Faire Un Choix Adéquat (FUCA)

• FUCA is a family of decision frameworks that range from rank aggregation MCDM methods to reforms in preference expressiveness for applications like admissions matching.
• The rank-aggregation MCDM approach converts raw criteria into ordinal ranks with weighted sums, offering simplicity, no need for normalization, and resilience to extreme values.
• In reform contexts such as the APB admissions system, FUCA enables the use of partial preorders to reduce random decisions and foster transparent, fair candidate matching.

Searching arXiv for the cited FUCA-related sources to ground the article. “Faire Un Choix Adéquat” (FUCA, more commonly spelled with accents as “Faire Un Choix Adéquat”) denotes a family of decision-oriented formulations rather than a single invariant concept across the arXiv literature. In the most explicit methodological sense, FUCA is an aggregation-type multi-criteria decision-making (MCDM) method based on rank aggregation, proposed by Fernando et al. (2011), in which alternatives are ranked criterion by criterion and then combined by a weighted sum of ranks (Wang et al., 8 Sep 2025). In a distinct but related institutional sense, FUCA names a proposed reform principle for the French Admission Post-Baccalaureat (APB) platform, where richer preference representations—specifically partial preorders rather than a forced strict total order—are used to reduce random selection and improve the expressiveness and transparency of admissions matching (Hamadi et al., 2017). In other papers, the phrase functions more loosely as a decision-support ideal: making a choice that is evidence-based, context-appropriate, collectively justified, or constructively legitimate, depending on the domain.

1. Terminological scope and principal meanings

The literature does not present FUCA as a single, universally standardized framework. Instead, the phrase appears in at least three levels of specificity.

First, FUCA is an explicitly named MCDM method. In the chapter on aggregation-type methods, FUCA is defined as a rank-based procedure that transforms an alternatives-criteria matrix into a final ranking by aggregating weighted ordinal positions. Its decision rule is to select the alternative with the smallest total weighted rank, and the chapter emphasizes that FUCA does not require normalization of criterion values (Wang et al., 8 Sep 2025).

Second, FUCA appears as a policy and preference-reasoning principle in admissions matching. In the APB study, the term is not introduced as a separate legal rule or national law, but as a reform idea: the admissions platform should allow applicants to state preferences more faithfully, and the platform should exploit that additional structure to make outcomes less dependent on randomness (Hamadi et al., 2017).

Third, several papers use the underlying idea of an “adequate choice” more generically. In software selection, it refers to a collective and operational choice among SEO analytics tools through brainstorming, committee evaluation, and a qualitative matrix (0905.4433). In student specialization, it refers to a concentration choice aligned with academic performance via Fuzzy Mamdani + C4.5 decision support (Harsiti et al., 2015). In high-speed machining, it refers to choosing a machining strategy that is not only geometrically feasible but dynamically compatible with the machine/NC pair (0904.1084). In multi-objective Bayesian optimisation from choice data, adequacy is expressed through choice functions and latent Pareto structure rather than full rankings or direct objective values (Benavoli et al., 2021).

A concise way to organize these usages is the following.

Usage of FUCA	Domain	Core meaning
Formal method	MCDM	Weighted aggregation of criterion-wise ranks
Reform principle	APB admissions	Richer preference expression to reduce random sorting
Generic decision-support ideal	Software, education, machining, choice learning	Context-appropriate, evidence-based, or structured choice

This distribution suggests that FUCA is best understood as a decision-adequacy motif whose most formal incarnation is the rank-aggregation MCDM method, while other uses preserve the semantic core of “making an adequate choice” but instantiate it with different technical apparatus.

2. FUCA as an aggregation-type MCDM method

In the MCDM chapter, FUCA is classified alongside SAW, MEW, AHP, ANP, COPRAS, MOORA, and WASPAS as one of eight aggregation-type methods (Wang et al., 8 Sep 2025). Its distinguishing feature is that it operates on ranks rather than normalized raw values.

For an alternatives-criteria matrix with $m$ alternatives and $n$ criteria, the inputs are criterion values $f_{ij}$ and criterion weights $w_j$, with

$\sum_{j=1}^{n} w_j = 1.$

Each alternative $i$ receives a rank $r_{ij}$ under each criterion $j$. Rank $1$ denotes the best alternative for that criterion, and rank $m$ the worst. For benefit criteria, larger values receive better ranks; for cost criteria, smaller values receive better ranks. The weighted aggregated rank of alternative $n$0 is then

$n$1

and the decision rule is to rank alternatives in ascending order of $n$2, so that smaller $n$3 indicates a better alternative (Wang et al., 8 Sep 2025).

The ranking logic is therefore purely ordinal. FUCA first discards magnitude information by replacing criterion values with criterion-wise orderings; it then restores inter-criterion importance through the weights $n$4. This produces a procedure that is computationally simple and conceptually transparent. The chapter explicitly lists three strengths: FUCA is easy to understand and implement, requires no normalization, and is less sensitive to extreme values because very large or very small raw values do not dominate once converted to ranks (Wang et al., 8 Sep 2025).

The same source also records the method’s limitations. FUCA loses magnitude information, so a tiny and a huge difference are treated identically if they induce the same rank order. Tie-breaking may be needed when alternatives share the same value on a criterion, and rank reversal can still occur if alternatives are added or removed (Wang et al., 8 Sep 2025). These limitations are not accidental; they follow directly from FUCA’s ordinal design. A plausible implication is that FUCA is most appropriate when the decision-maker trusts ordinal preference information more than inter-alternative distance information.

3. Worked example and comparative position among aggregation methods

The MCDM chapter illustrates FUCA on a common alternatives-criteria matrix with four alternatives $n$5–$n$6, three criteria $n$7, $n$8, $n$9, and weights

$f_{ij}$0

The criterion-wise ranks reported in the chapter are effectively:

Alternative	$f_{ij}$1	$f_{ij}$2	$f_{ij}$3
A1	1	1	4
A2	4	2	3
A3	3	3	2
A4	2	4	1

Using the weighted rank sum, the chapter reports the final FUCA scores

$f_{ij}$4

with the explicit calculation

$f_{ij}$5

The resulting ranking is

$f_{ij}$6

so A4 is selected as the top-ranked alternative (Wang et al., 8 Sep 2025).

The comparative significance of this example lies in its divergence from the value-based methods discussed in the same chapter. The summary states that SAW, MEW, COPRAS, MOORA, and WASPAS all ranked A3 as the top alternative, whereas FUCA ranked A4 first (Wang et al., 8 Sep 2025). This contrast is analytically important because it isolates the consequence of replacing normalized cardinal information with weighted ordinal ranks. FUCA rewards consistently favorable rank positions even when another method, using value magnitudes, prefers a different compromise.

This comparison also guards against a common misconception: FUCA is not merely a simpler implementation of SAW or a normalization-free restatement of standard weighted scoring. It is a substantively different decision rule. The chapter explicitly cautions that rankings cannot be generalized across methods and recommends trying multiple MCDM methods for a given application (Wang et al., 8 Sep 2025). In that sense, FUCA occupies a distinctive position within aggregation-type MCDM: not normalization-variant, but representation-variant.

4. Preference expressiveness and the APB reform principle

In the APB study, FUCA denotes a different but structurally related idea: the platform should support more faithful preference expression and use that structure to reduce random selection in admissions matching (Hamadi et al., 2017). The target problem is the French Admission Post-Baccalaureat Platform (APB), which allocates hundreds of thousands of first-year applicants to universities through a matching procedure.

The paper describes APB in three steps: applicants submit a ranked list of studies; institutions build preferences over applicants; APB allocates proposals in rounds. For selective studies, institutions already provide their own rankings. For non-selective limited-capacity studies, however, APB constructs institution preferences from applicant preferences. The original procedure groups applicants according to whether they are local to the academy, their relative rank of the study within a family of studies, and their absolute rank in the applicant’s list, before applying random sorting within each equivalence class (Hamadi et al., 2017).

The paper gives the corresponding construction: $f_{ij}$7 followed by

$f_{ij}$8

This is the point at which FUCA intervenes. The proposal is to move from forcing applicants into a strict total order toward allowing partial preorders with ties and incomparabilities. Formally, preferences over a finite set $f_{ij}$9 are expressed by a binary relation $w_j$0, with $w_j$1 for strict preference, $w_j$2 for indifference, and $w_j$3 for incomparability. A total preorder is reflexive, complete, and transitive; a partial preorder is reflexive and transitive but not necessarily complete (Hamadi et al., 2017).

The operational claim is that applicants who provide more specific preferences should receive priority over those who provide less specific information. To use partial preferences in matching, the paper extends them to a total preorder using optimistic reasoning rather than pessimistic reasoning. The total preorder is represented as an ordered partition

$w_j$4

where each $w_j$5 is an equivalence class of equally preferred objects (Hamadi et al., 2017).

The revised institution-preference construction introduces a disjunction size parameter: $w_j$6 Applicants are then sorted by relative rank $w_j$7, absolute rank $w_j$8, and disjunction size $w_j$9, with more specific statements considered first; randomness remains only as a last-resort tie-breaker via

$\sum_{j=1}^{n} w_j = 1.$0

The matching engine itself remains Gale–Shapley stable marriage, specifically the institution-proposing version used by APB (Hamadi et al., 2017).

A central example contrasts two candidates, $\sum_{j=1}^{n} w_j = 1.$1 and $\sum_{j=1}^{n} w_j = 1.$2. Under standard APB, both may be forced into the same strict list $\sum_{j=1}^{n} w_j = 1.$3, making them indistinguishable and exposing the allocation to randomization. Under the revised scheme, $\sum_{j=1}^{n} w_j = 1.$4 can instead submit

$\sum_{j=1}^{n} w_j = 1.$5

which indicates that $\sum_{j=1}^{n} w_j = 1.$6 and $\sum_{j=1}^{n} w_j = 1.$7 are equally top-level choices. The paper argues that $\sum_{j=1}^{n} w_j = 1.$8, who has only one top wish, should receive priority because she is more specific. In the example, the stable matching allocates $\sum_{j=1}^{n} w_j = 1.$9 to $i$0 and $i$1 to $i$2, eliminating random choice (Hamadi et al., 2017).

The paper claims several benefits for this FUCA-style reform: better expressiveness of applicant preferences, reduction of random decisions, more fairness, more transparency, and a better fit with AI preference theory (Hamadi et al., 2017). Here, adequacy does not mean weighted rank aggregation; it means a better alignment between stated preferences, matching semantics, and institutional legitimacy.

5. Domain-specific operationalizations of “adequate choice”

Beyond MCDM and APB, the phrase “Faire Un Choix Adéquat” functions as a broader decision-support theme in several technical domains.

In SEO software selection, the proposed method is explicitly pragmatic and collective. The decision process comprises: brainstorming, identification collective des critères, formation of a comité de pilotage, search for candidate software through a query in the leading search engine, retention of the 6 premiers résultats with an evaluation version, tests by the committee, evaluation through a matrice de pondération, and synthesis via tables and graphiques radar (0905.4433). The organizing acronym is CITIE: Comparatif, Interopérabilité, Tarif, Innovation, Ergonomie. Each software–criterion cell receives an evaluation de 1 à 3. The paper presents this not as a mathematically heavy optimization scheme, but as a way to structure a collective decision rapidly in a PME context (0905.4433).

In specialization choice for Informatics Engineering students, adequacy is defined as consistency between student interests and academic credentials. The proposed hybrid model combines C4.5 for rule extraction and Mamdani fuzzy inference for suitability evaluation (Harsiti et al., 2015). The dataset contains 126 students and 13 variables total, comprising 8 course grades, 4 GPA-related data values, and 1 specialization/classification variable. The system has 12 input variables and 1 output variable, CONCENTRATION, with fuzzy sets Multimedia, Web, and Programming. C4.5 uses

$i$3

and

$i$4

The study reports that 16 rules were produced from Weka, that 17 of the 126 students had chosen a specialization not matching the recommendation, and an overall suitability analysis accuracy of

$i$5

with mismatch rate

$i$6

FUCA here denotes a specialization choice aligned with demonstrated competence rather than preference alone (Harsiti et al., 2015).

In high-speed machining of aeronautical pockets, the adequate choice concerns machining strategy rather than ranking alternatives in an abstract decision matrix. The paper states that the “best” strategy is the one minimizing real machining time, not merely CAM-estimated time, and that this requires both geometric feasibility and dynamic efficiency (0904.1084). The tool-selection process introduces the diameters $i$7 and $i$8, uses dichotomy with

$i$9

computes machining time as

$r_{ij}$0

and employs a 5% improvement threshold in material removal rate for pruning intervals. The paper’s conclusion is that choosing an optimal strategy requires both geometric and dynamic analysis; geometry alone is insufficient (0904.1084).

These three examples share a family resemblance. In each, adequacy is achieved by enriching the representation of the decision problem: user-defined criteria and committee structure in software selection, historical academic patterns and fuzzy rules in specialization advice, and machine/NC dynamics in pocket machining.

6. Preference, uncertainty, and constructive choice as broader FUCA motifs

Two additional strands show how the “adequate choice” idea extends beyond explicit use of the acronym.

In choice-functions based multi-objective Bayesian optimisation, the relevant notion of adequacy is that a decision-maker need not provide full rankings or exact objective values. Instead, one observes a choice function

$r_{ij}$1

where $r_{ij}$2 is the set of all finite subsets of the input space and $r_{ij}$3 is the chosen subset from a query set $r_{ij}$4 (Benavoli et al., 2021). The paper assumes a latent vector-valued embedding

$r_{ij}$5

and interprets observed choices as Pareto non-dominated sets in that latent space. Pareto dominance is defined by

$r_{ij}$6

A Gaussian-process prior is placed on each latent component, and the model uses ADVI for kernel hyperparameters and elliptical slice sampling for posterior inference (Benavoli et al., 2021). This framework suggests that an adequate choice can be expressed through a compact rejection/retention judgment, provided the model can infer the latent multi-objective structure.

In effort-estimation model selection, the decision problem is explicitly multi-criteria and uncertain. The paper compares AHP and Fuzzy AHP for choosing among Expert Judgment, COCOMO, and a Fuzzy Neural Network / non-algorithmic model under four criteria: Reliability, MMRE, Pred(x/100), and Uncertainty (Sehra et al., 2013). Fuzzy AHP uses triangular fuzzy numbers

$r_{ij}$7

with membership function

$r_{ij}$8

and allows optimistic, moderate, or pessimistic attitudes (Sehra et al., 2013). The reported FAHP criterion weights are

$r_{ij}$9

and the final FAHP alternative weights are 0.2918, 0.3580, and 0.3434, making COCOMO / algorithmic model the best choice, whereas classical AHP instead favored the non-algorithmic model (Sehra et al., 2013). This is a direct instance where adequacy depends not only on alternatives and criteria, but also on the representation of uncertainty.

A final conceptual extension appears in constructive real algebra. The memoir on geometric theories for real algebra is not a paper that explicitly defines FUCA; however, it is relevant if adequacy is understood as making only those choices justified in a constructive proof setting (Lombardi et al., 2024). The work aims to build “une théorie dynamique aussi complète que possible” for real algebra sans axiome du choix dépendant, and it treats skolemization as acceptable when the existential witness is prouvablement unique (Lombardi et al., 2024). It introduces function symbols such as virtual roots $j$0 and a fraction-like function $j$1, with axioms including

$j$2

and, for $j$3,

$j$4

The conceptual link is not terminological but structural: an adequate choice is one that is canonical, conservative, and does not rely on non-constructive choice principles (Lombardi et al., 2024).

7. Synthesis, distinctions, and recurrent misconceptions

Across these sources, FUCA should not be treated as a single monolithic theory. The most common misconception is to assume that “FUCA” always denotes the rank-aggregation MCDM method. That is incorrect. The APB paper uses the phrase for a preference-expressiveness reform in admissions matching rather than for a weighted rank-sum method (Hamadi et al., 2017). Conversely, the MCDM chapter presents FUCA as a clearly delimited ordinal aggregation algorithm with formula

$j$5

which is a different object altogether (Wang et al., 8 Sep 2025).

A second misconception is to read all “adequate choice” formulations as equivalent. The evidence does not support that. In the SEO-software paper, adequacy is grounded in collective elicitation and qualitative scoring; in the specialization paper, it is grounded in historical academic patterns plus fuzzy inference; in machining, it is grounded in geometric and dynamic compatibility; in choice-learning, it is grounded in latent Pareto consistency (0905.4433, Harsiti et al., 2015, 0904.1084, Benavoli et al., 2021).

What does unify these disparate uses is a recurring technical intuition: a choice becomes more adequate when the decision procedure better represents the relevant structure of the problem. In the MCDM version, the relevant structure is ordinal ranking with criterion weights. In APB, it is the distinction between strict preference, indifference, and incomparability. In fuzzy and Bayesian decision-support settings, it is uncertainty, latent criteria, or non-dominance relations. In constructive mathematics, it is canonical witness construction rather than arbitrary selection.

This suggests a broad encyclopedic characterization. FUCA is not a single doctrine but a family of adequacy-oriented decision frameworks. Its most explicit formalization is a rank-aggregation MCDM method (Wang et al., 8 Sep 2025); its most developed socio-technical application in the supplied literature is a preference-theoretic reform of APB aimed at reducing random admissions decisions (Hamadi et al., 2017); and its wider use signals a consistent concern with replacing arbitrary, opaque, or overly rigid choice mechanisms by representations that are more expressive, more structured, or more faithful to the decision context.