XMCDA v4.0.0 Schema - Copyright (c) Decision Deck Consortium, 2020-2021
by Sébastien BIGARET and Patrick MEYER (IMT Atlantique)
Based on the previous works:
- XMCDA 3.x (2013-2019) by Sébastien BIGARET and Patrick MEYER (IMT Atlantique)
- XMCDA 2.x (2009-2012) by Raymond BISDORFF (University of Luxembourg), Patrick MEYER (Telecom Bretagne), Thomas Veneziano (University of Luxembourg), who formalised the collective work of the Decision Deck Consortium.
Root of the XMCDA tree.
Group of attributes used for all the main tags.
The id attribute allows to identify and reference the underlying piece of data.
The name attribute contains the human readable name of the object or concept.
The mcdaConcept attribute allows to specify to what mcda concept a tag is related. It is used by an algorithm to make choices which will have an influence on the output. The documentation of the program should therefore specify, if necessary, what mcdaConcept should be used for the input data. In particular, if an algorithm requires, among other things, twice the same input tag, they can be differenciated by the mcdaConcept (this is even mandatory if they are present in the same file, but should be optional if the two tags can be in different input files, or originate from two different programs). The algorithm should therefore not be too strict on these mcdaConcepts, as this will reduce the compatibility between the various programs.
Group of attributes used for all the main tags. The id attribute is mandatory.
The mandatory id attribute allows to identify and reference the underlying piece of data.
The name attribute contains the human readable name of the object or concept.
The mcdaConcept attribute allows to specify to what mcda concept a tag is related. It is used by an algorithm to make choices which will have an influence on the output. The documentation of the program should therefore specify, if necessary, what mcdaConcept should be used for the input data. In particular, if an algorithm requires, among other things, twice the same input tag, they can be differenciated by the mcdaConcept (this is even mandatory if they are present in the same file, but should be optional if the two tags can be in different input files, or originate from two different programs). The algorithm should therefore not be too strict on these mcdaConcepts, as this will reduce the compatibility between the various programs.
Description of most of the tags. The names of the elements suggest their standard use.
Bibliographical entries
The set of alternatives.
Definition of an alternative. The id attribute is mandatory.
Type Indicates whether an alternative is real or fictive. By default, if the tag is not present, an alternative is considered as real. Fictive alternatives are those which might not exist in reality, but can be used for some preference elicitation means.
Active indicates whether an alternative is active or not in the calculations. By default, if the tag is not present, an alternative is considered as active (true).
Definition of sets of alternatives.
Definition of a set of alternatives. Each element can be valued (to represent valued or ordered sets).
Values related to alternatives. Can for example be used to store overall values of alternatives, or their rank in a ranking.
Values related to sets of alternatives. Can for example be used to store overall values of sets of alternatives, or their rank in a ranking.
One or more values related to an alternative. Can for example be used to store the overall value of an alternative, or its rank.
One or more values related to a set of alternatives. Can for example be used to store the overall value of a set of alternatives alternative, or its rank.
Matrix of values related to alternatives.
Row of a matrix.
Matrix of values related to sets of alternatives.
Row of a matrix.
Linear constraints on the alternatives.
A linear constraint related to alternatives. It is composed of a sum of elements, each element being composed of an alternative or a variable and a coefficient. The rhs allows to specify the value of the right hand side. The values linked to each constraint can represent a credibility of this constraint in the eyes of a decision maker, for example.
Linear constraints on sets of alternatives.
A linear constraint related to sets of alternatives. It is composed of a sum of elements, each element being composed of a set of alternatives or a variable and a coefficient. The rhs allows to specify the value of the right hand side. The values linked to each constraint can represent a credibility of this constraint in the eyes of a decision maker, for example.
Definition of criteria.
Definition of a criterion. The id attribute is mandatory.
Indicates whether a criterion is active or not in the calculations. By default, if the tag is not present, the criterion is considered to be active (true).
Functions related to criteria. For example value functions.
Functions related to a criterion, to store for example a value function.
Thresholds related to criteria (typically discrimination thresholds).
Thresholds related to a criterion, to represent for example discrimination thresholds in outranking methods. The type of the threshold is specified in the mcdaConcept attribute of each threshold, and documented in each algorithm.
One or more discrimination thresholds related to a criterion.
Scales related to criteria (typically evaluation scales).
Evaluation scales related to a criterion.
One or more scales related to a criterion.
A reference to a scale defined anywhere in the parent criteriaScales.
Discrimination threshold (used to represent preferences on criteria).
The type of the affine threshold: either direct when the evaluation of the best of the 2 alternatives is taken into account in the affine formula, or inverse when the evaluation of the worst alternative is considered in the formula. If not present, it is assumed that the type is direct.
Definition of criteria sets.
A set of criteria, where the elements can be valued. An empty set of criteria can also be represented when no element is present.
A sequence of criteria-related values.
A sequence of criteria sets-related values.
One or more values related to a criterion.
One or more values related to a set of criteria.
Generic type for a matrix on criteria.
A row of the matrix.
A column of the matrix.
Generic type for a matrix on criteria sets.
A row of the matrix.
A column of the matrix.
Represents a set of linear constraints on the criteria.
A linear constraint related to criteria. It is constructed from a sum of elements, each element being composed of a reference to a criterion or a variable, and a multiplicative coefficient. The right hand side is specified separately, and the operator is either geq, leq or eq. Values can also be assigned to each constraint, to represent, e.g., a credibility of the constraint.
Represents a set of linear constraints on the criteria.
A linear constraint related to criteria. It is constructed from a sum of elements, each element being composed of a reference to a set of criteria or a variable, and a multiplicative coefficient. The right hand side is specified separately, and the operator is either geq, leq or eq. Values can also be assigned to each constraint, to represent, e.g., a credibility of the constraint.
A hierarchy of criteria.
A hierarchy of criteria sets.
A crosstable alternatives x criteria.
A row in a crosstable alternatives x criteria.
Definition of the categories.
Definition of a category. The id attribute is mandatory.
Indicates whether a category is active or not in the calculations. By default, if the tag is not present, the category is considered to be active (true).
A sequence of categories sets.
A set of categories, where each element can possibly be valued. An empty set of categories is obtained when no element is present.
A sequence of values related to categories.
A sequence of values related to sets of categories.
A value or values related to a category.
A value or values related to a set of categories.
A matrix of values related to categories.
A row of a matrix.
A matrix of values related to sets of categories.
A row of a matrix.
Represents a set of linear constraints on the categories.
A linear constraint related to categories. It is constructed from a sum of elements, each element being composed of a reference to a category or a variable, and a multiplicative coefficient. The right hand side is specified separately, and the operator is either geq, leq or eq. Values can also be assigned to each constraint, to represent, e.g., a credibility of the constraint.
Represents a set of linear constraints on sets of categories.
A linear constraint related to sets of categories. It is constructed from a sum of elements, each element being composed of a reference to a category set or a variable, and a multiplicative coefficient. The right hand side is specified separately, and the operator is either geq, leq or eq. Values can also be assigned to each constraint, to represent, e.g., a credibility of the constraint.
A sequence of category profiles
Description of a category by a bounding profile or a central profile.
A sequence of assignments of alternatives to categories.
An assignment of an alternative or a set of alternatives to a category, a set of categories or an interval of categories.
An interval of categories.
Evaluation scale. Choice between 3 types of scales: nominal, qualitative and quantitative.
Type of nominal scales, consiting in unordered labels.
The set of labels used by a nominal scale.
Type of qualitative ordered levels. It consists in a preference direction and either classical ranked labels or fuzzy labels (via valued labels in both cases).
Sequence of valued labels. Could either be labels to which are associated ranks, or fuzzy labels.
Type of quantitative scales. It consists in a preference direction (min or max), and a minimal and maximal quantitative evaluations (both of which are optional).
Preferential direction on a given quantitative or qualitative scale: either "max" for a scale where the higher values are preferred, or "min" for a scale where the lower values are preferred.
Restriction of the "value" type to simple values of type: integer, real, rational and NA.
General value type.
Sequence of values.
An interval consists in a lower bound and an upper bound, both of which are values. Neither the lower bound not the upper bound are mandatory, which allows to define unbounded intervals.
Bound of an interval.
If false (the default), the value is included within the interval, otherwise it is not.
Rational number described by the numerator and the denominator, which are both integers.
Label to which is associated a value. Can be used to represent ranks of ordinal labels, or fuzzy labels for example.
Restriction of the "function" type to functions compatible with fuzzy numbers.
Piecewise linear function defined by a sequence of segments.
Segment, defined by two points: one for the head, one for the tail. An attribute ("open") in the head and the tail allows to specify if the point is included in the segment or not.
Head point of a segment. All heads except the first one can be nil (no content): in that case, it is equal to the previous segment's tail.
Tail point of a segment..
Point defined by an abscissa and an ordinate.
The abscissa of a point.
The ordinate of a point.
An end point for segments.
If false (the default), the end point is included in the segment, otherwise it is not part of the segment.
An end point for segments which may have no coordinates. The meaning of an end point with no coordinates is defined by its parent.
The abscissa of a point.
The ordinate of a point.
If false (the default), the end point is included in the segment, otherwise it is not part of the segment.
A general function which can be of one of the following four types: either a constant, an affine function (defined by a slope and an intercept), a piecewise linear function (defined by a sequence of segments), or a set of points.
Affine function. Requires a slope and an intercept.
A discrete function defined by a sequence of points.
A crosstable alternatives x criteria showing the performances of the alternatives on the criteria.
The performances of an alternative on the criteria. Corresponds to a "row" in a performancetable.
Evaluation of an alternative on a criterion. If the evaluation is not available, put NA.
A variable (for linear constraints for example). The id is required to identify the variable.
A compound tag to store the definitions of the variables of a linear constraints system.
Input parameters for a program.
A parameter for a program.
Output of programs.
XMCDA generic method message structure.
.