a Basis for t h e Proposed Model 2.3 Comparison of Models for Multiobjective Decision Making 2.4 Risk Evaluation of Portfolios of Innovation Projects 3 THE LIGHTING INDUSTRY: A CLASSIC EXAMPLE OF INNOVATION
3.1 Developments in Lighting 3.2 P r o d u c t and Process Innovations
3.3 Classification of Innovations i n t h e Lighting Industry 3.4 Lighting Application Systems
4 A DECISION SUPPORT SYSTEM FOR THE LIGHTING INDUSTRY 4.1 The Basic Approach
4.2 Interactive Mode of Operation
4.3 The Basic Model and Different Versions 4.4 Quantifying Risk and Multiple Objectives
4.5 Computer P r o g r a m s for Two Versions of t h e Model 4.6 Results
REFERENCES
Research Report RR-83-29. December 1983
DECISION SUPPORT FOR INNOVATION MANAGEMENT:
APPLICATION TO THE LIGHTING INDUSTRY
Heinz-Dieter Haustein and Mathias Weber
h t e r n a t i o n a l I n s t i t u t e f o r Applied S y s t e m s A n a l y s i s , L a z e n b u r g , A u s t r i a
M a k i n g d e c i s i o n s a b o u t r e s o u r c e a l l o c a t i o n f o r f u t u r e i n n o v a t i o n s is a c h a l l e n g i n g t a s k in b o t h p l a n n e d a n d m a r k e t e c o n o m i e s . Not o n l y c a n s u c h d e c i s i o n s n o t be r e v e r s e d w i t h o u t c o n s i d e r a b l e l o s s o f e f f i c i e n c y , b u t t h e d e c i - s i o n m a k e r g e n e r a l l y f a c e s a n u m b e r o f c o n f l i c t i n g o b j e c t i v e s . In this r e p o r t t h e a u t h o r s t r y t o c o m b i n e t w o d i s c i p l i n e s that h a v e b e e n e v o l v i n g i n d e p e n - d e n t l y f o r a l o n g t i m e : i n n o v a t i o n t h e o r y a n d d e c i s i o n t h e o r y . A d e c i s i o n s u p p o r t s y s t e m f o r m a n a g i n g i n n o v a t i o n s s h o u l d r e f l e c t t h e m u l t i s t a g e n a t u r e o f t h e i n n o v a t i o n p r o c e s s a n d s h o u l d a l s o b e s u i t e d t o m u l t i o b j e c t i v e d e c i s i o n m a k i n g . At t h e s a m e t i m e i t is n e c e s s a r y t o s i m p l i f y t h e r e a l s i t u a - t i o n f o r t h e d e c i s i o n m a k e r in o r d e r t o a p p l y f o r m a l p r o c e d u r e s .
A v e r y p r o m i s i n g s c h e m e is t h e d e c i s i o n t r e e , t h o u g h i t h a s s h o r t c o m - i n g s . A p p l i c a t i o n o f d e c i s i o n t r e e s is c l o s e l y c o n n e c t e d with t h e e v a l u a t i o n p r o c e s s . A l m o s t a l l m o d e l s f o r e v a l u a t i n g i n n o v a t i o n p r o j e c t s o p e r a t e with o n l y o n e o b j e c t i v e . H o w e v e r , d i s c u s s i o n s with d e c i s i o n m a k e r s in t h e l i g h t - i n g i n d u s t r y , w h i c h s h o w s c l a s s i c f e a t u r e s o f t h e i n n o v a t i o n p r o c e s s , r e v e d e d t h e n e c e s s i t y t o i n c l u d e a t l e a s t t h r e e o b j e c t i v e s in t h e e v a l u a t i o n .
% r e f o r e , t h e a u t h o r s h a v e m a d e u s e o f t h e p o s s i b i l i t i e s o f m u l t i o b j e c t i v e d e c i s i o n m a k i n g .
The d e c i s i o n p r o b l e m in this w o r k c o n c e r n s t h e a l l o c a t i o n of r e s o u r c e s t o i n n o v a t i o n p r o j e c t s f o r t h e 1981-85 five-Year P l a n in t h e G e r m a n D e m o - c r a t i c R e p u b l i c . At p r e s e n t t h e m o d e l f o r e v a l u a t i n g i n n o v a t i o n p r o j e c t s i s b a s e d u p o n l i n e a r p r o g r a m m i n g a n d d e c i s i o n t r e e s . It w i l l be i m p r o v e d in c l o s e c o l l a b o r a t i o n with d e c i s i o n m a k e r s , u s i n g t h e r e s u l t s of g o d p r o g r a m - m i n g a n d o t h e r a s p e c t s o f d e c i s i o n t h e o r y .
1 INTRODUCTION
At one t i m e the decisions t h a t a firm m a d e or, r e s e a r c h and develop- m e n t , i n v e s t m e n t , production, a n d marketing were relatively independent of one a n o t h e r . Nowadays, however, i t is clear t h a t every decision m u s t t a k e i n t o a c c o u n t t h e whole process we call innovation. In addition, t h e changing a n d often t u r b u l e n t conditions of t h e world a n d national m a r k e t s have intro- duced m o r e risk i n t o decisions t o reallocate r e s o u r c e s among various innova- tive projects.
In t h i s study we review t h e main approaches of decision theory to t h e evaluation and selection of projects and link t h e m with innovation manage- m e n t . This is t h e first s t e p of a r e s e a r c h program t h a t is being c a r r i e d o u t a t t h e Economic University of Berlin i n t h e Germari Democratic Republic. The study was promoted by decision m a k e r s in t h e lighting i n d u s t r y of t h e GDR.
An analysis of t h e decision-making process in t h i s i n d u s t r y revealed t h e n e e d for a decision s u p p o r t s y s t e m . Our approach is t h u s tailored t o t h e needs of t h i s industry. Our u l t i m a t e goal is t h e development a n d i m p l e m e n t a t i o n of a decision s u p p o r t s y s t e m suitable for making decisions about innovation proj- e c t s a t t h e level of R&D m a n a g e m e n t using a portfolio a p p r o a c h .
Since innovations a r e closely linked t o national a n d i n t e r n a t i o n a l m a r - kets a n d r e s o u r c e s , t h e i n t e r a c t i o n between g o v e r n m e n t a l innovation policy and t h e technological policy of t h e individual firm is i m p o r t a n t . Although quality and consistency of corporate strategy a r e e s s e n t i a l t o t h e s u c c e s s of innovation, in practice c o r p o r a t e s t r a t e g y does n o t provide complete i n s u r a n c e , because inriovation is a complex p h e n o m e n o n , touching all s p h e r e s of technological, economic, and social activity. We c a n n o t hope t o i r ~ c o r p o r a t e all of t h e s e i n t e r r e l a t e d activities i n t o one quantitative decision support model. Moreover, i t is questionable w h e t h e r s u c h a n elaborate model would really assist t h e decision m a k e r in arriving a t b e t t e r decisions. Iri o u r view i t would b e better to include c e r t a i n c r u c i a l qualitative aspects, in t h e f o r m of judgments concerning expected f u t u r e s t a t e s of t h e world.
The m a n y factors t h a t influence t h e development of innovation c a n gen- erally be a t t r i b u t e d to t h e innovator,the organization, a n d t h e e n v i r o n m e n t . While no list of factors c a n be exhaustive, a brief survey, p r e s e n t e d below, will help t o indicate t h e advantages a n d shortcomings of t h e models proposed in t h e l i t e r a t u r e for aiding decision making on innovation projects, including o u r own a p p ~ . o a c h (Haustein e t aL. 1981).
I. Innovator
a . I n p u t , o u t p u t
a l . Input-related factors: necessary quantities a n d qualities for production
a2. Output-related factors: knowledge a n d utilization of properties a n d possible applications of t e c h n i q u e
b. Interaction of innovators
b l . Interplay of functional roles t h a t a r e n e c e s s a r y t o accomplish innovative activities
b2. Characteristics of innovators in t h e s e roles
11. Organization c . Resources
c 1. Material resources c2. Human resources c3. Information c4. Capacity
c5. Innovative potential d. Organizational dimensions
d l . Relationships with t h e e n v i r o n m e n t d2. Internal dimensions
e . Organizational m e a s u r e s e l . Planning
e2. Control 111. Environment
f . Resources
f 1. Natural resources f2. Human resources
g. Competitive situation: t i m e factor h . National needs and goals
i. Demand
Each factor influences innovative activities in a specific way, depending on t h e c i r c u m s t a n c e s ; no general p a t t e r n of influence can be found. The degree of influence also changes over t i m e , depending on t h e s t a g e of t h e par- ticular innovation. The concept of t h e efficiency of a factor, i.e. its degree of influence, is derived from a m i x t u r e of evidence from empirical studies and results of theoretical reasoning. Hypotheses about t h e efficiency of various factors a r e p r e s e n t e d by Haustein e t al. (1981).
Looking a t t h e innovation process within the whole social and n a t u r a l environment, t h e decision m a k e r m u s t identify socioeconomic opportunities by comparing needs, resources, a n d t h e s t a t e of t h e a r t in all processing sys- t e m s (Figure 1). We can investigate processing s y s t e m s using d a t a on, for example, material a n d energy flows o r bottlenecks in t h e replacement of t h e labor force with m o d e r n technology. We m a d e s u c h a n analysis of t h e innova- tion cycle in t h e textile industry of t h e German Democratic Republic (Hau- s t e i n 1974). This involved t h e following steps:
Draw up a s c h e m e of t h e technological s t r u c t u r e . Draw a n energy flow diagram.
Construct ii m a t e r i a l flow diagram.
Determine t h e a m o u n t of capital equipment p e r person o r t h e de- g r e e of automation in all e l e m e n t s of t h e technological s t r u c t u r e . Determine t h e range of potential innovations a n d t h e i r ability t o overcome loopholes and bottlenecks in t h e s y s t e m .
Evaluate potential innovations.
Estimate what innovations a r e lacking.
Processing systems Socioeconomic
FIGURE 1 Socioeconomic opportunities arising from t h e relationships between needs, resources, and processing systems.
Identify innovations of g r e a t importance, i.e. t h a t a r e able t o c r e a t e new imbalances.
Recommend a technological policy for t h e whole system.
Table 1 shows a possible s c h e m e for performing s u c h a n analysis.
From t h e standpoint of t h e organization, the innovation process can be divided i n t o t h r e e steps:
1. Establish and develop innovative potential.
2. Realize this potential by initiating and implementing innovations.
3. Ensure t h a t conditions within t h e organization will allow t h e new products and processes t o have a g r e a t e r impact on growth a n d efficiency of t h e organization.
The potential efficacy of a n innovation for a f i r m can be established only by taking s t r a t e g i c measures, for two reasons: a creative potential cannot be realized by s h o r t - t e r m activities; a n d , in m o s t cases, gaps a n d bottlenecks in technology t r a n s f e r cannot be overcome in one year. However, we do n o t i n t e n d t o deal h e r e with strategic problems of industrial organizations. Our aim is t o analyze t h e innovation process, which consists of t h e following steps:
preparation
r e s e a r c h and development i n v e s t m e n t
production
m a r k e t penetration phaseout.
TABLE 1 Scheme for analyzing socioeconomic opportunities according t o resource- processing systems.
I
Main obstacle t o efficiencyI I I l I I I l l
Energy (%) Material (%)
Labor time (man-hours)
1
Energy (Tcal)Weight of material (kt)
1
Main ~ r o c e s s e sI
A1
BI
I
Phase of r e s o u r c e - ~ r o c e s s i n e cvcle1
1I
11 1(We can also distinguish three more general stages: invention, technical real- ization, and commercialization.) The steps are interconnected and often overlap. Only t h r e e of t h e m (preparation, production, and phaseout) are inevitable. The time aspect of t h e innovation process is generally well known, b u t from t h e point of view of management i t is necessary also to analyze t h e whole process in t e r m s of t h e requirements and opportunities for decision making. Therefore, we have combined two dimensions in Table 2: t h e steps of t h e innovation process and t h e stages of t h e decision process. In this way we can identify where t h e main delays occur. In t h e example shown, t h e whole
innovation cycle from event l a t o event 9f lasts m o r e t h a n 19 years. The delay in m a s s production caused by foreign competitors is relatively high (nine years). The reason for this is t h e retardation in t h e first two or t h r e e s t e p s of t h e innovation process a n d in t h e preparation of decisions. However, t i m e is only one e l e m e n t of t h e innovation process.
TABLE 2 Congruence between the innovation and decision processes in the textile in- dustry of the German Democratic Republic.
Innovation Decision process process
a. b. c. d. e. f . Total
Appear- First Prepara- Decision Start of End of period ance of external Lion of making irnplemen- irnplernen- (yr) problem information decisions tation tation
1. Preparation 1959
2. Research 1962 3. Development 1963 4. Investment 1963
preparation 5. Investment 1964
realization 6. Start of 1967
production 7. Mass 1969
production 8. Market 1973
penetration 9. Phaseout
-
Theoretically, i t is not difficult to include i.nput- a n d output-related fac- t o r s s u c h a s labor, capital equipment, raw m a t e r i a l s , technological risk, u n i t scale, a n d funding. Certain relations with t h e business e n v i r o n m e n t can be modeled fairly accurately, b u t m a n y o t h e r factors have r e m a i n e d outside t h e project evaluation and selection models reported in t h e l i t e r a t u r e , s u c h a s interplay of functional r u l e s , c h a r a c t e r i s t i c s of innovative persons, t h e econo-mic m e c h a n i s m , and t h e m a n a g e m e n t s y s t e m . We regard s u c h shortcomings t o be theoretically r a t h e r t h a n practically i m p o r t a n t : a deci- sion m a k e r in a particular firm is probably not very c o n c e r n e d a b o u t m o s t of t h e s e factors; in his daily work h e c o n c e n t r a t e s on input- a n d output-related factors.
Some of t h e m o s t i m p o r t a n t relations between t h e early, predictable c h a r a c t e r i s t i c s of an innovation project and o t h e r variables a r e p r e s e n t e d in Figure 2. Using t h e model t h a t we have developed, we will look a t the early s t a g e s of a n innovation project, when only r o u g h predictions of t h e scientific/ technological level a n d range of application of t h e innovation exist.
The scientific/technological level a n d t h e r a n g e of application d e t e r m i n e t h e n e x t s e t of variables describing t h e innovation project:
- - - - - -
-- -
- 1I I I I I I I I I I I
Expected economic I
I
I I
L - - - - -
-
--
- - - - - -- -
- - - - - - -I Social needsFIGURE 2 The decision-making process for innovation projects. The features and re- lationships within the broken perimeter are included in the decision support system t h a t has been developed. Open circles indicate decisions.
Business
complexity
compatibility with existing equipment risk
expected R&D time expected lifetime
expected resource requirements.
I
environmentEstimates of these variables, and of expected economic benefits and expendi- t u r e s , become more and more accurate as the project progresses. One h a s also t o take into account the efficiency of t h e firm producing the innovation, because the speed with which a new product or process is adopted depends
greatly on the benefits t o the consumer. This is termed t h e socioeconomic effectiveness of the innovation.
We shall now summarize t h e features of innovation decisions t h a t should be taken into account when devising a decision support system for innovation management.
Decisions on innovation can only be reversed with considerable losses of efficiency. The further an innovation advances, t h e more difficult i t becomes to reverse t h e decision to adopt i t , because of t h e manpower involved.
Innovation decisions are affected by problems in all economic activities of t h e firm, e.g. in investment policy, t h e hiring of man- power, procurement policy, and m a r k e t strategy (Hennecke 1975).
Great uncertainty about further development of adopted projects, f u t u r e market conditions, e t c . complicates decision making. Even in planned economies, resource allocation cannot be predicted exactly.
Decision makers have t o deal with many conflicting objectives representing both qualitative a n d quantitative aspects of business.
Measurement and comparison of these objectives combine objective and subjective elements. The importance of experience in t h e s e m a t t e r s cannot be overemphasized. The evaluation of alternatives can change rapidly as a result of unforeseen events.
Innovations a r e created not by chemical reactions, but by people (decision makers, research and development specialists, workers), who form groups with often conflicting goals. To be successful, management m u s t create a n atmosphere of c o m m i t m e n t t o t h e projects eventually selected and weigh t h e i n t e r e s t s of all groups.
An innovation project in the lighting industry normally lasts for t h r e e t o seven years and consists of many steps, although t h e methodology described in this article does not consider explicitly steps preceding project proposals or following implementation.
Hence making decisions on innovation is by n a t u r e dynamic and multistage. Every stage has particular problems and sources of uncertainty. Therefore, a lot of partial decisions have to be made in t h e iterative process of decision making during a project.
Decisioris have t o be made within a c e r t a i n period, sometimes r a t h e r brief. Thus decisions are made sequentially; task specifications rnay change over t i m e , e i t h e r independently or a s a result of previous decisions; information available for l a t e r deci- sions may be contingent upon t h e outcomes of earlier decisions;
and implications of any decision may affect t h e future of the project (Rapoport 1975).
In a planned economy, innovation decisions depend on consulta- tions with higher levels of administration. The more important t h e innovation, t h e more time is needed for consultations.
In t h i s study we cannot consider all levels of innovation management (Twiss 1974, ch. 2 ) . There are so many peculiarities among different levels of
management t h a t no general recommendations can be given. The higher t h e level in t h e m a n a g e m e n t hierarchy, the more complex t h e decisions become.
This is reflected in t h e number of admissible alternative decisions, t h e number and quantifiability of t h e objectives, t h e complexity of t h e inter- dependences among t h e objectives, and t h e scope of long-lasting effects (which is difficult to predict). In addition, a t higher levels t h e problems fac- ing m a n a g e m e n t become less s t r u c t u r e d . This considerably affects the appli- cability of economic-mathematical methods to t h e management of innova- tion projects.
2 A REVIEW OF DECISION SUPPORT MODELS AND THEIR RELEVANCE TO INNOVATION MANAGEMENT
Keen and Morton (1978) define a decision support system (DSS) as computer-based support for management decision makers who are dealing with s e m i s t r u c t u r e d problems. The problem of designing an optimal R&D portfolio is often considered unstructured, but t h i s depends in e a c h case on t h e features of t h e innovation decisions, which are determined by:
t h e complexity of t h e technical field (number a n d n a t u r e of t h e relations t o other scientific disciplines, technical fields, and indus- trial branches);
t h e age a n d m a t u r i t y of t h e most important product and process innovations, which determine t h e profile of t h e technical field under consideration and t h e dynamics of i t s d.evelopment ( t h e r e is a n excellent study by Filippovskii 1978);
t h e class of innovations prevailing;
t h e position and importance of t h e technical field (or industrial branch) in t h e economy a s a whole (Haustein and Maier 1980).
Decisions about innovation projects will also display features of uniqueness a n d / o r repetitiveness, with obvious consequences for t h e degree of support t h a t formal models c a n give.
Important operations in t h e decision process are comparing resource requirements and availability and assessing t h e degree to which new projects can m e e t t h e goals of t h e firm. Thus innovation decisions rely on searches for information on previous experience a s well a s t h e application of analytic techniques. Some of t h e steps in decision processes of this type can be partly delegated t o t h e computer for solution by a n interactive mode of operation.
The general approach of DSS s t a r t s with the investigation of t h e key deci- sions to be made and t h e determination of which parts of t h e process a r e s t r u c t u r e d and which p a r t s rely on subjective judgment. The decision m a k e r t h e n tries t o organize st.ructured subproblems for solution by computerized methods on t h e basis of appropriate models. a
We believe t h a t our approach fits well into t h e concept of DSS. A decision analysis of the proposed innovation projects (based on decision trees) m a y serve a s a convenient starting point for f u r t h e r analysis, using other interac- tive procedures t o be discussed l a t e r . The first s t e p of the analysis is like t h e
rational framework of decision making; the second step resembles
"satisficing" a n d is closer to real decision making. We do n o t see DSS as a replacement for widely accepted management tools but a s a n extension of t h e m .
A DSS based only on an outcome-oriented approach is too limited. Like Zeleny (1976), we define a decision as a dynamic process with feedback loops, s e a r c h detours, sequential exploration of preferred and feasible s e t s of alter- natives, information gathering, assessment of t h e s t r u c t u r e s and goals of t h e alternatives, and the addition and exclusion of alternatives. Optimization of s u c h a complex system is only possible with a highly simplified model based on many assumptions. Figure 3 presents a simplified version of the process- oriented approach t o decision making (details a r e given by Zeleny 1976).
2.1 Models f o r Evaluation a n d Selection of Innovation Projects A decision support system for innovation management should
combine outcome-oriented and process-oriented approaches;
reflect the multistage n a t u r e of innovation;
reflect the uncertainty affecting innovation;
reflect the mutual dependence between innovation projects;
take into account t h e main kinds of resources required in a n inno- vation project;
be suitable for multiobjective decision making;
be m o r e or less compatible with existing planning and m a n a g e m e n t systems;
be suitable for man-machine interaction;
be based on easily accessible data;
be based on existing problem-solving techniques t h a t can be easily computerized.
To date no decision support model meeting all these requirements h a s been constructed.
In practice, the evaluation and selection process consists of a t least two steps. The first is a qualitative screening of t h e proposed innovation projects.
Ranking and scoring can help to reject proposals t h a t do n o t m e e t certain m i n i m u m requirements or t h a t a r e inferior to other candidates. In this s t e p one can adopt risky basic research or applied r e s e a r c h projects with highly uncertain economic parameters. A final decision about whether t o continue or reject a proposal is delayed until major uncertainties can be clarified o r disappear. In t h e second step, which is quantitative in n a t u r e , t h e proposed methodology is applied t o support the final decision.
Our approach to decision support is based on decision t r e e s . We a r e con- vinced t h a t t h i s methodology can be used by t h e decision maker to coordi- n a t e corporate strategy and resource allocation to new and ongoing projects if i t is combined with a model for forecasting long-term effects of innovation projects t h a t have been adopted by t h e firm. The approach is based on c e r - tain principles t h a t a r e common in dynamic and complex situations (e.g.
Preliminary set of alternatives
of ideal
Displacement o f ideal Predecision
conflict
Search for additional
t
New alternatives
I 4
,
FIGURE 3 Simple process-oriented model of decision making, based on Zeleny's (1976) theory of the displaced ideal. The ideal i s defined as the alternative, infeasible in general, t h a t provides the highest score with respect to all individual attributes considered.
I
Belyaev 1977). Faced with the complexity of t h e problem, the decision maker and t h e analyst a r e forced to simplify reality. The simplifications affect t h e projects t o be considered, t h e time periods (model horizon and benefit hor- izon), t h e number of objectives and their formal representation, t h e decision maker, and t h e resources required.
Our model applies only t o medium-sized and large projects. A Axed per- centage of the budget is spent on all remaining R&D projects and on highly uncertain basic research, which sometimes cannot be related to particular products and processes o r h a s ill defined economic and technical parameters.
Research and development management is represented in our model by one b
alternatives
Good compromise found Selection o f
alternatives
Additional information
Search for new information, reassessment o f goals and alternatives, return of
closest t o ideal
4
Partial decisions (discard inferior alternatives) discarded
alternatives
4
Displacement of ideal closer t o feasible set of alternatives;
conflict reduced
decision maker, whose preferences are assumed t o be typical of R&D manage- m e n t as a whole. This assumption may be relaxed in the future
Most of t h e variables in the proposed model a r e of a continuous n a t u r e . In order to handle t h e problems, we discretize all continuous variables and functions (for instance, probability distributions) and consider only a limited number of options, in most cases not more than five, including mean and extreme values. This simplification greatly eases the task of assessing t h e probability of future events, because the decision maker is able to perceive significant differences between t h e options. In discretizing time, we have selected periods of half a y e a r .
In a dynamic environment, where objectives, sets of feasible alterna- tives, and preferences are constantly changing, optimization of t h e evalua- tion and selection process over t h e whole planning horizon is almost impossi- ble. Under s u c h circumstances, it is necessary t o make priority decisions.
The optimal solution refers only to t h e first period. Decisions affecting more distant periods will be reconsidered when t h e information on t h e m becomes more reliable. The decision process is divided into stages, similarly to t h e innovation process (Table 2). This is the main idea of the law defining the gen- eral s t r u c t u r e of t h e decision-making process for innovation projects in t h e GDR.
Decisions about innovation projects cannot be made independently, because projects compete for scarce resources, especially for manpower and investments. For this reason, we use a portfolio approach. In order t o find an approach appropriate to t h e problem of decision making in t h e GDR lighting industry, we shall t r y to split t h e problem into classes of decision situation, which will throw light upon possible difficulties in handling i t . Danilov- Danilyan (1980) based his classification upon a description of the alternatives (good or bad) and a description of t h e preferences (good or bad), thereby dis- tinguishing four classes of decision situation. In our case t h e number of feasi- ble alternatives (project proposals) is known, but a description of t h e m in t e r m s of resource requirements, development t i m e , probabilities of f u t u r e events, and short- and long-term effects on the firm and on society can only be r a t h e r sketchy, a t least in t h e early stages of the innovation process.
Obviously, preferences a r e even less clearly defined. Hence our problem belongs to Danilov-llanilyan's class
IV
(bad descriptions of both alternatives and preferences), like almost all problems in socioeconomic decision making.Von Winterfeldt and F'ischer (1975) classify decision situations on t h e basis of three featu-res of the alternatives: the number of attributes, uncer- tainty, and t i m e (Table 3). An optimal portfolio of innovation projects is characterized by t h e presence of all three complicating features. The works of Danilov-Danilyan and of von Winterfeldt and Fischer indicate t h a t appropri- a t e models for our case a r e still. lacking a t present. The only way to apply for- mal methods is to neglect one of t h e features of t h e preference system, for instance t h e t i m e variability
Models for project evaluation and selection have been reviewed else- where (Gear e t al. 1971, Souder 1973a, b, 1978, Clarke 1974, Schwartz 1976) and have been classified by Moore and Baker (1969), Gear e t d. (1971), Souder (1972), and others. Only very few formal models are in use. Successful implementations of project evaluation models have been reported by Souder
TABLE 3 A classification of choice situations and models (from von Winterfeldt and Fischer 1975).
Case The choice alternative is: Model.
Multi- Uncertain Time-
attributed variable
-
1 Yes No No 1. Simple-order
2. Riskless trade-off model 3. Additive conjoint m e a s u r e m e n t
2 Yes Yes No 1. Simple expected utility model
2. Riskless decomposition - expected utility model
3. Multiplicative expected utility model 4. Additive expected utility model 3 Yes
4 Yes
5 No
N 0
Yes No Yes
Yes No model a t present Yes No model a t present
No 1. Simple-order
2. Difference s t r u c t u r e s
KO 1. Expected utility a n d simple expected utility models
2. Minimax and minimax-regret models 3. Portfolio theory
Yes 1. Additive tirne preferences 2. Additive time preferences with
variable discounting r a t e s 3. Additive time preferences with
constant discounting r a t e s Yes Yes 1. Additive time preferences -
expected utility model (constant o r variable discounting rates)
2. Multiplicative time preferences - expected utility model (constant or variable & ~ c o u n t i n g rates)
(1968), Atkinson and Bobis (1969), Bell and Read (1970), Cochran e t al. (1971).
and Grossman and Gupta (1974). Baker and Pound (1964), Rubenstein (1966).
and Ritchie (1970) have cited t h e following reasons for managers' ignorance of almost all of t h e models proposed:
Important aspects of t h e decision-making process (for instance, uncertainty, t h e sequential n a t u r e of decision making, t h e inter- dependence of projects, and multiple criteria) a r e absent or han- dled inadequately.
The models fail to r e p r e s e n t t h e real evaluation and selection pro- cess, particularly the roles of experience, intuition, and judgment.
The necessary input d a t a a r e lacking.
There is a lack of m u t u a l understanding between decision m a k e r s and analysts.
2.2 The Decision Tree a s a Basis for t h e Proposed Model
Recent developments in modeling t h e evaluation and selection of
R&D
portfolios are encouraging (Hespos and Strassman 1965, Gear e t al. 1970, 1972, Lockett and Freeman 1970, Allen a n d Johnson 1971, Gillespie a n d Gear 1972, Lockett and Gear 1972, Gear and Lockett 1973, c e a r 1974, Chiu and Gear 1979). Clarke (1974) s t a t e d t h a t models involving decision tree analysis have been receiving increasing attention from management scientists. A comprehensive literature survey led us t o conclude t h a t for our specific pur- pose a model using decision t r e e s is most suitable.
A decision t r e e is a convenient tool for s t r u c t u r i n g all of a decision m a k e r ' s ideas about a prolect. With t h e help of a decision tree one can r e p r e s e n t and analyze a series of partial decisions t o be made over time.
Decision t r e e s reflect one of the most i m p o r t a n t features of innovation deci- sions: their sequential c h a r a c t e r .
A formal method based on decision t r e e s can be applied successfully only when t h e innovation project h a s reached a c e r t a i n degree of maturity and ideas about basic construction, project versions, resource requirements, main sources of uncertainty, development time scale, e t c . are relatively well defined. We assume t h a t projects a r e evaluated and selected over a certain planning horizon, which is divided into periods T. A decision m u s t be made on Nprojects, each of t h e m with a n u m b e r of possible paths t o completion.
Projects c a n branch o u t whenever decision nodes or chance nodes occur.
A decision node on t h e time scale i s any point a t which the decision maker can influence t h e progress of the project by making a decision, a s a result of which a branch of a given set of possible paths will be selected. Chance nodes a r e beyond t h e control of t h e decision maker and depend on chance events, such as a n increase in t h e price of raw materials or t h e inability to obtain t h e necessary machinery within a certain t i m e .
The length of the periods in t h e model can be chosen so t h a t a decision is made a t t h e beginning of a period. The same assumption can be made about chance events t h a t are supposed to occur before a partial decision is made. The resource requirements a r e assumed t o be known for each time interval and for each version of t h e project. The n u m b e r of resource types is specific t o each case.
Another model assumption requires t h a t t h e decision m a k e r be able t o assign probabilities to t h e outcomes of a chance node. This problem will be discussed later (Section 2.3.3.1). All combinations of particular decisions and chance events have some result, which is measured according t o scales t h a t correspond t o the chosen multiple objectives.
The presentation of innovation projects in t h e form of decision t r e e s pro- vides t h e decision maker with several advantages:
I t allows him t o s e e all projects as a whole.
It allows the representation a n d adequate handling of interrelated decisions t h a t occur a t different times.
I t omits all less important project features.
It forces t h e decision maker to use notions, judgments, experience, intuition, and quantitative data for constructing decision t r e e s in a n interactive m a n n e r .
It allows early detection of feasible options and bottlenecks.
It shows t h e connections between partial decisions and t h e main sources of uncertainty.
It combines outcome- and process-oriented approaches to decision mak- ing.
Schwartz and Vertinsky (1980) found t h a t the selection of R & D projects is largely dependent on project-specific considerations, s u c h as probability of success (technical and commercial), r a t e of r e t u r n , and payback period.
Broader economic indicators a r e often ignored. "R&D decision making is ...
stimulated by t h e opportunity of particular
R&D
projects r a t h e r t h a n being part of a n integral environmental adaptation strategy." This observation sup- ports our a r g u m e n t for t h e application of t h e decision t r e e to t h e evaluation and selection of innovation projects, because i t provides a b e t t e r representa- tion of project-specific attributes than of environmental ones.However, we c a n n o t overlook t h e several weaknesses and problems i n h e r e n t in this application of decision trees:
a . Decision t r e e s cannot depict accurately t h e complexity of factors influencing t h e real decision-making process. This is t r u e even of quantitative models. Building qualitative factors into t h e decision t r e e is not easy and is often a m a t t e r of subjective judgment. The problem of whether or not i t is possible to apply decision t r e e s t o t h e situation described here is discussed in t h e l i t e r a t u r e . Larichev (1979), for instance, questions t h e value of decision t r e e analysis for unique decisions. On t h e other hand, many applications can be cited for problems of this kind (Keeney and Raiffa 1976, Bell e t al.
1977, Howard 1980).
b. The construction of a decision t r e e is time-consuming. Often deci- sion makers are unwilling t o spend t h e t i m e necessary t o answer analysts' questions about their preference systems, or to provide all of t h e necessary d a t a a t t h e same time.
c. It is particularly difficult to construct decision t r e e s for t h e very cases where t h e i r application would be most useful: i n topics of basic a n d applied research in their early stages. One m u s t be willing to place a c e r t a i n degree of confidence in both t h e objectives and t h e technical/commercial p a r a m e t e r s of t h e projects.
d . Certain methodological problems have to be solved in a specific way for e a c h case. Among t h e m are inclusion of new project proposals in f u t u r e periods, t h e length of t h e planning horizon (the problem of projects t h a t are not completed within t h e planned period), t h e interdependence of projects, transfer of resources, and t h e degree of detail in t h e decision tree.
e . Decision t r e e s do not take into a c c o u n t strategic considerations, which often greatly influence t h e selection of innovation projects. A n u m b e r of i m p o r t a n t aspects of decision making on innovation proj- e c t s a r e not quantifiable. For this reason m a t h e m a t i c a l models may be misleading in s o m e applications (Roman 1980).
f . Decision t r e e s c a n n o t be used t o r e p r e s e n t t h e whole lifetime of a n innovation. It is impossible to specify t h e resource r e q u i r e m e n t s m o r e t h a n five t o seven years in advance. The kinds of r e s o u r c e s required differ considerably from stage t o stage. Hence t h e analyst is forced t o aggregate, thereby losing m u c h of t h e information available. Only very rough figures c a n be calculated for models based on decision trees. However, this is t r u e of all economic-mathematical models intended for supporting innovation decisions.
g. Sometimes decision t r e e s c r e a t e t h e illusion of a freedom of choice, which in reality does not exist because of c o n s t r a i n t s not formally included in the analysis.
h . The basic model is linear (Section 4).
There a r e probably o t h e r limitations t o t h e approach described in t h i s r e p o r t , yet, despite its shortcomings, we a r e convinced t h a t t h e model c a n be useful for case studies other t h a n t h a t of t h e lighting industry, with which our work is concerned.
Not every problem c a n be solved by applying decision t r e e methodology alone. For example, Smallwood a n d Morris (1980) used decision t r e e s only for s t r u c t u r i n g t h e decision; they t h e n used underlying and i n t e r c o n n e c t e d m a t h e m a t i c a l models to g e n e r a t e t h e n u m b e r s . First a t t e m p t s t o realize t h i s approach were reported by Gear e t al. (1970). Other models and techniques, widely accepted in industry, have t o be used t o provide information:
models of innovation diffusion (Mansfield e t al. 1971, Davies 1979);
models for forecasting manpower r e q u i r e m e n t s ;
models of technological substitution (Linstone and Sahal 1976);
models for optimal timing of innovations (Barzel 1960, Kamien a n d Schwartz 1974);
scenario analysis.
Much r e s e a r c h has been carried o u t on how t o facilitate the application of decision t r e e s t o innovation m a n a g e m e n t . This work is aimed a t
developing efficient m e t l ~ o d s for analyzing decision t r e e s (Moskowitz 1971. Marien and Jagetia 1972);
synthesizing several approaches, including decision t r e e s (Chapman 1979);
developing new m e t h o d s for extracting subjective probabilities f r o m t h e decision m a k e r (Yager 1977);
lending a foundation to fuzzy decision analysis (Chang and Pavlidis 1977, Watson e t d. 1979).
On t h e whole t h e s e new efforts mitigate several of t h e disadvantages of deci- sion t r e e s and make t h e t r e e s more useful. However, some recently obtained results do n o t go beyond t h e stage of theoretical investigations or laboratory t e s t s and a r e far from being applicable in business (e.g. Watson e t d. 1979).
Finally, t h e s e developments rely on equipment t h a t - i s not yet widely avail- able, even in large firms (e.g. video projectors) (Levin e t d. 1978).
Chapman (1979) demonstrated t h e flexibility of decision t r e e analysis, combining i t with key characteristics of network approaches. His methodol- ogy "reflects a strong belief in approaches which are interactive, nested, a n d intuitively driven, integrating model selection and solution in a modular fashion, with diagrams and computations emphasizing communication and robustness r a t h e r t h a n precision and generality."
In t h e i r combination of fuzzy-sets theory and decision analysis, Watson e t d. (1979) allow for fuzziness in probabilities and utilities. The a u t h o r s s t r e s s t h e difference between t h e imprecision of t h e i n p u t data and t h e u n c e r t a i n t y of t h e f u t u r e s t a t e of t h e world. These qualities a r e modeled in different ways, using fuzzy-sets theory and probability theory, respectively.
Critics a t t a c k decision analysis for t h e imprecision of t h e data provided by t h e decision m a k e r ("garbage in - garbage out"). This problem cannot be solved simply with a variable-by-variable sensitivity analysis as i t is normally performed, because in reality variables change i n combination with one another. Many decision makers a r e p u t off by t h e necessity to provide infor- m a t i o n in numerical form. Watson e t al. show t h a t this r e q u i r e m e n t can be diminished o r even replaced. I t c a n be expected t h a t in t h e f u t u r e decision m a k e r s will provide their assessments of values, utilities, and probabilities in verbal form. The a u t h o r s point o u t t h a t they cannot offer a n all-purpose tool, b u t t h a t they c a n outline t h e general direction for improving decision analysis.
A t Stanford University in California interactive computer graphics a r e used t o compose, decompose, simplify, transform, merge, and regenerate net- work pictures, including decision trees. The purpose of this system is to accelerate convergence in man-computer experiments, for example by eas- ing t h e t a s k of drawing decision t r e e s for all projects under consideration.
Some of our initial thoughts about t h e s t r u c t u r e of a man-machine system based on decision t r e e s (Section 4) for t h e selection of innovation projects have beer1 corroborated by t h e US study. We plan to use some suggestions i n t h e r e p o r t t o improve o u r system.
Similar efforts were reported by Lewis (1975), Leal and Pearl (1977), a n d Thompson and Kirschner (1978). Lewis's interactive system for editing t r e e s t r u c t u r e s allows insertion, deletion, s e a r c h , and display of any branch of a given s t r u c t u r e . Leal and Pearl described a n interactive c o m p u t e r program t h a t was designed and implemented to elicit decision t r e e s from decision m a k e r s . This automation of t h e tedious process of drawing decision t r e e s in a natural-language conversation between decision maker and c o m p u t e r greatly facilitates t h e distribution of decision analysis techniques.
The t e c h n i q u e of Leal a n d P e a r l d o e s n o t d e p e n d on t h e a r e a of applica- t i o n . All i n p u t d a t a provided by t h e u s e r a r e m a p p e d i n t o o n e of t h e d a t a t y p e s ( e v e n t s , a c t i o n s , likelihoods, r e l a t i o n s , e t c . ) . One of t h e biggest disad- v a n t a g e s of t h e m a n u a l eliciting of decision t r e e s is t h e d a n g e r of s p e n d i n g too m u c h t i m e on d e t a i l s t h a t a r e i r r e l e v a n t t o t h e final s o l u t i o n Leal a n d P e a r l u s e a n efficient t r e e expansion m e t h o d t h a t d i r e c t s effort toward t h e m o s t c r i t i c a l t i p n o d e , defined a s t h e node t h a t is m o s t likely t o c h a n g e t h e first-step s o l u t i o n c u r r e n t l y c o n s i d e r e d b e s t . The t r e e e x p a n s i o n m e t h o d is based o n a s e n s i t i v i t y analysis a l g o r i t h m a n d on t h e analogy between decision t r e e e l l c i t a t i o n a n d h e u r i s t i c s e a r c h i n g on g a m e t r e e s t h a t was first m e n - t i o n e d by Leal a n d P e a r l .
A g e n e r a l i z a t i o n of t h e s e efforts i s r e p o r t e d by Levin e t al. (1978), who developed a s y s t e m f o r i n t e r a c t i v e c o m p u t e r aiding of g r o u p decision m a k i n g based on decision t r e e s . Decision t r e e s a r e c o n s t r u c t e d using v a l u e a n d prob- ability i n p u t s f r o m all g r o u p m e m b e r s . The s y s t e m does n o t a s s u m e famil- i a r i t y of t h e decision m a k e r s with decision a n a l y s i s a n d c o i n p u t e r p r o g r a m - m i n g .
The s y s t e m s being developed a r e becoming i n c r e a s i n g l y user-friendly a n d a r e likely t o realize t h e f o r e c a s t by Matheson a n d Howard (1968) t h a t
"soon t h e logical s t r u c t u r e of a n y decision a n a l y s i s m i g h t be a s s e m b l e d f r o m s t a n d a r d c o m p o n e n t s . " While we c a n n o t overlook t h e d i s c r e p a n c y b e t w e e n t h e inspiring o p p o r t u n i t i e s opened u p by r e s e a r c h e r s a n d t h e a c t u a l applica- tion of t h o s e s y s t e m s in daily decision m a k i n g , t h e g e n e r a l d i r e c t i o n of c o m - p u t e r i z e d decision s u p p o r t s y s t e m s based o n decision analysis s e e m s c l e a r .
2.3 C o m p a r i s o n of Models f o r Multiobjective Decision Making
Almost all m o d e l s for innovation p r o j e c t e v a l u a t i o n a n d s e l e c t i o n o p e r a t e w i t h o n e objective only. However, d i s c u s s i o n s with t h e decision m a k - e r s i n t h e lighting c o m p a n y u s e d for o u r c a s e s t u d y r e v e a l e d t h e n e c e s s i t y t o i n c l u d e a t l e a s t t h r e e objective f u n c t i o n s , which a r e n o t c o m m e n s u r a b l e . We s h a l l d i s c u s s l a t e r which of t h e methodologies for m u l t i o b j e c t i v e decision m a k i n g (MODM) is b e s t s u i t e d for t h e c a s e s t u d y . The e x c e l l e n t reviews by MacCrimmon (1973) a n d Hwang e t al. (1900) will h e l p t o solve o u r p r o b l e m of c h o i c e b e c a u s e t h e y a r e based o n different classification p r i n c i p l e s . MacCrim- m o n s t r e s s e s t h e s t r u c t u r a l differences between t h e v a r i o u s m e t h o d s (Table 4); Hwang e t al. s t r e s s t h e s t a g e a t which t h e i n f o r m a t i o n is n e e d e d a n d t h e type of i n f o r m a t i o n (Table 5); a n d Larichev (1979) c o n c e n t r a t e s o n t h e t y p e of i n f o r m a t i o n provided by t h e decision m a k e r a n d i t s m o d e of u s a g e (Table 6).
A f i r s t g l a n c e a t t h e m o d e l s proposed i n t h e l i t e r a t u r e i n d i c a t e s that. t h e following c l a s s e s a r e worth considering for o u r c a s e s t u d y :
t h e m e t h o d of t h r e s h o l d s of i n c o m p a r a b i l i t y (Roy a n d B e r t i e r 1971, Roy 1977, Larichev 1979);
goal p r o g r a m m i n g (Section 2.3.1);
TABLE 4 Multiobjective/multiattribute decision-making models (MacCrimmon 1973) A. Weighting methods
1. Inferred preferences a. Linear regression b. Analysis of variance c . Quasilinear regression
2. Directly assessed preferences: general aggregation a. Trade-offs
b. Simple additive weighting c. Hierarchical additive weighting d. Quasiadditive weighting
3. Directly assessed preferences: specialized aggregation a. Maximin
b. Maximax
B. Sequential elimination methods
1. Alternative versus standard: comparison of a t t r i b u t e s a. Disjunctive and conjunctive constraints
2. Alternative versus alternative: coniparison of attributes a. Dominance
3. Alternative versus alternative: comparison of a t t r i b u t e s a. Lexicography
b. Elimination by aspects C. Mathematical programming methods
1. Global. objective function a. Linear programming 2. Goals in constraints
a. Goal programming 3. 1,ocal objectives: interactive
a. Interactive, multicriterion prograrnming D. Spatial proximity methods
1. lsopreference graphs a. Indifference map 2. Ideal points
a. Multidimensional, n o n m e t r i c scaling 3. Graphic preferences
a. Graphic overlays
t h e s t e p m e t h o d (Section 2.3.2);
decision analysis (Section 2.3.3);
t h e r e f e r e n c e point approach (Section 2.3.4)
We shall consider briefly the s t r e n g t h s a n d weaknesses of four oi t h e classes listed above in o r d e r t o define options for o u r case study. We a r e convinced t h a t not just any model will solve all t h e problems. For t h i s reason we shall t r y t o i m p l e m e n t two o r t h r e e of t h e m a n d c o m p a r e t h e r e s u l t s obtained (Section 4).
TABLE 5 A taxonomy of m e t h o d s for rnultiobjective decision making ( f r o m Hwang e t al. 1980). MOLP, multiobjective goal programming; SEMOPS, sequential multiobjective problem-solving technique;.GPSTEM, a c o m b i n a t i o n of goal-programming a n d STEM m e t h o d s ; SIGMOP, s e q u e n t i a l information g e n e r a t o r for multiobjective problems.
S t a g e a t which Type of Major c l a s s e s of m e t h o d s i n f o r m a t i o n is n e e d e d i n f o r m a t i o n
1. No a r t i c u l a t i o n of p r e f e r e n c e information
Apriori articulation 2.1. Cardinal of p r e f e r e n c e i n f o r m a t i o n i n f o r m a t i o n
2.2. Ordinal a n d c a r d i n a l i n f o r m a t i o n 3. Progressive 3.1. Explicit
a r t i c u l a t i o n of trade-off p r e f e r e n c e i n f o r m a t i o n
( i n t e r a c t i v e m e t h o d s )
3.2. Implicit trade-off
4. Aposteriori 4.1. Implicit
a r t i c u l a t i o n of trade-off p r e f e r e n c e i n f o r m a t i o n
(nondoniirlated solutions g e n e r a t i o n m e t h o d )
1.1.1. Global c r i t e r i o n m e t h o d 2.1.1. Utility f u n c t i o n
2.1.2. Bounded objective m e t h o d 2.2.1. Lexicographic m e t h o d 2.2.2. Goal p r o g r a m m i n g 2.2.3. Goal a t t a i n m e n t m e t h o d 3.1.1. Method of Geoffrion a n d
i n t e r a c t i v e goal p r o g r a m m i n g 3.1.2. S u r r o g a t e worth trade-off m e t h o d 3.1.3. Method of s a t i s f a c t o r y goals 3.1.4. Method of Zionts-Wallenius 3.2.1. STEM a n d r e l a t e d m e t h o d s 3.2.2. SIGMOP m e t h o d
3.2.3. Method of displaced ideal 3.2.4. GPSTEM m e t h o d
3.2.5. Method of S t e u e r ( I n t e r a c t i v e MOLP m e t h o d )
4.1.1. P a r a m e t r i c m e t h o d 4.1.2. & - c o n s t r a i n t m e t h o d 4.1.3. MOLP m e t h o d
4.1.4. Adaptive s e a r c h m e t h o d
2.3.1 Goal Programming
Goal programming (GP) is frequently proposed t o deal with problems with multiple objectives. Surveys of t h e s t a t e of t h e a r t have been made by Korn- bluth ( 1 9 7 3 ) and Nijkamp and Spronk ( 1 9 7 7 ) . The approach has been applied to a large n u m b e r of practical problems i n a wide variety of fields, ranging from manpower planning t o environmental protection. Goal programming minimizes a weighted combination of t h e deviations from a n u m b e r of goals ( t a r g e t levels, aspiration levels) s e t by t h e decision maker. This a s p e c t dis- tinguishes GP from t h e theory of t h e displaced ideal (Zeleny 1 9 7 6 ) .
The large n u m b e r of applications c a n be explained by t h e flexibility of t h e m e t h o d a n d by its correlation to r e c e n t results of behavioral theory. The following versions have been developed:
i n t e r a c t i v e GP (Dyer 1972): according t o t h e classification of Lari- chev and Polyakov ( 1 9 8 0 ) , Dyer's method is pseudostructured and t h e information required is difficult t o obtain (Spronk 1979);
TABLE 6 Larichev's (1979) classification of m e t h o d s of multiobjective decision making.
Class Basic i d e a
1. Axiomatic methods Several axioms a r e i n t r o d u c e d and t h e i r validity is t e s t e d i n order t o c o n s t r u c t a m u l t i a t t r i b u t e utility function of a specific type (von Winterfeldt a n d Fischer 1975, Keeney a n d Raiffa 1976, Hum- p h r e y ~ 1977).
Direct m e t h o d s
Prescription of both t h e f o r m of t h e aggregation function a n d all i t s p a r a m e t e r s .
Application of specific cri- t e r i a (Savage. Wald. La- place. Hurwicz) according t o t h e wishes of t h e deci- sion m a k e r u n d e r condi- tions of unknown proba- bilities of t h e s t a t e s of t h e world.
Postulation of t h e aggre- gation rule; p a r a m e t e r s a r e d e t e r m i n e d by t h e de- cision m a k e r .
Postulation of t h e aggre- gation rule; p a r a m e t e r s a r e d e t e r m i n e d by calcu- lations.
Postulation of t h e r u l e of maximization of expected value (utility).
Compensation methods Method of t h e thresholds of incomparability
5. Interactive m e t h o d s
Decision m a k e r prescribes t h e form of t h e aggre- gation function for t h e m e a s u r e m e n t (or assess- m e n t ) in t e r m s of t h e individual objectives.
Decision m a k e r defines s t e p by s t e p a compro- m i s e between t h e objectives.
Comparisons of t h e alternatives a r e made in pairs, separately for e a c h criterion. An index i s calculated a n d t e s t e d a g a i n s t t h r e e thresholds s e t by t h e decision m a k e r . The relationship between alternatives is d e t e r m i n e d a s "strongly preferred," "weakly preferred," or "no prefer- e n c e . " A ranking is developed from t h e prefer- e n c e m a t r i x (Roy 1977, Larichev 1979).
Interactive m e t h o d s a r e applied when t h e model of t h e choice situation i s only p a r t l y known. The relations between t h e c r i t e r i a a r e described in a n i n t e r a c t i v e process between decision m a k e r a n d c o m p u t e r .
- --
