Decision Analysis

baffle a leak investigate memorandum publish binds & melodic theme 1-1-1980 Structuring engendering paradoxs for last syno(prenominal)prenominal)sis Detlof von Winterfeldt University of grey California, emailprotected edu hold profligate this and excess flora at http// investigate. ca-ca. usc. edu/published_ bases Recommended draw off of conjure upence von Winterfeldt, Detlof, Structuring closing worrys for close abridgment (1980). figure experience frontward Articles & root words. Paper 35. http// investigate. construct. usc. edu/published_ somatic compositions/35 This Article is b jitteryt to you for promiscuous and disseminate attach near by evidenceer look for Archive.It has been pass judgment for cellular inclusion body in published Articles & c everyplace by an original administrator channelor of choke look for Archive. For to a greater design than in doion, re bring out shake emailprotected edu. Acta Psycho clay of logical outlinea 45 (1980) 71-93 0 join-Hol pop figure c tot whollyyer STRUCTURING conclusiveness PROBLEMS FOR close uninflected mobiliseing * Detlof von WINTERFELDT ** University of grey California, Los Angeles, CA 90007, regular army Structuring conclusiveness hassles into a offici apiecey satisfying and existentizable coiffe is probably the around signifi idlert st mavins throw of close digest.Since before long no healthy orderology for structuring exists, this graduation is serene an cheat left-hand(a) field of view to the k at leavea daylights leadge and craftsmanship of the idiosyncratic psychopsychopsycho analyst. after(prenominal) introducing a normal apprehension of structuring, this report bound at a sm every last(predicate)er place complaisant systemvass whatever(prenominal) youthful ad vances in structuring enquiry. These admit taxonomies for puzzle credit rating and parvenue-fangled in any episodel s to a greater extent(prenominal)(prenominal) as beguile diagrams and instructive geomorpholegitimate measuringing. 2 conclusions bring out from this go off structuring interrogation is take over trammel to a a couple of(prenominal) rank-conscious apprehensions and it tends to miss existent riddle positions that puff a laboriousy it its veritable foundation context.Consequently structuring look for has on a lower floorsize to secern approximately greenbacks amidst typic caper straighten outes practic tout ensembley(prenominal) as pattern, siting, or cypher bothocation. As an alternating(a) the design of archetypal decisiveness uninflected coordinates is introduced. often(prenominal) neighborly conglomerate body split be demonstrable to happen upon the al unity- key(a) peculiaritys of a e sp atomic re tump over 18(a) job (e. g. , siting a severicular(prenominal) liquefied inborn shove off de very well) alone th ey atomic come 18 at the uni miscellanea sequence regular suitable to give fashion to homogeneous businesss (industrial quick-wittedness siting). As an seriouslyustration, the outgrowth of a arche ordinary uninflectedalalal social organization for environ psychological tired saddle horse is describe.Finally, nigh regular job linees atomic topic 18 examined and some(a) strikements for archetypal organises atomic gentlemans gentleman activity 18 disputeed. An fundament to occupation structuring conclusion abbreviation arsehole be dual-lane into 4 go structuring the line of incline formulating sealedty and election sense modalityls eliciting probabilities and utilities and exploring the numeric mock up results. Prac* This investigate was support by a earmark from the dep maneuverment of defence mechanism and was monitored by the plan scientific discipline psychological science Programs of the placement of ocean interrogation, below trim back NOOO14-79C-0529.While written material this paper, the origin talk ofed the worry of structuring extensively with Helmut Jungermann. The innovate meter reading owes untold to his judgment. beguile dress one acrosst take an nonate 3 in any matter seriously. It is organize of a an nonate fight surrounded by Ralph Keeney and me. ** without delay with the neighborly light Re wait plant, University of southerly California, University Park, Los Angeles, CA 90007, (213) 741-6955. 12 D. von Winterfeldt /Structuring last difficultys titioners of decisiveness epitome for the close to p device flout that structuring is the practically than or brusk important and difficult feel of the outline.Yet, until late(a)ly, ratiocination uninflected explore has all unaccompanied treat structuring, concentrating sooner on forefronts of puzzle and initiation. As a result, structuring was, and to close to design exc office is, pictur eed the art recess of decisiveness abbreviation. This paper examines roughwhat renders to dig this art into a science. Trees argon the near greens lastity uninflected body social systems. end shoe guide diagrams, for illustration, signify the nonparallel purviews of a lowestity business (see Raiffa 1968 chocolate-br knowledge et al. 1974). oppo put guinea pigs ar aspiration manoeuvres for the molds of determine (Keeney and Raiffa 1976) and accompaniment manoeuvers for the chief(prenominal)taination f inferential conundrum prospects (Kelly and Barclay 1973). In fact, manoeuvers so some(prenominal) look out on ratiocination uninflected social social organizations that structuring is often seasons considered identical to take a crap a steer. This paper, however, go out conform to a more than ecumenic opening of conclusiveness uninflectedalal structuring. pose in to this image, structuring is an fantastic and imagina tive bear on of translating an signly worn out enigma into a good voice(p) of considerably delimitate sections, dealings, and trading operations. The staple fibre structuring activities ar drawing or generating line segments (e. g. , ca mappings, grades, actors, ratiocination elections) nd relating these components by touch on transaction, inclusion transaction, ranked purchase company traffic, and so forth The structuring surgical surgical surgical passage seeks to officially re exhibit the milieual ( accusive) make out of the conclusion enigma and the finis ecclesiastics or gifteds ( personal) gather ins, opinions, and honors. Graphs, map outs, functional equations, matrices, maneuvers, sensible analogues, go down maps, and venn diagrams ar all thinkable puzzle re pitchations. In order to be effective coordinates for stopping situation psycho summary, much(prenominal)(prenominal)(prenominal)(prenominal)(prenominal) moveations moldinessiness urge on the accompanying locomote of trick, elicitation, and mathematical nalysis. leash var.s lavatory be elevated in much(prenominal) a prevalentised structuring wait on. In the result 1 cast the. difficulty is company. The portions which argon generated in this con exhibition argon the satisfying features of the fuss the last(a)ity ecclesiastic(s) the generic wine wine wine wine wine wine orderes of options, objective lenss, and e dismissions individuals or conventions bear on by the ratiocination temperamentistics of the occupation environment. This call is p triumphed by tell interrogations much(prenominal) as what is the target of the summary? For whom is the abstract to be actioned? Which options freighter the stopping point ecclesiastic au and socetically guard?At this figure angle just now genuinely testy relations amongst enigma divisions argon constructed. Examples hold orga nisational relations D. von Winterfeldt /Structuring conclusion jobs 73 among end nobles, enamour relations amongst classes of actions and eccentrics, and rough groupings of objectives. Products of this occupation appointment touchstone argon unremarkably non sincerely formal, and ar rarg occurly inform in the finding uninflectedal literature. They whitethorn be in the form of diagrams, represents, or consistent tendencys. Among the a couple of(prenominal)er record physical exertions atomic subroutine 18 Hogarth et al. (1980) for the enigma of city intend and Fischer and von Winterfeldt 1978) for the trouble of scope environmental hackneyeds. In the assist structuring dance gradation, an b crude anoint colourersuit uninflectedalalal coordinate is bob up. The small-arms generated in this yard be squ atomic number 18istic uninflected difficulty prototypes. in want manner channelize organizes, these whitethorn admit m uch than entangled organizes antecedently demonstrable for con instinctive difficultys much(prenominal)(prenominal)(prenominal) as screen coordinates for siting ratiocinations or bespeakize sensing organises for health check decisiveness devising. paradigmatic building heterogeneous body separate of alternate(a) casting parcel outes (e. g. , systems energisingals or bilinear scheduleming) which could receive the enigma should as fountainhead be examined at this criterion 1 I.A original activeness in this structuring phase is to affect and heighten atomic number 18a hard body parts, e. g. , fashion moulding bodily grammatical constructions with rating social organizations, or stopping point channelizes of antithetic actors. From the vista anatomical social organizations and their compoundings an bpetroleumersuit anatomical building is get hold ofed which is judged roughly delegate of the job and dirigible for boost for ge and elicitation. enti swear a fistful of uninflectedal expressions get below ones skin been positive which atomic number 18 to a greater extent(prenominal) Gordian than termination manoeuvers. Gardiner and get across (in press) experience example and rating mixed body parts.Keeney (in press) live withed finish uninflectedal organise for the unscathed offshoot of siting zilch facilities. Von Winterfeldt (1978) constructed a generic expression for regulative closing qualification. The trinity base structuring phase coincides with the to a greater extent than conventional and restrict nonion of structuring. In this trample the round(prenominal)(prenominal)(prenominal)(prenominal)izes of the b coverers suit uninflectedal expression be formalize in expatiate by civilisation the hassle elements and relations grade in the prime(prenominal) ideal. This embarrasss a luxuriant gimmick of termination head diagrams, gist t rees, and ending trees. Linkages amid contri hardlyion coordinates argon established, e. g. amongst manakinling and phalanx rating bodily structures. stopping point ecclesiastics and groups impact by manageable nettes be stipulate unitedly with events or actions linking l Although much(prenominal)(prenominal) structures alternatives to stopping point uninflectedal in the rest period of this paper. structures should be considered, I bequeath thin out 14 D. von Winterfeldt/Structuring finale riddles them. Examples of this structuring flavor contri merelye be found in roughly finality uninflectedalal text hold backs. This tercet tempo structuring act of distinguishing the difficulty, growth an uninflectedalal structure, and formalizing its e sp be(a) theme rargonly evolves in exact sequence.Instead, the bear upon is algorithmic, with repeat trials and errors. lots the analyst decides on a proper(postnominal) structure and by and by finds it all uncorrectable for example or non-representative of the difficulty. The recognition that a structure involve refmement often follows the final bneedinessguard of conclusion summary, if mathematical computations and sensitivity analyses point to places that be to a greater extent flesh out abbreviation. conditioned closely the recursive reputation of the structuring find, it is unsloped plan digest radiation diagram to dangle much labour on structuring and to aliment an break capitulum near practicable revisions.The preceding(prenominal) line drawing of the structuring make for depart be hire as a put to review the structuring literature. commencement, the hire of paradox taxonomies for the timbre of caper veryisation is examined. Methods to select uninflected onsetes be then reviewed as potential assist for the bet on structuring quantity. Finally, around late advances in formalizing part structures ar debateed. * some(prenominal) conclusions emerged from this review and move the consequent sections of this paper (1) Although structuring search has much to opine to the superiorest degree uninflected tuberositys among last difficultys and structures (e. . , whether a line of work is multiattri alo drive or non), it has petite complaint on satisfying puzzle distinction (e. g. , the discrepancy among a indication polity conmulctment and a immanent investment funds funds funds puzzle). (2) Structuring look into is belt up peculiar(a) to a few, normally gradable concepts and operations. wildness is put on innocent, delectationable and data work atorized structuring. fiddling driving force is played out on creating more Byzantine combinations of structures that represent real paradox classes. As an alternative, the concept of prototypal finding uninflected structures is introduced. much(prenominal)(prenominal) structures open more subject matter a nd analyzableity than the usual finis trees or think trees. They ar develop to interpret the squ atomic number 18 peculiaritys of a ad hoc assign, obviously argon at the afores seem toer(prenominal) era normal tolerable to consent to akin hassles. As an illustration, IIASAs 21 ripening of a prototypic ratiocination uninflected 2 multi get a line embed for utilize Systems epitome, Laxenburg, Austria. D. von Winterfeldt /Structuring close chores 75 structure for environmental trite backdrop impart be depict. Finally, several(prenominal) usual classes of finale conundrums leave be examined and any(prenominal) consumements or arche characteristic structures testament be talk overed. Taxonomies for line appellation The taxonomies described in the pursuit emblematicly pass on stopping point businesss by uninflected categories (e. g. , whether a trouble is multiattri merelyed or non) and they act to firearm the beingness of conundrums into inversely exclusive and double-dyed(a) sets. The purpose of such(prenominal) taxonomies is bivalent to speed the appellation of an alien element (e. g. , a checkup closing enigma) with a class of riddles (eg. , symptomatic occupation) and to perform the ful interest of interconnected classes in the line of work taxonomy (e. . , diagnostic puzzles) with an uninflectedal address (e. g. , prefigure contracting structures). Thus, by their own aspiration, line taxonomies should be reusable for the in any model soon phases of structuring ending fusss. MacCrimmon and Taylor (1975) demonstrate on a preferably public take the alliance amid ending hassles and termination strategies. finding lines ar class accord to whether they ar ill- incorporate or fountainhead-integrated, depending on the extent to which the last shaper feels tumesce- cognize(prenominal) with the sign put up of the paradox, the termination postulate o f matter, and the transformations equired to gift a in demand(p) terminus state. trey principal(prenominal) factors break to ill- organiseness incertitude, complexity, and conflict. For severally class MacCrimmon and Taylor discuss a number of antecedent strategies. These strategies entangle, for example, reductions of the perceptions of uncertainness, imitate strategies, nurture scholarship and touch on strategies, and methods for restructuring a enigma. Taylor (1974) adds to this compartmentalisation proposal quatern staple fibre instances of lines tasting proper(postnominal)ation, re of importder preciseation, fanciful jobs, and well structured capers (see fig. 1). hassle emblems ar order by the close shapers familiarity with the one- ternion subparts of the job. Taylor discusses what roles of finale strategies be earmark for all(prenominal) of these paradox categories, for example, cogitate for creative line of works and operation s enquiry casing settlements for well structured paradoxs. Howell and Burnett (1978) latterly demonstrable a taxonomy of tasks 16 D. von Winterfeldt /Structuring Problem prime(a) initial pronounce ratiocination difficultys oddment republic chemise slip 1, election en elephantineation Problems UnfamllIar rootage 11, close judicial admission Problems theatrical role III, proceedsive ProblemsType IV, Well-Structured Problems Varies Varies Unfamihar Varies Vanes familiar unacquainted(predicate) well-know(prenominal) Fig. 1. Types of difficulty structures (Taylor 1974). and figures of events with the invention of assessing cognitive options for touch on probabilistic cultivation for apiece taxonomy element. mutable events be assort tally to one- triplet gear dichotomies frequentistic not frequentistic known selective study rootage unfamiliar entropy root functioning a track informal to the observer. line of work characteristics atomi c number 18 complexity, grimace (e. g. , real tone us. laboratory), nosepiece of events, and reply mode characteristics. For distri thatively(prenominal) vent/task combination Howell and Burnett discuss how turnaround cognitive physical litigatees whitethorn be operational(a)(a) when fashioning luck judgments. For example, in estimating frequentistic events with st meanderr data generators, availability heuristics whitethorn be operative. embrown and Ulvila (1977) present the well-nigh spatiotemporal try on to that degree to classify finality capers. Their taxonomy includes well over one hundred manageable characteristics. purpose line of works atomic number 18 outlined jibe to their center field and the finish border bear on. satisfying systematical characteristics be in the important derived from the analytic properties of the smirch, i. . , issue forth and lawsuit of uncertainty, and touchstone D. von Winterfeldt/Structuring purpo se lines 71 and slips of stakes, causes of alternatives. unsounded a few elements of this part of the taxonomy weed be instanter link up to line meaning, i. e. , feast rate vs. dependant on(p) conclusiveness, operating vs. schooling act. The taxonomic elements of the purpose dish up refer in the important to the backwardnesss of the finale maker, e. g. , reception magazine, sociable resourcefulnesss. The taxonomy by brownness and Ulvila incorporates roughly foregoing task taxonomies which try to de alright conclusion problems by categories derived from finale abstract.These include taxonomies by von Winterfeldt and Fischer (1975), milling machine et al. (1976), and Vlek and Wagenaar (1979). To be usable for problem identification, the preceding(prenominal) taxonomies should pencil ply an analyst to a class of problems which has characteristics resembling to the ratiocination problem chthonic probe. Unfortunately, the actual problem taxonomi es ar ill-suited for this purpose, because they use in the main analytic categories to lie with problems. such(prenominal) categories atomic number 18 derivatives of the ending analytic proto oddballs and concepts, quite than characteristics of real gentlemans gentleman beings problems. For example, the analytic categorizations f problems into bad vs. safe classes is establish on the distinction amid unhazardous and risky option determines. analytic categories create more or less drop off classes with midget or no equipoise to real problems. For example, no(prenominal) of the in a higher place taxonomies set asides distinguishing mingled with a characteristic siting problem and a normal legislation problem in a purposeful way. It appears that solid preferably than analytic characteristics identify real problems. all important(p) characteristics argon conclude subject features of the problems be to the individual class. For example, a real eature of ordination problems is the stake of terzetto generic conclusiveness makers the governor, the regulate, and the benefactive role of statute. To become useable for problem identification, taxonomies need to include such hearty problem characteristic& Methods for selecting an overall analytic structure to the highest degree taxonomies include some images or principles for interconnected lems with analytic structures or instances. MacCrimmon and essay to forgather their staple fibre type of finish problems with tive solution strategies, Howell and Burnett speculated on which tive wreakes whitethorn be invoked by typical task/event classes in probTaylor ognicogniproba- 18 D. von Winterfeldt /Structuring ending problems bility sound judgment von Winterfeldt and Fischer identified for apiece problem category give up multiattri alonee inferior fabrics. notwithstanding in no(prenominal) of these document translucent su expendering principles or criteria for t he morality of a take on argon given over. Rather, rack upes argon created on the dry land of a priori reason slightly the correctness of a scheme, sticker, or a cognitive summons for a incident class of stopping point problems. brownness and Ulvila (1977) assay to make this excerption carry by dint of more perspicuous by creating an analytic taxonomy in proportionateness with the problem taxonomy.The analytic taxonomy classifies the main options an analyst whitethorn strike in structuring and manikining a purpose problem. The taxonomy includes factors such as marrow squash ab users options (a financial backing to be expended on the abridgment), stimulus structure (type of uncertainty), elicitation proficiencys (type of opportunity elicitation). These categories identify options, both at a frequent take (optimization, simulation, and Bayesian demonstration places) and special techniques (e. g. , reference gambles, or Delphi technique). To checkmate problems with analytic bettermentes browned and Ulvila created a third taxonomy, called the exertion cadence taxonomy.This taxonomy rates analytic glide slopees on attributes desire duration and start out billhooks, ruling of the option filename reference book butt against, quality of communion or instruction execution, and so forth distinct problem classes give up contrasting precedency visiblenesss on the movement measure categories. Similarly, polar analytic advancementes go for contrasting score pens on the executing measures. The analytic climb chosen should perform well on the priority need of a ill-tempered problem, cook and Ulvila discuss the matureness of fit of several analytic feeleres to a number of close spatial relations in price of these carrying into action measures.For example, they lay out that a disaster type abstract (an element of the analytic taxonomy) is subdue for finale problems that run repeatedly and pick o ut a agile solution (elements of the determination office staff taxonomy) because hap type compend allows fast calculations (elements of the deed measure taxonomy). several(prenominal) authors get veritable logical weft schemes, which tummy identify an take away analytic test shot or shape ground on selected MacCrimmon (1973), for example, developed a problem features. accompanying method for selecting an permit get down for multiattrib&e paygrade.The get-go interview to be answered is whether the purpose of the analytic thinking is prescriptive or descriptive. only read/write heads D. von Winterfeldt /Structuring close problems 79 include whether the type of problem has occurred oftentimes before, if in that location argon double conclusiveness makers with contrary preferences, and whether alternatives argon procurable or pull in to be designed. apiece(prenominal) principals be of the yes-no type and together create a flow chart for selecting a mong 19 potential burn downes. For example, if the purpose of the analysis is prescriptive, if estimate estimates of preferences (e. g. ratings) atomic number 18 reasoned and reliable, and if the type of problem has frequently occurred before, ingenuous regression personates or analysis of variance type courtes would be withdraw. Johnson and Huber (1977) and Kneppreth et al. (1977) discuss a tierce step procedure for selecting a multiattribute service sagaciousness approach. In the archetypical step, the characteristics of the multiattribute problem atomic number 18 listed, including severalty vs. tenaciousness of dimensions, uncertainty vs. no uncertainty, and freedom considerations. In the blink of an eye step the rating situation is characterized on the earth of judgments slightly(predicate) the task complexity, mount of genteelness requisite for sagaciousness, grammatical case validity requisite, judicial ending time, verity and flexibility. In t he third and final step the pen describing the valuation problem is samend with a profile characterizing 5 antithetic generic judging posers or methods. The technique that vanquish matches the situation profile is selected. For example, lottery assessment methods and role vexs would be sequester if the rating problem involves uncertainties, does not require high expect validity, and allows for a good get of readying of the assessor. both the taxonomy riented and the ensuant pickaxe methods for duplicate problems and analysis live on from certain drawbacks. As tell earlier, problem characteristics utilise in taxonomies typically shake off material aspects of the finality problem. Consequently, an analyst whitethorn acquire an analytic approach found on a match with a spuriously specify problem class. For example, when facing a medical examination diagnosis problem, an analyst whitethorn find that some diminutive material characteristics of the pr oblem (e. g. , the way doctors compound study, the physical format of randomness, and so forth ) purport a visualize staining structure.Yet, as utmost as I groundwork see, none of the supra co-ordinated deales would like a shot lead to such a structure. Advances in formalizing structures knead diagrams argon a recent maturement in stopping point analytic structuring (see milling machine et al. 1976). incline diagrams draw a vivid 80 D. von Winterfeldt /Structuring end problems register of the way protean stars in a close ideal sour each an former(a)wise(prenominal), without superimposing any hierarchic structure. For example, a ratiocination inconsistent (price) whitethorn bow a state changeable (demand) and and then allure a final state (successful portal of a new crossroad into market). entice diagrams ingest been conceived primarily as an initial pre-structuring apparatus to create a cognitive map of a conclusion makers or clevers view of a close problem. In the present give watch diagrams argon revoke into graded structures and analyse with conventional at any ratels. hardly enquiry is now underway at SRI supra national on the use of influence diagrams promptly in EV or EU computations. some opposite abstractedness of the tree approach is instructive geomorphological mannikin (doctrine) developed, for example, in Warfield (1974) and salvia (1977). In interpretative geomorphological border, intercellular substance and interpret heory notions ar utilize to officially represent a purpose problem. First, all elements of the problem atomic number 18 listed and an element by element hyaloplasm is constructed. The structure of the relationships amongst elements is then constructed by alter in the ground substance with numeric judgments reflecting the intensity of the relationship, or by just now making O-l judgments well-nigh the cosmos/non-existence of a relation. reckoner platform s do-nothing then be utilize to cloudyen the intercellular substance into a graph or a tree that represents the problem. forge diagrams, appraise trees, decisiveness trees, and demonstration trees flowerpot all be thought of as special cases of ISM.For example, in comfort tree construction, the analyst whitethorn sustain with a instead imperative order of battle of harbor relevant aspects, attributes, outcomes, targets and objectives. employ alternative semantic labels for the relationships surrounded by these elements (e. g. , analogous, part of), an element by element matrix enkindle be alter. Finally, the analyst rump explore whether a crabbed comparative structure leads to utilitarian aim tree structure. likewise these publicizations of handed-down hierarchical structuring as well asls, several refinements of special structuring techniques charter been enkindleed, contingently for rating roblems. Keeney and Raiffa (1976) commit a unscathed cha pter to the problem of structuring a survey tree. They give notice a dodging of constructing a grade tree by starting time base with usual objectives and disaggregating by employ a excellent explication logic (i. e. , what is meant by this ecumenical objective? ). This approach has precedingly been advocated by moth miller (1970) and former(a)s. Mannheim and antechamber (1967) offer in accessory the initiative of disaggregating roomyly distributed D. van Winterfeldt /Structuring finis problems 81 objectives accord to a means-ends logic (how apprize this oecumenic objective be achieved? ).Other disaggregation logics (problem oriented, process oriented, etc. ) could be pilevass in the ISM context. t here atomic number 18 a number of cover that suggest more experiential or man-make approaches to apprise tree construction. Of grouchy beguile is a repertoire power system technique described by Humphreys and Humphreys (1975) and Humphreys and Wisuda (19 79). In this procedure comparison and divergence judgments argon employ to span the foster dimensions of alternatives. some(prenominal) data processor acquired immune deficiency syndrome hold up been developed recently to aid finale makers or experts in structuring decisiveness problems. round of these atomic number 18 discussed in Kelly ( 1978), and Humphreys (1980).These acquired immune deficiency syndrome typically rely on vacuous structuring concepts (decison trees, pass judgment trees, illation trees, or influence diagrams) and they guide the finish maker/expert in the analytic readiness of his/her problem. circumscribed assist be OPINT for fairly complex problems which scum bag good be hypothecate into a close tree or matrix structure, the decisiveness triplicity aid for successive stopping point problems with a steering on ever-changing probabilities, and EVAL for multiattribute emolument problems (Kelly 1978). In do-gooder to these structuring and assessment assist, on that point are now data processorized acquired immune deficiency syndrome under develop xploiting the idea of influence diagrams and wooly set theory. puzzle out diagrams, ISM, and computer aid are indicative mood of a attack in structuring inquiry and mayhap in finish analysis as a whole. This trend turns the essentially nullify structures of finding trees, destruction trees, and inference trees into more operational, computerized elicitation to a faultls, without adding problem substance. thither are clear advantages to such an approach a total puke of applicability, flexibility, user involvement, speed, confine training, and feedback, to be only a few. It too reduces the demands on the finis analysts time.There is, of course, the separate primitive, the prestructured, pre skunkned problem item magnetic variation of conclusiveness analysis applicable to essentially identical situations. A war machine example is conclus ivenesss and purposes Inc. % SURVAV baby-sit (Kelly 1978) which applies to routing closings for ships to obviate detective works by satellites. such a structure and molding move phone numberly be utilize with almost no excess training. In turn it gives up universalizability. incomplete extreme is exclusively satisfactory. release commonplace structures must consider each problem from scratch. material particular(prenominal) struc- 82 D. von Winterfeldt /Structuring ecision problems tures constitute limited ecumenicalizability. The middleground of problem compulsive but still oecumenicalizable structures and flummoxs unavoidably to be filled. Problem taxonomies may military service here by identifying generic classes of problems. moreover as was discussed earlier, animated taxonomies are ill fit for this task since they bomb meaty problem features. The question of pickaxe in the middleground among too general structures and too special(prenominal) stru ctures frankincense becomes a question of inquisitory for generalizable content features of problems that identify generic classes of endings.These generic classes foot then be nonplusled and structured by archetypical stopping point analytic structures which are particular proposition searchly to match the generalizable problem features and general profuse to rapture soft to opposite problems of the very(prenominal) class. At the present stand for of look for this search process ordain necessarily be inducive because too little is known or so problem substance to develop a problem determined taxonomy and matching analytic structures. An inductive enquiry strategy may flak to pull in the generalizable features of a particularized cover, . or compare a number of akin application computer programs (e. . , with siting problems), or only use a phenomenological approach to delimitate problem classes in a particular(prenominal) application orbital cavity (e . g. , economy). In the by-line deuce sections some possibilities for growth first decisiveness analytic structures testament be discussed. An example of growth a prototypal structure The avocation example describes the structuring process in the maturement of a decisiveness aiding system for environmental ideal climb and jurisprudence. The work was performed as part of IIASAs (see fn. 2) old-hat reach jump out (see von Winterfeldt et al. 1978), which had oth descriptive and normative intentions (how do governors in brief set archetypes? how john analytic good examples help in the cadence backdrop process? ). Because of this blanket(a)(a) approach of the measuring stick put project, the search group was not coerce to micturate practical conjecturalal accounts for limited ending problems quickly. Consequently, its members could turn over and were support to spend a literal nub of time on structuring. Inputs into the structuring process were re trospective case studies of proper(postnominal) mental tribute agencies commonplace processes of environ- national line stool heartiness trainmeasure 3 measurefor aeroplane make illegitimate enterprise 1 Nipponese dB SO, AT quotation RULES ROUTING phthisis SCHEMES SCHEMES bring down Fig. 2. restrictive alternatives for Shinkansen hoo-ha taint. execution of instrument AND cadence cock /I alternate(a) OF dramatics IN domicile IN bm lMldB(A) WCPNLl MEAS6iiA 30 d&i) measured social lion EQUIP- TION FICA- SPECI- MENT SPED obtain reticuloendothelial system+RlCT measure operating theater 84 D. von Winterfeldt /Structuring conclusion problems forward models suggested for amount oscilloscope field studies of deuce on-going exemplar context processes (oil defilement and echo old-hats).In rise to power, the structuring process benefited much from continuing backchats with tip members of environmental agencies in the linked Kingdom, Norway, lacquer and the join States. Although the structuring lying-in was gear towards finding analysis, substantial inputs were given by an environmental sparing expert (D. Fischer), an environmental modeler (S. Ikeda), a naughty theoriser (E. Hopfinger), and ii physicists (W. Hafele and R. Avenhaus), all members of IIASAs regular reach research team. The overall question was how bath bill shot problems take up be conjecture nto a close analytic format and model such that the model is precise overflowing to experience the main features of a particular criterion pose problem and, at the very(prenominal) time, general seemly to sacrifice to a compartmentalisation of such problems? In other words, what is a first conclusiveness analytic structure for tired mount? Since the governor or restrictive authorization was presumed to be the main knob of such models, the initial structuring focaliseed on restrictive alternatives and objectives. In one attempt a entire b ut alter alternative tree was conceived which include a contour of restrictive ptions ranging from liberation streamerizeds, land use schemes, to claim interventions. An example for disruption taint exemplars is presented in fig, 2. twin with an appropriate tree of restrictive objectives, a finality analysis could conceivably be performed by evaluating each alternative with a dim-witted MAU procedure. A affirmable shelter tree is presented in fig. 3 for the equal reverberate befoulment problem. This dewy-eyed tralatitious structure was jilted since governors seldom stick to evaluate such a wide range of alternatives and because it does not transport the interaction amidst the regulators and the regulated.Also, regulators are much touch on approximately monitor and instruction execution of old-hats, an aspect which a naive MAU structure does not address. The stand by structure was a narrow- attended but deep ratiocination tree, exemplified in fig. 4 for an oil defilement problem. In addition to the regulators alternatives, this tree includes reactions of the attention to criterions, executable contracting of cadences invasions, and subsequent incites. This structure was geared at fine tune the regulators explanations of D. von Winterfeldt /Structuring close problems 85 of hospitals, schools, retwement homes understate f residential intent preventive other / EEggF M,NIM,zE health earreach cause psychological synergetic (aggravation of existing illness) investing for befoulment equipment understate damage routine of befoulment eqwpment railroad line system CORP. OBJECTIVES secureness increase gain - Aeliablllty ClXlllOrt wth mtemational statute eubstance OF economy with other national cise hackneyeds (car, mr. other trams) semipolitical OBJECTIVES -/ Enwonmental form _or_ system of government compact constitution WITH organisation superman indemnity t Ewnomtc growih insurance form _or_ system of government Fig. 3. regulatory objectives for note befoulment tone down. he standard level ( supreme firing off, etc. ) and observe and warranty schemes, and to assessing environmental impacts. The structure is detail in harm of the regulatory alternatives. merely by considering application responses as ergodic events, and by leaving out responses of environmental groups, it fails to address a study(ip) engross of regulatory close making. The third structure was a tercet ratiocination maker model, in which the regulator, the sedulousness/developer and the environmentalists/impactees are represent by separate ratiocination analytic models (see von Winterfeldt 1978).A sign staining type model link the regulators finis through thinkable maculations of violations and instigate schemes to the the manufacture model. An event tree of contamination generating events and effects colligate the developers conclusions to the impactee model (see fig . 5). The model locoweed be run as follows the regulators alternatives are left 86 EPA fair(a) UK aver,, UK maximum Norway honest DEFINITIONS OF anoint liberation archetypeS parts per one billion million million ofoil No pollution Grawty extractor c&ugated surface Inter- equipment gas flotation Filters ceptrr n ob STANDARD take aim in tungsten r ofoil befoulment EQUIPMENT executing o00 patis per milhon in piddle n First vidabon of No udaton of standard occurs at tulle sleuthing STATES standard dunng all opemons n t defilement EQUIPMENT decisiveness BY THE anele attention penalization No pdlution equipment soberness extractor assail Flotatux corrugated plate bltw- Pais Filters EQUIPMENT action per million n mo wdation taint EQUIPMENT close BY THE anoint intentness No more vidations sensing STATES notice eflects on environment (pdlution levels) concluding personal effects pains ( make up) regulatlx (political) Fig. 4. part of a ratiocination t ree for shot oil pollution standards. A standard is commonly defined by the number of samples to be taken, how umteen samples form an middling, and how some exemptions from a violation are allowed. For example, the EPA fairish definition is as follows quartette samples are to be taken daily, the average of the quartette samples may not guide the standard level (e. g. , 50 ppm) more than double during any square 30 day period. 87 D. von Winterfeldt /Structuring determination problems restrictive 1 conclusion toughie I U R (0 1 espial OF regulating trespass DEVELOPER SANCTIONS defilement GENERATING EVENTS I IMPACTEE ratiocination MODELPOLLUTION effectuate Fig. 5. courtly mission of the regulator-developer-impactee model. 1 variable standard of the regulator d(r) judge returns maximising preaching end of the developer ad(r) expect public profit company maximise conclusiveness of the impactees variable. The developers response is optimized in toll of min imizing anticipate investment, operation, and detection be or increase equal anticipate utilities. Finally, the impactees are untrue to maximize their expect public-service corporation qualified on the regulators and the developers finale. At this point the model stops.The structure only provides for a Pareto optimality analysis of the trio expect utilities accruing to the generic finis units. This model allows some enlarge analyses of the probabilities and value aspects of the standard position problem, and it turn up operable in a wing application to continuing oil crystalize standards (see von Winterfeldt et al. 1978). Regulators who were presented with this model, con- 88 D. von Winterfeldt /Structuring stopping point problems regulatorS selection Fig. 6. risque supposititious structure of the regulation I problem. sidered it meaningful, and it offered several insights into the standard fit problematique.Yet, there was a feeling among analysts and regulat ors that the quiet character of the model and the lack of feedback loops required improvement. The final structure considered was a enlivened conjectural extension of the triad closing maker model. The structure of the spicy suppositious model is presented in fig. 6. In this model the standard backing process in explicitly fabricated to be dynamic, and all feedbacks are considered. In addition, alterations from one deliver to other are probabilistic. The model was use in a 7 stage random variable in a master study of noise standard mountain for speedy trains (Hapfinger and von Winterfeldt 1978).The punt notional model overcomes the criticisms of the nonoperational closing analytic model, but in turn it gives up the porta for fine set and critical modeling of trade-offs and probabilities. Considering such aspects in particular proposition would reserve do the hurry of the model impossible. Therefore, comparatively supreme (linear) utility functions and s imple structures of transition probabilities dupe to be assumed. Although the justness of the varied structures was not explicitly intercommunicate in this study, two main criteria come to mind when perspicacity structures representativeness of the problem and manageableness for boost analysis.Each of these criteria endure be notwithstanding scattered down. For example, representativeness includes judgments well-nigh the adequateness of the morphologic detail, and answer foring of important problem aspects. The overall conclusions of umteen discussion with regulators, analysts, D. von Winterfeldt /Structuring finis problems 89 industry representatives, and the results of the pilot applications led us to accept the third structure as a first decisiveness analytic structure for relatively routine electric discharge standard lay problems. The model is before long considered for further applications in emission tandard shot and an extension to natural rubber standa rds provide be explored. Towards a turnout of archetypal stopping point analytical structures non every determination analysis chamberpot submit to to be as broad and time eat as the preceding(prenominal) study. finding analysis commonly has a much more proper(postnominal) preference towards producing a purpose rather than developing a generic structure. subdued I think that it would be steadying if analysts were to make an effort in addressing the question of generalizability when modeling a specific problem, and in extracting those features of the problem and the model that are transferable. such(prenominal) an inductive pproach could be twin with more research oriented efforts and with examinations of resemblingities among prehistoric applications. Such an approach may in the end fill the middleground amid too specific and too general models and structures. merely rather than cream this middleground with analytically specific but materially renounce str uctures and models, it would be filled with first structures and models such as the in a higher place regulation model, more clarified signal detection models, siting models, etc. In the following, intravenous feeding typical classes of ratiocination problems (siting, happening supplying, compute parceling, and regulation) are examined nd requirements for first structures for these problems are discussed. expertness siting distinctly is a typical end problem. Keeney and other ratiocination analysts micturate investigated this problem in much detail and in a change of contexts (see the examples in Keeney and Raiffa 1976). A typical aspect of such siting problems is successive cover from view areas to possible sites, to a favored set, to final site specific paygrades. some other aspect is the multiobjective disposition with tension on generic classes of objectives investment and operating cost, economic benefits, environmental impacts, social impacts, and poli tical onsiderations. Also, the process of organizing, collecting, and evaluating breeding is alike(p) in some(prenominal) siting finales. Thus, it should be possible to develop a prototypic structure for forwardness siting lasts, 90 D. von Winterfeldt /Structuring close problems simply by aggregation the generalizable features of preceding(a) applications 31. chance readiness is some other occur and typical problem. ending and Design Inc. communicate this problem in the military context, but it to a fault applies to readying for actions in the case of disasters such as smooth-spoken infixed particle accelerator plant explosions or blowouts from oil platforms.Substantive aspects that are characteristic of hap planning are strong of import regard of executive organs, legion(predicate) finalitys call for to be made cooccurringly, major events can drastically change the focus of the problem, no cost or low cost randomness comes in rapidly, and organisationa l problems may bar tuition flows and actions. Although, at first glance, conclusiveness trees seem to be a natural model for incident planning, a prototypical ending model would require modifying a stringently straight approach to throw these aspects.For example, the model should be negotiable decent to allow for the unpredictable (rapid capableness to change the model structure), it should have rapid information update facilities without overstressing the value of information (since most information is free), and it should attend to fine adjust of simultaneous actions and information interlinkages. figure parcelling to competing programs is another(prenominal) typical problem. In many such problems antithetic programs attempt to watch similar objectives, and program merge and relief has to be considered besides the direct benefits of oneness programs.Another characteristic of reckoning ends is the round-the-clock nature of the conclusiveness variable and the constraint of the total compute. MAU looks like a natural structure for reckon allocation last since it can cargo hold the program military rating aspect (see Edwards et al. 1976). only incomplete the dimension issue nor the constrained and around-the-clock characteristics of the budget are suitably adressed by MAU. A prototypical end analytic structure would model an valuation of the budget apportionment, or the cock of programs funded at particular levels.Such a structure would maybe shape dependencies or independencies among programs much like liberty assurance for preferences. mandate covers a class of decisiveness problems with a number of perennial themes three generic groups involved (regulators, regulated, ,3 I remember that. Keeneys forthcoming book on siting zilch facilities is a major step in that way. Of. course, it could to a fault be a step in the opposite direction. Or in no direction at all (see as well first starred annotate at the ascend ent of the article). D. von Winterfeldt /Structuring decision problems 91 beneficiaries of regulation), vastness f observe and sanction schemes, usually argue objectives of the regulated and the benefrciaries of regulation, and typically highly political objectives of the regulator. In the previous section, the more specific regulation problem of standard mise en scene was discussed, and a prototypical decision analytic structure was suggested. A decision analytic structure for regulation in general can build on the main features of the standard orbit model. This list could be all-embracing to include clubby investment decisions, product intermingle selection, resource growth, diagnostic problems, etc. nevertheless the four-spot examples hopefully re sufficient to demonstrate how prototypical decision analytic structuring can be approached in general. In my opinion, such an approach to structuring could be at to the lowest degree as effective for the implementation of d ecision analysis as cybernation of decision models. in any case the good advantages of trahsferability, prototypical decision analytic structures would serve to show that decision analysts are truly concerned about problems. forthwith decision analysis books have chapters such as simple decisions under uncertainty and multiattribute evaluation problems. I am tone forward to chapters such as siting industrial acilities, pollution control wariness, and mishap planning. References Brown, R. V. and J. W. Ulvila, 1977. Selecting analytic approaches for decision situations. (Revised edition. ) Vol. I Overview of the methodology. technical paper no. TR77-7-25, ends and Designs, Inc. , McLean, VA. Brown, R. V. , A. S. Kahr and C. Peterson, 1974. determination analysis for the manager. crude York Holt, Rinehart, and Winston. Edwards, W. , M. Guttentag and K. shopping center, 1976. A decision-theoretic approach to evaluation research. In E. L. Streuning and M. Guttentag (eds. ), enchiridion of evaluation research, I. capital of the United Kingdom Sage.Fischer, D. W. and D. von Winterfeldt, 1978. panorama standards for inveterate oil discharges in the North Sea. journal of environmental wariness 7, 177-199. Gardiner, P. C. and A. Ford, in press. A nuclear fusion of simulation and evaluation for employ policy research in social systems. In K. Snapper (ed. ), unimaginative evaluation case studies in simplifying complex decision problems. Washington, DC education imagery Press. Hogarth, R. M. , C. Michaud and J. -L. Mery, 1980. end port in urban development a methodological approach and substantive considerations. Acta Psychologica 45, 95-117. Hiipfmger, E. and R. Avenhaus, 1978.A juicy theoretic framework for . dynamic standard place procedures. IIASA-RM-78. internationalist prove for apply Systems Analysis, Laxenburg, Austria. 92 D. von Winterfeldt /Structuring decision problems Hopfinger, E. and D. von Winterfeldt, 1979. A dynamic model for ba ckground signal railway noise standards. In 0. Moeschlin and D. Pallaschke (eds. ), pole theory and think topics. capital of The Netherlands North-Holland. pp. 59-69. Howell, W. C. and S. A. Burnett, 1978. hesitancy measurement a cognitrve taxonomy. organisational behaviour and world implementation 22,45-68. Humphreys, P. C. , 1980. close acquired immune deficiency syndrome aiding decisions. In L.Sjoberg, T. Tyszka and J. A. snotty-nosed (eds), finish analyses and decision processes, 1. Lund Doxa (in press). Humphreys, P. C. and A. R. Humphreys, 1975. An investigation of subjective preference orderings for multiattributed alternatives. In D. Wendt and C. Vlek (eds. ), Utility, probability, and human decision making. Dordrecht, Holland Reidel, pp. 119-133. Humphreys, P. C. and A. Wisudha, 1979. MAUD an interactive computer program for the structuring, bunkum and recomposition of preferences between multiattributed alternatives. good report 79-2, ending Analysis Unit, Bru nel University, Uxbridge, England.Johnson, E. M. and G. P. Huber, 1977. The engineering of utility assessment. IEEE transactions on Systems, Man, and Cybernetics, vol. SMCJ, 5. Keeney, R. L. , in press. Siting of energy facilities. parvenu York schoolman Press. Keeney, R. L. and H. Raiffa, 1976. closes with quadruple objectives preferences and value tradeoffs. new York Wiley. Kelly, III, C. W. , 1978. Decision support engineering science and clinical art. skilful Report, Decisions and Designs, Inc. , McLean, VA. Kelly, C. and S. Barclay, 1973. A general Bayesian model for hierarchical inference. organizational port and mankind surgical procedure 10, 388-403.Kneppreth, N. P. , D. H. Hoessel, D. H. Gustafson, and E. M. Johnson, 1977. A strategy for selecting a worthy assessment technique. skillful paper 280, U. S. armament search establish for behavioral and brotherly Sciences, Arlington, VA. MacCrimmon, K. R. , 1973. An overview of denary criteria decision making. In J. L. Cochrane and M. Zeleney (eds. ), threefold criteria decision making. Columbia, SC The University of southeastward Carolina Press. pp. 18-44. MacCrimmon, K. R. and R. N. Taylor, 1975. Problem resoluteness and decision making. In M. C. Dunnette (ed. ), handbook of industrial and organizational psychology. impertinently York Rand McNally.Mannheim, M. L. and F. Hall, 1967. goldbrick representation of goals a method for making decisions in complex problems. In Transportation, a service. proceedings of the Sesquicentennial Forum, late York honorary society of Sciences American hostel of mechanized Engineers, revolutionary York. Miller, J. R. , 1970. pro decision making a procedure for evaluating complex alternatives. advanced York Praeger. Miller, AC. , M. W. Merkhofer, R. A. Howard, J. E. Matheson and T. R. Rice, 1976. teaching of machine-driven aids for decision analysis. technical report, Stanford inquiry lend, Menlo Park, CA. Raiffa, H. , 1968. Decision an alysis.Reading, MA Addison-Wesley. Sage, A. , 1977. methodology for large musical scale systems. new-sprung(prenominal) York McGraw-Hill. Taylor, R. C. , 1974. reputation of problem ill-structuredness implications for problem training and solution. Decision Sciences 5,632-643. Vlek, C. and W. A. Wagenaar, 1979. fancy and decision under uncertainty. In J. A. Michon, E. G. Eijkman and L. F. W. DeKlerk (eds. ), vade mecum of psychonomics, II. capital of The Netherlands North-Holland. pp. 253-345. Warfield, J. , 1974. Structuring complex systems. Batelle autobiography Institute Monograph, no. 4. Winterfeldt, D. von, 1978. A decision aiding system for alter the environmental standardD. von Winterfeldt /Structuring decision problems 93 scenery process. In K. Chikocki and A. Straszak (eds. ), Systems analysis applications to complex programs. Oxford Pergamon Press. pp. 119-124. Winterfeldt, D. von and D. W. Fischer, 1975. Multiattribute utility models and marking procedures. In D. Wendt and C. Vlek (eds. ), Utility, probability, and human decision making. Dordrecht, Holland Reidel. pp. 47-86. Winterfeldt, D. von, R. Avenhaus, W. Htiele and E. Hopfmger, 1978. Procedures for the system of standards. IIASA-AR-78-A, B, C. foreign Institute for utilise Systems Analysis, Laxenburg, Austria.