Delphi is a common method for reaching group consensus, and
it is implemented through several rounds of opinion presentation
and anonymous feedbacks [25]. The decision makers will adjust
their opinions according to the anonymous feedbacks from the
previous round to reach a convergence. Although generally three
or four rounds will lead to the consensus, there is an obvious doubt
if it can satisfy the limit of pressing time. The dynamic decisionmaking
situation is also a big obstacle for designing the structured
interaction survey.
For the outranking methods like ELECTRE [26] and PROMETREE
[27], there are also many applications in GDSS. Such methods can
rank alternatives without converting the original values to normalized
ones, and the indifference and preference thresholds can be
taken into account. However, such methods request to make
pairwise comparisons between each two alternatives, which may
need a lot of time, especially when many restoration schemes need
to be evaluated and verified. As restoration proceeds, the pressing
time and large number of alternatives limits the utilization of such
kinds of methods. And the definition of indifference and preference
thresholds can also be regarded as their disadvantages in some
meaning, because the thresholds and models may introduce too
much subjectivity and uncertainty.
For the demerits of abovementioned methods, technique for
order preference by similarity to an ideal solution (TOPSIS) is a
good choice because it is easy to assimilate and implement [28].
The method is intuitively appealing for its visualization, at least
for two dimensions. The distances considering both the best and
worst solutions make TOPSIS reasonable and rational. And the
most important is that it exhibits the fewest rank reversals among
the common MADM methods, thus it can determine the restoration
schemes more stably [29].
Delphi is a common method for reaching group consensus, andit is implemented through several rounds of opinion presentationand anonymous feedbacks [25]. The decision makers will adjusttheir opinions according to the anonymous feedbacks from theprevious round to reach a convergence. Although generally threeor four rounds will lead to the consensus, there is an obvious doubtif it can satisfy the limit of pressing time. The dynamic decisionmakingsituation is also a big obstacle for designing the structuredinteraction survey.For the outranking methods like ELECTRE [26] and PROMETREE[27], there are also many applications in GDSS. Such methods canrank alternatives without converting the original values to normalizedones, and the indifference and preference thresholds can betaken into account. However, such methods request to makepairwise comparisons between each two alternatives, which mayneed a lot of time, especially when many restoration schemes needto be evaluated and verified. As restoration proceeds, the pressingtime and large number of alternatives limits the utilization of suchkinds of methods. And the definition of indifference and preferencethresholds can also be regarded as their disadvantages in somemeaning, because the thresholds and models may introduce toomuch subjectivity and uncertainty.For the demerits of abovementioned methods, technique fororder preference by similarity to an ideal solution (TOPSIS) is agood choice because it is easy to assimilate and implement [28].The method is intuitively appealing for its visualization, at leastfor two dimensions. The distances considering both the best andworst solutions make TOPSIS reasonable and rational. And themost important is that it exhibits the fewest rank reversals amongthe common MADM methods, thus it can determine the restorationschemes more stably [29].
การแปล กรุณารอสักครู่..
