Preface

A useful answer can be smaller than the question that first brought people together. “Which pump should we buy?” may first need a distinction between a pump’s rated capacity and whether assistance can actually reach the affected people. “What should we develop next?” may first need a worthwhile future contribution rather than a training plan or a model upgrade. Problem Structuring and Decision Support helps people make those differences usable without claiming that one discussion, model or recommendation settles the whole situation.

The framework serves practitioners and assisting agents who formulate questions, conduct inquiry, compare alternatives and prepare qualified advice. They may be internal analysts, facilitators working with affected groups, or professionals advising another person or organization. A team can investigate a question for its own later decision while keeping the analysis and choice distinct. The development-direction advising profile adds a distinct recommending performer where that service relationship is the question.

Problem Structuring and Decision Support also supports the maintenance of the Methods and working arrangements that make this work dependable. Its pattern language offers a repertoire for recurring difficulties. The Parts group those difficulties; pattern numbers are addresses, and an actual inquiry selects contributions by the results it needs.

PSD.Preface:1 - Start from the missing result

A pattern describes a recurring difficulty and a useful way to respond. A Method is a reusable way of doing the work; a workshop appointment, diagram or software package is not that way merely because it carries a familiar name. Open the pattern that can change the current answer. If an adequate result already exists, use it within its conditions instead of restarting the inquiry.

Some questions are contested. People can disagree about what is wrong, who bears a consequence, what counts as useful service or which evidence should matter. Preserve differences that change action. Making them explicit does not require every position to have equal evidential support, and a shared sentence does not establish unanimous endorsement.

Other questions are already sufficiently formed. They can enter directly at a missing alternative, specialist premise, comparison or recommendation. A person exploring an opportunity can stop even earlier, without asking another performer for advice.

PSD.Preface:2 - How the results connect

Problem structuring makes a difficult situation investigable. It can change which question is worth answering, whose consequences are considered, what counts as the problem and which interventions become plausible. Decision support uses suitably qualified formulations, alternatives, value accounts and evidence to help a receiving decision. A problem formulation can be a complete useful return before a recommendation is possible; a well-formed question can also enter comparison directly.

PSD.1 establishes the receiving use when that use is unclear. PSD.2 supplies attributed concerns and participation limits, while PSD.3 shows how rival formulations change evidence needs, boundaries and possible responses. PSD.4 uses the needed engagement and participation results directly, alongside the formulations, to choose a usable inquiry boundary. A formulation alone cannot establish who receives the result or whether an affected person’s concern has been represented adequately.

Within that boundary, a model can answer one question and expose another. PSD.5 makes the contributions and limits of complementary models usable. PSD.6 compares the ways of obtaining the needed results and qualifies any combination. PSD.7 helps participants examine the developing meanings and return shared and contested claims. A corrected concern can therefore change a model or reopen the boundary while independent analysis continues. PSD.16 becomes useful when the actual interaction between these activities threatens their results.

Formulations and models inform the alternative construction in PSD.8. A consequential difference may then need a value account from PSD.9, an uncertainty account from PSD.10, a consequence comparison from PSD.11, or the conditions under which a comparison holds or reverses from PSD.12. Obtain only the result that is missing. For example, a sensitivity result can expose a missing road-access premise without requiring another round of alternative generation. A new service concern that the option set cannot address gives a reason to revisit that set or its formulation.

PSD.13 connects the qualified inputs to what the recipient can use now: a direction, a retained set, a conditional comparison, a probe, a request or an abstention with its reason. PSD.14 connects a material later observation to the advice it can change when follow-up is needed. Neither relation requires performing all preceding patterns. The full flood-pump application shows the connections, including a useful first engagement question and a later whole-investment blocker.

PSD.Preface:3 - Keep unlike contributions distinct

A concern map can preserve what different participants mean. A hydraulic model can estimate a stated physical consequence. A value account can explain which differences matter to the receiving decision. These accounts can inform one another, but none becomes the represented situation or acquires the other’s evidential authority.

Likewise, human capability and transfer, organization arrangements, AI evaluation, legal conditions and safety are different questions. This framework helps specify the result needed from each practice and how its absence affects advice. It does not supply those practices’ conclusions. Qualify a source’s contribution for the subject and configuration of the advice; similar wording does not establish that its result applies elsewhere.

State what each result supports. A description may explain a proposed arrangement; a claim that its relations already hold needs evidence about that arrangement. A score may inform a choice, which remains with the authorized chooser. Advice for a particular client needs results qualified for that case; an available Method alone does not supply them.

PSD.Preface:4 - Qualify the whole combination for the receiving question

Use the questions in PSD.6, Combine contributions through explicit joins for the combination you actually propose. Each supplying result must be qualified for its own claim, fit the receiving subject, configuration, service meaning and horizon, and meet the conditions of joint use. State any conversion that needs additional work: a concern may suggest an objective, but its frequency of mention supplies no value weight; a scenario can condition analysis, but its presence supplies no probability.

Check shared conditions at the scope where they matter. The needed participants must have usable ways to contribute or correct attribution; specialist evidence must cover the relied-on claim; the Methods must be jointly possible with the available competent effort, data access and time. Include preparation, interpretation and coordination where they consume the same limited resources. Results from one dataset retain that shared evidence dependence. Three compatible pairs establish neither a feasible whole arrangement nor three independent confirmations.

In a constructed budgeting variant of APP-PSD-01, attributed concern recovery needs four team-hours, conditional technical analysis needs six, and preparation of the integrated return needs five. These estimates include their preparation and joining effort, require suitably competent staff, and draw on one twelve-hour budget before the board’s deadline. The hours cannot be shared or credited to two activities. Every pair needs at most eleven hours, yet the whole needs fifteen. The proposed combined return therefore needs three more hours, a supported arrangement with lower total effort, or a narrower answer or later delivery agreed within the engagement. Omitting a required contribution leaves that whole return unsupported. The individual inquiry or calculation can still be useful at its stated scope. These numbers illustrate a joint condition; a real engagement supplies its own estimates and resource rules.

Evidence effort follows the claim someone will rely on. A sketch may locate a disputed service meaning, and a conditional calculation may identify the assumption worth investigating. A recommendation that relies on physical performance, safety, rights or professional competence needs the relevant practice’s qualified result. PSD.5, Match assurance to the claim being used states that boundary. There is no framework-wide sample size, confidence score or completed-workshop count that can replace it. Ask what possible finding could change the bounded return; acquire evidence for that question and stop collecting when further work cannot improve that return enough to justify its burden. Retain a mandatory domain check whenever the intended reliance requires it.

The common practical tension is between obtaining a timely usable answer and preserving the distinctions that can defeat it. More participation can reveal a missing concern while consuming scarce attention. More models can widen the answer while adding interpretation work. A complete ranking can be easier to communicate while requiring judgements the evidence does not supply. State the accepted cost or narrower result, and return to the affected contribution when its conditions fail.

Constituent actions in ongoing work. During a decision-support conversation, distinguishing two meanings of a proposal can constitute part of constructing the problem frame, while that framing is part of supporting the pending decision. If the decision concerns allocation of a shared resource, treating each party’s local success as the shared criterion can defeat the whole comparison. The facilitator needs the intermediate ability to elicit and compare these meanings, in addition to subject knowledge and the ability to run the meeting. FPF B.1.5.EW makes that vertical recoverable. The PSD Methods structure the decision support; the participants, subject expertise, authority and resources needed by the resulting course of action must still be available.

PSD.Preface:5 - Complete a bounded answer in an open situation

Useful inquiry does not require certainty about everything. It requires knowing what the present answer covers, what it leaves open and what could change it. A partial comparison, a retained pair of alternatives, a narrowly qualified recommendation or an exact request can complete the current question.

Keep protected conditions outside a compensating score when their governing source requires that protection. If the requirement’s justification is in question, use PSD.9 §4.4 and C.11.DUA: compare its protective contribution with its full burden. State separately what change is justified, whether the requirement remains in force, and how it could be amended.

If plausible variation in the estimates or value judgements used by a comparison or recommendation could change its conclusion, test that variation at proportionate effort through PSD.12. Report whether and under what conditions the conclusion changes. When the response at a central input can hide a consequential nonlinear difference, PSD.12 section 4.3.1 supplies a small paired-response check. Its constructed equal weights are not actual event probabilities. State when a weight or probability is assumed. A missing premise should block only the claims that need it; independent content can remain useful.

A recommendation belongs to the advising activity. Prepare it for the recipient’s intended use; leave the later choice and any authorization to the actors entitled to make them. Plans, performed work and observed effects need their own basis. This separation protects the recipient’s agency and makes a later change intelligible: the new evidence may narrow advice without rewriting what was previously chosen or done.

PSD.Preface:6 - Architectural Rationale

The language joins problem structuring and decision support through their results because a technically sound answer can address an inadequate question, and an insightful reframing can leave the recipient with no usable comparison. Keeping the contributions individually addressable lets a practitioner repair the missing connection without repeating adequate work. It also preserves a direct inquiry result when no later recommendation is required.

These serious alternatives explain when the arrangement is worth its cost:

The scope also includes improving and continuing this practice. A better way to elicit concerns, a repair to simultaneous inquiry, and the successful transfer of that way to another team are different achievements. PSD.15–PSD.17 preserve their separate questions and evidence. They become relevant when the repertoire, interaction or continuation is itself at issue; an ordinary comparison need not become a professional-development programme.

PSD.Preface:6.1 - The development-direction profile and neighboring results

The development-direction advising profile is a bounded use of the shared PSD contributions. Its difference is consequential: advice concerns a developing subject whose capability, support configuration, horizon and professional relationship can change what is supportable. A sponsor may differ from the person whose private assessment matters; human transfer evidence differs from an organization’s arrangement comparison and from an AI configuration’s evaluation. The profile’s guidance on evidence about the developing subject, comparison guidance, worked uses, and recognition and assurance account supply those narrower answers. Its professional source account retains the actual service boundaries; they do not become entry requirements for every PSD inquiry.

The independently published Development Opportunity Construction and Development-Direction Advising framework offers directly addressed contributions when the missing result concerns development opportunities themselves. DOCA.3 constructs a worthwhile receiving problem and proposed contribution; DOCA.5 checks whether a development direction and its required supports can be achieved together; DOCA.7 bounds the distinct advising relationship. A qualified opportunity can enter PSD.8, and a bounded service can use the needed PSD comparison or recommendation. Use PSD.3 to examine a contested problem formulation. Use the contribution that is missing; a formed PSD question with adequate inputs needs no preceding DOCA traversal. The complete opportunity MethodDescription and advising profile included here retain their narrative uses.

The direct-practice boundary account identifies the result needed from human development, organization arrangements, technical evaluation and other practices. Such a result supplies a premise for this inquiry; the supplier retains its Method, evidence and conclusion. For example, an organization’s qualified comparison of provider and internal arrangements can support advice, while a publication describing how to make that comparison cannot settle the client’s case.

PSD therefore supplies a repertoire and explanations of its connected uses. A claim that a selected combination is one composite Method needs the whole action, identified part Methods, permitted variations and joins specified in PSD.6. A Part groups reading material; a profile narrows use; a result dependency states what one contribution needs from another. Those relations do different work. Reconsider the organization of the language when its field boundary, a profile’s independently useful result, or a recurring contribution changes.

PSD.Preface:6.2 - Why the source lines are combined

Problem-structuring methods give the framework a substantive answer before optimization begins: make rival interpretations and their practical consequences available for inquiry. Smith and Shaw’s 2019 characterization and Kogetsidis’s 2025 application review support a diverse continuing practice, with exploratory and published-application limits. The framing comparisons used in PSD.3 add the need to make contextual changes visible. PSD uses these contributions to construct materially different formulations and stop when they are sufficient for the inquiry or decision they serve. It rejects both a sponsor’s brief as an unquestioned final frame and endless collection of perspectives that changes no next action. The exact comparison and its source limits remain in PSD.3, SoTA-Echoing.

The older critical review of combining problem-structuring methods with multiple-criteria decision analysis (MCDA) by Marttunen, Lienert and Belton (2017) makes the interfaces between structuring and analysis part of the problem. The exploratory workshop comparison by Lami and Tavella (2019) provides a bounded counterexample to uniform workshop usefulness. PSD adapts their contributions into the smallest complementary model set and explicit Method joins in PSD.5–PSD.6. This costs more interpretation than applying one familiar technique. Retain the cheaper technique when it supplies the same result and conditions; the reviewed applications establish no universal winning combination.

The analytical line supplies ways to build alternatives, express values, compare consequences and identify useful uncertainty questions. The 2025 MCDA synthesis by Greco, Słowiński and Wallenius informs matching preference information to the comparison and recommendation form. Borgonovo and colleagues’ 2026 decision-analysis synthesis connects alternative construction, uncertainty, sensitivity and information acquisition. PSD.8–PSD.13 adopt these as selective contributions: a justified scalar model remains usable, incomplete preferences can leave a retained set, and one missing branch limits the whole recommendation. The cost of this arrangement is a less decisive-looking return when the premises warrant less decisiveness.

Under unsettled futures, Lempert and colleagues’ 2024 DMDU analysis supports decision-relevant inquiry despite low-confidence knowledge. The older 2019 DAPP account contributes pathway timing and failure conditions. PSD.10, PSD.12 and PSD.14 use those distinctions to expose holding regions, reversals and attainable follow-up. They leave probabilities, local thresholds and authority to their actual sources. A qualified probabilistic analysis remains preferable when its premises fit; calling uncertainty difficult is not a reason to discard it.

Professional sources add conditions on the return that a technically correct comparison may leave open. PSD.13 uses selected ICMCI engagement and conflict clauses within consulting and NCDA provisions within human career services. The profile source-responsibility account separately qualifies AI used by an adviser. These bounded normative contributions address competence, intelligible reliance and the service relationship. They supply neither a universal advising protocol nor effectiveness evidence for an intervention. The wider source map returns to the facilitation, follow-up, repertoire and cultural-continuation arguments in their own bodies.

Reconsider the affected arrangement when a simpler Method supplies the same useful return, a shared condition or join fails, a material perspective cannot be preserved, or better evidence changes a source contribution. A local case may need a different Method without defeating the whole language. Repeated failures of the same boundary or profile, or a recurring practical need that the existing contributions do not meet, are reasons to reopen the architectural choice.

PSD.Preface:7 - Whose conditions and evidence shape the answer

The sponsor’s first question, the participants who can attend, and the analyst’s familiar models all affect what is made visible. Record whose concerns are attributed, which affected people or consequences remain unrepresented, and how a material correction can reach the result. PSD.2 and PSD.7 provide that work. More participation does not make all claims equally supported, and a dissenting position has only the practical or institutional effect justified in the case. Domain rules determine a right, duty or protected condition.

The worked applications emphasize institutional investment and development advice. They are constructed teaching cases, and the source studies have their own populations, professional settings and access limits. A practitioner transferring the guidance must qualify local meanings, communication access, evidence and service conditions that can alter use. Method familiarity, fluent AI output, participant confidence and publication visibility can all be mistaken for support of a different claim. The relevant pattern’s evidence and bias questions help locate that claim; no single confidence measure qualifies the whole inquiry.

PSD.Preface:8 - Questions before relying on a combined return

The opening and direct entries help recognize the first useful action. Assurance asks whether the result is supported for its intended reliance. Use the checklists in the patterns actually selected, and reuse their qualified answers when the subject, content, receiving use and conditions remain the same. For the combination, answer the following questions:

• Can the recipient recover what this return answers, whose later decision or inquiry it serves, and what remains outside it? An early inquiry may end with a formulation or precise question.
• Do material concerns, competing frames and boundary choices survive into the models, alternatives and comparison that depend on them? A supported correction must reach the claim it changes.
• Does each relied-on result cover the needed subject, configuration, meaning and horizon, with its actual evidence limits? Have shared assumptions or sources been counted only for the support they supply?
• Do the selected Methods meet the joint conditions, including total effort and required interpretation? Feasibility of each pair does not establish that all the required work can be done together.
• Does the scope of the recommendation match the alternatives and conditions actually examined? In APP-PSD-01, two pump options can be compared conditionally while the whole investment question remains open. In APP-PSD-02, a qualified provider exclusion can close that branch, while an unexamined provider cannot simply disappear.
• Is it clear what information to request, when the inquiry can finish, and which advice must be reconsidered if a premise changes? Does the authorized decision-maker retain the later choice or action? The development profile also needs its stated service conditions and evidence about the developing subject; those qualifications are not inherited merely from generic PSD use.

These questions address failures shown in the applications: one service phrase hiding unlike concerns, an unjustified conversion of qualitative meaning into a score, and a partial comparison presented as a whole recommendation. Correct the particular meaning, obtain the missing contribution or narrow the return. Answer the questions in the working account the recipient needs; actual performance and later effects require their own evidence.

PSD.Preface:9 - Let practice change without making every case start over

A corrected concern, altered access condition or new model configuration can change the current result. Identify the claims that depend on the change and ask the responsible practitioner to reconsider them. Obtain and interpret the relevant observation before choosing a response; the mere arrival of a notice is not evidence that the new configuration is safe or effective.

A different problem concerns how inquiry itself is conducted. Use PSD.15 when the professional repertoire needs a justified change, PSD.16 when simultaneous inquiry activities interfere, and PSD.17 when a practice must be continued or changed across practitioners. These are conditional uses, not a required tail on every recommendation.

Contemporary source comparisons belong with the pattern claims they support. Earlier traditions remain valuable where their limits are understood. Neither a fashionable school nor an old success is enough to establish the present Method’s fit. What changes in practice is the reader’s ability to return a useful answer with visible limits—and to change that answer, or the way of obtaining it, when the actual basis changes.

PSD.Preface:End