improveAssignment: Post factum analysis: improve assignment
In kciomek/rorutadis: Robust Ordinal Regression UTADIS

Description Usage Arguments Value See Also Examples

This function calculates minimal rho by which alternative evaluations on selected criteria have to be multiplied for that alternative to be possibly (or necessarily) assigned to at least some specific class (rho >= 1). Note! This function works for problems with only non-negative alternative evaluations.

1 2	improveAssignment(alternative, atLeastToClass, criteriaManipulability, necessary, problem)

`alternative`	An alternative for assignment improvement.
`atLeastToClass`	Desired assignment.
`criteriaManipulability`	Vector containing a logical value for each criterion. Each value denotes whether multiplying by `rho` on corresponding criterion is allowed or not. At least one criterion has to be available for that manipulation.
`necessary`	Whether necessary or possible assignment is considered.
`problem`	Problem for which improvement will be performed.

Value of rho or NULL if given assignment is not possible in any scenario.

deteriorateAssignment

perf <- matrix(c(8, 2, 1, 7, 0.5, 0.9, 0.4, 0.5), ncol = 2)
problem <- buildProblem(perf, 3, FALSE, c('g', 'g'), c(0, 0))
problem <- addAssignmentsUB(problem, c(1, 2), c(2, 3))

# a_1 dominates a_4 and a_1 is assigned at most to class C_2
# How many times evaluations of a_4 should be improved
# that a_4 will be assigned possibly to class C_3?
rho <- improveAssignment(4, 3, c(TRUE, TRUE), FALSE, problem)